This action might not be possible to undo. Are you sure you want to continue?

“The Modal Assurance Criterion (MAC)”

Profesor: R. Avilés Asignatura: Dinámica y Fatiga Alumnos: Rivero González, Álvaro 18 de Diciembre de 2009

Modal Assurance Criterion (MAC)

2

Index

Abstract 1. State of the Art. Antecedents. review of existing methods 1.1. Historical Development of MAC 1.1.1. Modal Vector Orthogonality 1.1.2. Modal Vector Consistency 1.1.3. Modal Assurance Criterion (MAC) Zero 1.1.4. Modal Assurance Criterion (MAC) Unity 1.1.5. MAC Presentation Formats 1.2. Other Similar Assurance Criteria 1.3. Uses of the Modal Assurance Criterion 1.4. Abuses of the Modal Assurance Criterion 1.5. Current Developments 2. Software 2.1. LMS Virtual.Lab Overview 2.2. LMS Virtual.Lab Correlation 2.2.1. Systematic Validation from the bottom up 2.2.2. LMS Virtual.Lab Correlation. Features and Benefits 2.2.3. LMS Virtual.Lab Model Updating. Features and Benefits 4 5 5 6 8 10 11 12 14 17 18 20 21 21 25 26 27 29

Modal Assurance Criterion (MAC)

3

3. Uses for Aircraft Design and Testing/Certificating Companies 3.1. Using MSC/NASTRAN and LMS/PRETEST to find an optimal sensor placement for modal identification and correlation of aerospace structures 3.1.1. MSC/NASTRAN and LMS/PRETEST 3.1.2. Target Mode Selection 3.1.3. Sensor Placement 3.1.4. Shaker Positioning 3.2. Modal Test of L-610G Aeroplane Conclusions Bibliography

31

31 33 33 37 41 43 47 48

Modal Assurance Criterion (MAC)

4

Abstract

This report firstly reviews the development of the original modal assurance criterion (MAC) together with other related assurance criteria that have been proposed over the last twenty years. Some of the other assurance criteria that will be discussed include the coordinate modal assurance criterion (COMAC), the frequency response assurance criterion (FMAC), partial mode assurance criterion (PMAC) and modal assurance criterion using reciprocal modal vectors (MACRV). Several uses of MAC that may not be obvious to the casual observer will be identified; the common problems with the implementation and use of modal assurance criterion computations will also be identified. Afterward, the LMS Virtual.Lab is briefly presented, stressing the module related to the modal correlation, the LMS Virtual.Lab Correlation. Finally, this dossier introduces some particular uses of the model assurance criterion that could be interesting for aircraft design and testing/certificating companies.

Este informe revisa en primer lugar el desarrollo del criterio de confianza modal (MAC) junto con otros criterios de confianza o correlación relacionados que han sido propuestos a lo largo de los últimos veinte años. Otros criterios de correlación que serán tratados son el criterio de correlación modal por coordenadas (COMAC), el criterio de correlación de respuesta en frecuencia (FMAC), el criterio de correlación de modo parcial (PMAC) y el criterio de correlación modal utilizando vectores modales recíprocos (MACRV). Varios usos del MAC que pueden no ser obvios para el observador casual serán identificados; los problemas habituales con la implementación y uso de computaciones del criterio de confianza modal serán también identificados. Seguidamente, el LMS Virtual.Lab es brevemente presentado, haciendo hincapié en el módulo relacionado con la correlación modal, el LMS Virtual.Lab Correlation. Finalmente, este informe introduce algunos usos particulares del criterio de correlación modal que podrían resultar interesantes para empresas de diseño y de prueba/certificación de aeronaves.

it is necessary to present the nomenclature that will be used in the following paragraphs: L A* Ni No Ne Na 𝐻𝑝𝑞 (𝑤) 𝐻𝑝𝑞 (𝑤) 𝜑𝑞𝑟 𝜑𝑝𝑞𝑟 𝜑 𝑇 𝜑 𝐻 𝜑𝑟 𝜑𝑝𝑟 = Number of matching pairs of modal vectors. This first section will first review the historical development of the modal assurance criterion. Many times. 1. mode r. therefore. mode r. was the orthogonality check. just like ordinary coherence. = Number of analytical modal vectors. The orthogonality check. not always the best approach. = Modal coefficient for degree-of-freedom q. = Number of outputs (assumed to be larger than Ni). linear regression correlation coefficient. = Complex conjugate of A. when the modal assurance criterion was developed. = Modal coefficient for reference p.Modal Assurance Criterion (MAC) 5 1. = Number of inputs. typical abuses will be identified. . degree-of-freedom q. Other similar assurance criteria will then be identified although the list is not intended to be comprehensive. = Complex conjugate transpose (Hermitian) of {φ}. = Transpose of {φ}. an analytical model was not available which renders the orthogonality check impractical. It is important to recognize the mathematical similarity of these varied criteria in order to be certain that conclusions be correctly drawn from what is essentially a squared. coupled errors in the analytical model development.1. = Modal vector for mode r. State of the art. Historical Development of MAC The historical development of the modal assurance criteria originated from the need for a quality assurance indicator for experimental modal vectors that are estimated from measured frequency response functions. Antecedents. Typical uses of the modal assurance criterion will be discussed and finally. Review of existing methods The development of the modal assurance criterion over twenty years ago has led to a number of similar assurance criteria used in the area of experimental and analytical structural dynamics. = Synthesized frequency response function. The modal assurance criterion is a statistical indicator. = Number of experimental modal vectors. = Measured frequency response function. which can be very powerful when used correctly but very misleading when used incorrectly. however. The standard of the late 1970s. = Modal vector for reference p. mode r. the reduction of the analytical model and the estimated modal vectors into a single indicator and was. Before starting with the review of the methods.

1. a little modal assurance criterion (Little MAC). With this form of scaling. Variations of this process include using analytical modal vectors together with experimental modal vectors and the appropriately sized mass or stiffness matrix. In the internal development of the modal assurance criterion at the University of Cincinnati. For this reason. Modal Vector Orthogonality The primary method that has historically been used to validate an experimental modal model is the weighted orthogonality check comparing measured modal vectors and an appropriately sized (the size of the square weighting matrix must match the length and spatial dimension of the modal vector) analytical mass or stiffness matrix (weighting matrix). any further discussion of . the experimental modal vectors are used together with a mass matrix. In the traditional orthogonality check. stiffness or damping matrix. Theoretically. 1. to evaluate orthogonality between the experimental and analytical modal vectors. In the original thought process. It is important to recognize that this least squares based form of linear regression analysis yields an indicator that is most sensitive to the largest difference between comparative values (minimizing the squared error) and results in a modal assurance criterion that is insensitive to small changes and/or small magnitudes. This latter comparison is normally referred to as a pseudo-orthogonality check (POC). normally derived from a finite element model.Modal Assurance Criterion (MAC) 6 The original development of the modal assurance criterion was modeled after the development of the ordinary coherence calculation associated with computation of the frequency response function. normally derived from the same finite element model. this was considered an advantage since small modal coefficient values are often seriously biased by frequency response function (FRF) measurements or modal parameter estimation errors. the experimental modal vectors are used together with a mass matrix. to evaluate orthogonality of the experimental modal vectors. In practice.1 (10 percent of the diagonal terms).1. normally derived from a finite element model. the off-diagonal values in the modal mass matrix are expected to be less than 0. Since the “Big Mac” acronym was already in use at that time. Structural Dynamics Research Lab (UCSDRL). MAC is the designation that has persisted. The modal assurance criterion that survives today is what was originally identified as Big MAC. and the analytical modal vectors. a big modal assurance criterion (Big MAC) and a multiple modal assurance criterion (MultiMAC) were formulated as part of the original development. these matrices are made available by way of a finite element analysis and normally the mass matrix is considered to be the most accurate. In the pseudo-orthogonality check. each modal vector of a system will be orthogonal to all other modal vectors of that system when weighted by the mass. The experimental and analytical modal vectors are scaled so that the diagonal terms of the modal mass matrix are unity. Little MAC and Multi-MAC were primarily testing methods and are not discussed further here. for the case of proportional damping.

Eq. the mass matrix can be invalid. 1 is not sufficiently satisfied. First. Note that. even under this criterion. As a result. it is important to try to develop methods that indicate confidence that the modal vector is. This can be due to measurement errors or problems with the modal parameter estimation algorithms. The accepted criterion in the aerospace industry. . 1 and 2 or by applying an equivalent procedure involving the state-space form of the weighted orthogonality relationship. When the orthogonality conditions are not satisfied. From an experimental point of view. This is a very common assumption and many times contributes to the problem. This complication has to be resolved by a process of real normalization of the measured modal vectors prior to utilizing Eqs.Modal Assurance Criterion (MAC) 7 orthogonality will be made with respect to mass matrix weighting. one (or more) of three situations may exist. Second. experimental modal vectors are not always real-valued and Eqs. this probably contributes significantly to the problem. It is a common procedure to form the modal vectors into a normalized set of mode shape vectors with respect to the mass matrix weighting. 1 and 2 are developed based upon normal or real-valued modal vectors. the reduction of the mass matrix can be invalid.1. This normally involves some sort of reduction algorithm as well as interpolation of grid locations to match the measurement situation. this result does not indicate where the problem originates. in general. is for all of the generalized mass terms to be unity and all cross orthogonality terms to be less than 0. The most obvious example of this situation would be when a relatively large amount of mass is reduced to a measurement location that is highly flexible. In Eqs. the result of zero for the cross orthogonality calculations (r ≠ s. an attempt is made to adjust the modal vectors so that the cross orthogonality conditions are satisfied. Since the mass matrix does not always represent the actual physical properties of the system when it is built or assembled. the orthogonality relations can be stated as follows: For r ≠ s: For r = s: 𝜑𝑟 𝜑𝑟 𝑇 𝑇 𝑀 𝜑𝑠 = 𝑂 𝑀 𝜑𝑠 = 𝑀𝑟 (1) (2) Experimentally. all three situations contribute to the failure of orthogonality or pseudoorthogonality criteria on occasion. 2. part of the problem. the mass matrix must be an No×No matrix corresponding to the measurement locations on the structure. Third. Often. such as the center of an unsupported panel. In all probability. where this confidence check is made most often. the modal vectors can be invalid. When Eq. This means that the finite element mass matrix must be modified from whatever size and distribution of grid locations required in the finite element analysis to the No×No square matrix corresponding to the measurement locations. This can certainly be a realistic problem and cause severe errors. but may represent only incidental motion of the overall modal vector. the measurement location is weighted very heavily in the orthogonality calculation of Eq. or is not. 1 and 2. In such a situation. 1) can rarely be achieved but values up to one tenth of the magnitude of the generalized mass of each mode are considered to be acceptable.

2. different subsets of the frequency response function matrix and/or different modal parameter estimation algorithms are utilized to estimate separate.1. differenced or sorted to provide a best estimate of the modal vector or to provide an indication of the type of error vector superimposed on the modal vector. the modal vectors can be compared and contrasted through an evaluation that consists of the calculation of a complex modal scale factor (relating two modal vectors) and a scalar modal assurance criterion (measuring the consistency or linearity between two modal vectors). In any case. In these cases. elements of each vector can be averaged (with or without weighting). The function of the modal assurance criterion (MAC) is to provide a measure of consistency (degree of linearity) between estimates of a modal vector. In terms of modern. The modal scale factor and the modal assurance criterion also provide a method of easily comparing estimates of modal vectors originating from different sources. multiple reference modal parameter estimation algorithms. If the modal vectors are not well represented in the frequency response function matrix. The function of the modal scale factor (MSF) is to provide a means of normalizing all estimates of the same modal vector. The modal vectors from a finite element analysis can be compared and contrasted with those determined experimentally as well as modal vectors determined by way of different experimental or modal parameter estimation methods. . the modal scale factor is a normalized estimate of the modal participation factor between two references for a specific mode of vibration. the analytical modal vector. The estimation of modal vectors from this frequency response function matrix will be a function of the data used in the modal parameter estimation algorithms and the specific modal parameter estimations algorithms used. Frequently. redundant modal vectors for comparison purposes. taking into account magnitude and phase differences. Once two different modal vector estimates are scaled similarly. can be used to complete the experimental modal vector if some degrees of freedom could not be measured. In this approach. if different estimates of the same modal vectors are generated. Modal Vector Consistency Since the frequency response function matrix contains redundant information with respect to a modal vector. the consistency of the estimate of the modal vector under varying conditions such as excitation locations (references) or modal parameter estimation algorithms can be a valuable confidence factor to be utilized in the process of evaluation of experimental modal vectors. the estimation of the modal vector will contain potential bias and variance errors. together with the modal scale factor. If an analytical and an experimental vector are deemed consistent or similar.Modal Assurance Criterion (MAC) 8 1. The common approach to estimation of modal vectors from frequency response functions is to measure several complete rows or columns of the frequency response function matrix. the modal vectors will contain potential variance errors. methods can be compared and contrasted in order to evaluate the mutual consistency of different procedures rather than estimating the modal vectors specifically. This provides an additional confidence factor in the evaluation of a modal vector from different excitation (reference) locations or different modal parameter estimation algorithms.

Modal Assurance Criterion (MAC) 9 The modal scale factor is defined. The second part is the part that is not correlated with modal vector d and is made up of contamination from other modal vectors and any random contribution. as follows: 𝑁 0 ∗ 𝑞 =1 𝜑 𝑐𝑞𝑟 𝜑 𝑑𝑞𝑟 𝑁 0 ∗ 𝑞=1 𝜑 𝑑𝑞𝑟 𝜑 𝑑𝑞𝑟 𝑀𝑆𝐹𝑐𝑑𝑟 = (3a) or: 𝑀𝑆𝐹𝑐𝑑𝑟 = {𝜑 𝑐𝑟 }𝑇 {𝜑𝑑𝑟 } ∗ 𝑑𝑟 𝑑𝑟 {𝜑 }𝑇 {𝜑 ∗ } (3b) Since the modal scale factor is a complex-valued scalar. this is also equivalent to: {𝜑 }𝐻 {𝜑 ∗ } 𝑑𝑟 𝑐𝑟 𝑀𝑆𝐹𝑐𝑑𝑟 = {𝜑 𝑑𝑟 }𝐻 {𝜑 ∗ } 𝑑𝑟 (3c) Eq. this is also equivalent to: 𝑀𝐴𝐶𝑐𝑑𝑟 = or: 2 ∗ 𝜑 𝑑𝑟 𝐻 𝜑 𝑐𝑟 𝜑 𝑑𝑟 𝐻 𝜑 𝑑𝑟 𝜑 𝑐𝑟 𝐻 𝜑 𝑐𝑟 (4c) . The first part is the part correlated with modal vector d. This error vector is considered to be noise. The modal assurance criterion is defined as a scalar constant relating the degree of consistency (linearity) between one modal and another reference modal vector as follows: 2 𝑁 0 ∗ 𝑞 =1 𝜑 𝑐𝑞𝑟 𝜑 𝑑𝑞𝑟 𝑁 0 𝑁 0 ∗ 𝑞=1 𝜑 𝑐𝑞𝑟 𝜑 𝑐𝑞𝑟 𝑞 =1 𝜑 𝑑𝑞𝑟 𝑀𝐴𝐶𝑐𝑑𝑟 = or: 𝑀𝐴𝐶𝑐𝑑𝑟 = ∗ 𝜑 𝑑𝑞𝑟 (4a) 2 ∗ 𝜑 𝑐𝑟 𝑇 𝜑 𝑑𝑟 ∗ ∗ 𝜑 𝑐𝑟 𝑇 𝜑 𝑐𝑟 𝜑 𝑑𝑟 𝑇 𝜑 𝑑𝑟 (4b) Since the modal assurance criterion is a real-valued scalar. modal vector c can be considered to be made of two parts. In the general case. 3 implies that the modal vector d is the reference to which the modal vector c is compared. according to this approach.

this is not delineated by the modal assurance criterion. is bounded between zero and one.3. exist in all modal vector estimates. Note that. unlike the orthogonality calculations. to one – representing a consistent correspondence. linear relationship. The modal assurance criterion can only indicate consistency. There is noise on the reference modal vector.Modal Assurance Criterion (MAC) 10 𝑀𝐴𝐶𝑐𝑑𝑟 = or: 𝜑 𝑑𝑟 𝐻 𝜑 𝑐𝑟 𝜑 𝑐𝑟 𝐻 𝜑 𝑑𝑟 𝜑 𝑑𝑟 𝐻 𝜑 𝑑𝑟 𝜑 𝑐𝑟 𝐻 𝜑 𝑐𝑟 (4d) 𝑀𝐴𝐶𝑐𝑑𝑟 = 𝑀𝑆𝐹𝑑𝑟 𝑀𝑆𝐹𝑑𝑟𝑐 (4e) The modal assurance criterion takes on values from zero – representing no consistent correspondence. the modal assurance criterion is normalized by the magnitude of the vectors and. Even though the modal assurance criterion is unity. This can occur if the system is nonlinear and two data sets have been acquired at different times or excitation levels. This case is the same as noise on the input of a frequency response function measurement. thus. 1. Invalid assumptions are normally the cause of this sort of potential error. this is an indication that the modal vectors are not consistent. the assumptions involving the system or the modal parameter estimation techniques are not necessarily correct. If the same errors. random or bias. In this manner. The modal parameter estimation algorithms will also not handle the different nonlinear characteristics in a consistent manner. The modal parameter estimation is invalid. Modal Assurance Criterion (MAC) Zero If the modal assurance criterion has a value near zero. No amount of signal processing can remove this type of error. the modal assurance criterion should approach unity and the value of the modal scale factor can be considered reasonable. .1. not validity or orthogonality. The frequency response function measurements may contain no errors but the modal parameter estimation may not be consistent with the data. System nonlinearities will appear differently in frequency response functions generated from different exciter positions or excitation signals. The assumptions may cause consistent errors in all modal vectors under all test conditions verified by the modal assurance criterion. if the modal vectors under consideration truly exhibit a consistent. This can be due to any of the following reasons: The system is nonstationary.

This would be the situation if.4. which is a function of the common modal coefficients between the rows or columns. If the first four reasons can be eliminated. The modal vectors can be consistent for any of the following reasons: The modal vectors have been incompletely measured. Therefore. this modal vector may not be one of the true modal vectors of the system. The modal vectors represent the same modal vector with different arbitrary scaling. The modal vectors are primarily coherent noise. this is an indication that the modal vectors are consistent. the modal assurance criterion will only reflect a consistent (linear) relationship to the reference modal vector. This can be done by computing the modal assurance criterion between Ne modal vectors estimated from experimental data and Na modal vectors estimated from a finite element analysis evaluated at common stations. Hopefully. Modal Assurance Criterion (MAC) Unity If the modal assurance criterion has a value near unity. This situation can occur whenever too few response stations have been included in the experimental determination of the modal vector. Since the reference modal vector may be arbitrarily chosen. this measure of inconsistency will imply that the modal vectors are orthogonal.Modal Assurance Criterion (MAC) 11 The modal vectors are from linearly unrelated mode shape vectors. Under the constraints mentioned previously. the modal assurance criterion indicates that the modal scale factor is the complex constant relating the modal vectors and that the modal scale factor can be used to average. If the two modal vectors being compared have the same expected value when normalized. 1. the modal assurance criterion can be interpreted in a similar way as an orthogonality calculation. In any case. It could simply be a random noise vector or a vector reflecting the bias in the modal parameter estimation algorithm. The modal vectors are the result of a forced excitation other than the desired input. This does not necessarily mean that they are correct. if the first three reasons can be eliminated. This process results in a Ne×Na rectangular modal assurance criterion matrix with values that .1. during the measurement of the frequency response function. the two modal vectors should differ only by the complex valued scale factor. since the different modal vector estimates are from different excitation positions. a rotating piece of equipment with an unbalance is present in the system being tested. The modal assurance criterion can be used to verify or correlate an experimental modal vector with respect to a theoretical modal vector (eigenvector). difference or sort the modal vectors. the modal assurance criterion can be applied in many different ways.

The modal assurance criterion can be used to evaluate modal parameter estimation methods if a set of analytical frequency response functions with realistic levels of random and bias errors is generated and used in common with a variety of modal parameter estimation methods. Once the modal assurance criterion establishes that two vectors represent the same information. provide a useful tool in the processing of experimental modal vectors. assuming the modal assurance function indicates consistency. the vectors can be averaged. the error pattern often gives an indication that the error originates due to the location of the excitation or due to an inadequate modal parameter estimation process. a table of numbers was usually presented as shown in Table 1. MAC Presentation Formats One of the big changes in the application of the Modal Assurance Criteria over the last twenty years is in the way the information is presented. differenced or sorted to determine the best single estimate or the potential source of contamination using the modal scale factor. The computation of modal scale factor and modal assurance criterion results in a complex scalar and a correlation coefficient that does not depend on weighting information outside the testing environment. . The concept of consistency in the estimate of modal vectors from separate testing constraints is important considering the potential of multiple estimates of the same modal vector from numerous input configurations and modal parameter estimation algorithms. Since the modal scale factor and modal assurance criterion are computed analogous to the frequency response function and coherence function. into a complete estimate of each modal vector at all measurement stations. These characteristics. If the error appears to be random and the modal assurance criterion is high.1. If the error appears to be biased or skewed. modal vectors can be sorted. Based upon partial but overlapping measurement of two columns of the frequency response function matrix. Additionally. the modal vectors can be averaged (using the modal scale factor) to improve the estimate of a modal vector. agreement between existing methods can be established and new modal parameter estimation methods can be checked for characteristics that are consistent with accepted procedures. as well as others.5. this approach can be used to evaluate the characteristics of each modal parameter estimation method in the presence of varying levels of random and bias error. Historically.Modal Assurance Criterion (MAC) 12 approach unity whenever an experimental modal vector and an analytical modal vector are consistently related. Since the modal scale factor is a complex scalar that allows two vectors to be phased the same and to the same mean value. these vectors can be subtracted to evaluate whether the error is random or biased. both the advantages and limitations of the computation procedure are well understood. 1. In this way.

that MAC is a discrete calculation and what appears as a color contour plot really only represents the discrete mode to mode comparison.2-D and 3-D presentation of MAC values. however.. It is important to remember. Figure 1.Numerical presentation of MAC values. . Nevertheless.. Today.Modal Assurance Criterion (MAC) 13 Table 1. most computer systems routinely utilize color to present magnitude data like MAC using a 2D or 3D plot as shown in Figures 1 and 2. a color plot does allow for more data to be presented in an understandable form in a minimum space.

which tends to highlight the cross terms (off diagonal) that are generally very small MAC values. This is required since the modal assurance criterion minimizes the squared error and is dominated by the larger values. In this case. and to adjust the modal assurance criterion to weight the degrees-of-freedom in the modal vectors accordingly. the WMAC becomes a unity normalized orthogonality – or pseudo-orthogonality – check where the desirable result for a set of modal vectors would be ones along the diagonal (same modal vectors) and zeros offdiagonal (different modal vectors) regardless of the scaling of the individual modal vectors. least squares computation approach to the analysis (projection) of two vector spaces as the modal assurance criterion. The equations for each assurance criterion are not repeated unless there is a significant computational difference that needs to be clarified or highlighted. The subset is chosen based upon the user‟s interest and may reflect only a certain dominant sensor direction (X. WMAC is proposed for these cases. This development is needed whenever different data types (with different engineering units) are included in the same modal vector to normalize the magnitude differences in the vectors. Note that the weighting matrix is applied as an inner matrix product for the single numerator vector product and both vector products in the denominator. Other Similar Assurance Criteria The following brief discussion highlights assurance criteria that utilize the same linear. Essentially this approach utilizes the square root of the MAC calculation.Modal Assurance Criterion (MAC) 14 1. Partial Modal analysis Criterion (PMAC): The partial modal assurance criterion (PMAC) was developed as a spatially limited version of the modal assurance criterion where a subset of the complete modal vector is used in the calculation. . Weighted Modal Analysis Criterion (WMAC): A number of authors have utilized a weighted modal assurance criterion (WMAC) without developing a special designation for this case. just sensor distribution. Y and/or Z) or only the degrees-offreedom from a component of the complete modal vector.2. Modal Assurance Criterion Square Root (MACSR): The square root of the modal assurance criterion (MACSR) is developed to be more consistent with the orthogonality and pseudoorthogonality calculations using an identity weighting matrix. This list is by no means comprehensive nor is it in any particular order of importance but includes most of the frequently cited assurance criterion found in the literature. Scaled Modal Assurance Criterion (SMAC): The scaled modal assurance criterion (SMAC) is essentially a weighted modal assurance criterion (WMAC) where the weighting matrix is chosen to balance the scaling of translational and rotational degrees-of-freedom included in the modal vectors. The purpose of the weighting matrix is to recognize that MAC is not sensitive to mass or stiffness distribution.

The coordinate modal assurance criterion (COMAC) is calculated using the following approach. The two modal vectors in each mode pair represent the same modal vector. Modal Assurance Criterion with Frequency Scales (FMAC): Another extension of the modal assurance criterion is the addition of frequency scaling to the modal assurance criterion. therefore. The COMAC is calculated over a set of mode pairs. the degree of spatial aliasing and the frequency comparison in a single plot. The enhanced coordinate modal assurance criterion (ECOMAC) was developed to extend the COMAC computation to be more aware of typical experimental errors that occur in . experimentally based.Modal Assurance Criterion (MAC) 15 Modal Assurance Criterion Using Reciprocal Vectors (MACRV): A reciprocal modal vector is defined as the mathematical vector that. The reciprocal modal vector can be thought of as a product of the modal vector and the unknown weighting matrix that will produce a perfect orthogonality result. assessment of parameter variation. there will be a value of COMAC computed for each (measurement) degree-of-freedom. For two sets of modes that are to be compared. The modal assurance criterion using reciprocal modal vectors (MACRV) is the comparison of reciprocal modal vectors with analytical modal vectors in what is very similar to a pseudoorthogonality check (POC).). when transposed and premultiplied by a specific modal vector.” This development is particularly useful in model correlation applications (model updating. Reciprocal modal vectors are computed directly from measured frequency response functions and the experimental modal vectors and are. etc. The Enhanced Coordinate Modal Assurance Criterion (ECOMAC): One common problem with experimental modal vectors is the potential problem of calibration scaling errors and/or sensor orientation mistakes. The COMAC attempts to identify which measurement degrees-of-freedom contribute negatively to a low value of MAC. once the mode pairs have been identified with MAC or some other approach: 𝐶𝑂𝑀𝐴𝐶𝑞 = 𝐿 𝑟=1 𝜑 𝑞𝑟 𝜙 𝑞𝑟 2 𝐿 ∗ 𝑟=1 𝜑 𝑞𝑟 𝜑 𝑞𝑟 𝐿 ∗ 𝑟=1 𝜑 𝑞𝑟 𝜑 𝑞𝑟 (5) Note that the above formulation assumes that there is a match for every modal vector in the two sets and the modal vectors are renumbered accordingly so that the matching modal vectors have the same subscript. analytical versus analytical. experimental versus experimental or experimental versus analytical. This extension of MAC “offers a means of displaying simultaneously the mode shape correlation. but the set of mode pairs represents all modes of interest in a given frequency range. the result is zero. Only those modes that match between the two sets are included in the computation. Coordinate Modal Assurance Criterion (COMAC): An extension of the modal assurance criterion is the coordinate modal assurance criterion (COMAC). The reciprocal modal vectors are utilized in controls applications as modal filters and the MACRV serves as a check of the mode isolation provided by each reciprocal modal vector compared to analytical modes expected. yields unity. When the same computation is performed with this reciprocal modal vector and any other modal vector or any other reciprocal modal vector.

The simplest example is a validation procedure that compares the FRF data synthesized from the modal model with the measured FRF data. This approach has been utilized in the modal parameter estimation process for a number of years under various designations (parameter estimation correlation coefficient. Modal Correlation Coefficient (MCC): One of the natural limitations of a least squares based correlation coefficient like the modal assurance criterion is that it is relatively insensitive to small changes in magnitude. etc. Mutual Correspondence Criterion (MCC): The mutual correspondence criterion (MCC) is the modal assurance criterion applied to vectors that do not originate as modal vectors but as vector measures of acoustic information (velocity. 𝑤 2 𝑤 =𝑤 1 𝐻𝑝𝑞 𝑤 2 ∗ 𝑤 =𝑤 1 𝐻𝑝𝑞 𝑤 𝐻𝑝𝑞 (𝑤 ) 2 ∗ 𝑤 𝐻𝑝𝑞 (𝑤 ) 𝐹𝑅𝐴𝐶𝑝𝑞 = ∗ 𝑤 𝐻𝑝𝑞 (𝑤) 𝑤 2 𝑤 =𝑤 1 𝐻𝑝𝑞 (6) . The equation in this formulation utilizes a transpose and will only correctly apply to real valued vectors. Inverse Modal Assurance Criterion (IMAC): An alternative approach to increasing the sensitivity of the modal assurance criterion to small mode shape changes is the inverse modal assurance criterion (IMAC). Naturally. Frequency Response Assurance Criterion (FRAC): Any two frequency response functions representing the same input-output relationship can be compared using a technique known as the frequency response assurance criterion (FRAC). synthesis correlation coefficient and response vector assurance criterion (RVAC)). in the vector comparisons. pressure. small modal coefficients become significant in the least squares based correlation coefficient computation. The basic assumption is that the measured frequency response function and the synthesized frequency response function should be linearly related (unity scaling coefficient) at all frequencies. This serves as an independent check of the modal parameter estimation process. Naturally. the FRFs can be compared over the full or partial frequency range of the FRFs as long as the same discrete frequencies are used in the comparison. The modal correlation coefficient (MCC) is a modification of MAC that attempts to provide a more sensitive indicator. position by position. This procedure is particularly effective as a modal parameter estimation validation procedure if the measured data were not part of the data used to estimate the modal parameters. intensity. This approach uses essentially the same computational scheme as MAC but utilizes the inverse of the modal coefficients. Therefore.). This approach is particularly important when using modal vectors in damage detection situations where the magnitude changes of the modal vectors being measured are minimal. this computation suffers from the possibility that a modal coefficient could be numerically zero.Modal Assurance Criterion (MAC) 16 defining modal vectors such as sensor scaling mistakes and sensor orientation (plus or minus sign) errors.

combines analytical modal vectors with measured frequency response functions (FRFs) in an extension of FRAC and FDAC that weights or filters the FRF data based upon the expected. the CCF will detect the common problem of a constant phase shift that might be due to experimental signal conditioning problems. Correlation with analytical modal models (mode pairing). analytical modal vectors. which is computed without squaring the numerator term. comparing analytical to analytical modal vectors. Frequency Domain Assurance Criterion (FDAC): A similar variation in the frequency response assurance criterion is the frequency domain assurance criterion (FDAC). A partial list of the most typical uses that have been reported in the literature are as follows: Validation of experimental modal models. comparing measured to analytical modal vectors. Coordinate Orthogonality Check (CORTHOG): The coordinate orthogonality check (CORTHOG) is a normalized error measure between the pseudo-orthogonality calculation. which is a FRACtype of calculation evaluated with different frequency shifts. etc. Several different normalizing or scaling methods are used with this calculation. 1. In situations where analytical and experimental FRFs are compared. the FDAC is formulated to identify this problem. The magnitude of the coefficient is the same as the FRAC computation but the phase describes any systematic phase lag or lead that is present between the two FRFs. Correlation with operating response vectors. and the analytical orthogonality calculation. Modal vector error analysis. . the modal FRF assurance criterion (MFAC).Modal Assurance Criterion (MAC) 17 Complex Correlation Coefficient (CCF): A significant variation in the frequency response assurance criterion is the complex correlation coefficient (CCF).3. Since the difference in impedance (FRF) model updating is often an FRF that is in question due to frequencies of resonances or anti-resonances. A related criterion. Uses of the Modal Assurance Criterion Most of the potential uses of the modal assurance criterion are well known but a few may be more subtle. Mapping matrix between analytical and experimental modal models. thus yielding a complex valued coefficient.

These issues can be summarized as: The modal analysis criterion is not an orthogonality check. Experimental modal vector completion and/or expansion. The modal vectors have been zero padded. Modal vector consistency/stability in modal parameter estimation algorithms. . Abuses of the Modal Assurance Criterion Many of the alternate formulations of the modal assurance criterion were developed to address some of the shortcomings of the original modal assurance criterion formulation.Modal Assurance Criterion (MAC) 18 Modal vector averaging. The number of elements in the modal vectors (space) is small. Structural fault/damage detection. Weighting for model updating algorithms. a poor result will often follow. These issues can be further explained in the following paragraphs. The wrong mathematical formulation for the modal assurance criterion is used.4. Optimal sensor placement. 1. The misuse or abuse of the modal assurance criterion generally results due to one of five issues. When users utilize the original modal assurance criterion in these situations. For the purposes of this discussion. this is referred to as misuse or abuse. The modal assurance criterion is sensitive to large values (wild points?) and insensitive to small values. Quality control evaluations. Repeated and pseudo-repeated root detection.

This means that the degrees-of-freedom involving the largest magnitude differences between the two modal vectors will dominate the computation while small differences will have almost no effect. these wild points may dominate the MAC calculation. if the modal vectors have only a limited number of degrees-offreedom. In these situations. users implement the modal assurance criterion. there may be great variability in the MAC computation.. this does not give the correct result. The typical example involves the engine of an automobile. therefore. several modal vectors at different modal frequencies will have very high MAC numbers indicating that the modal vectors are the same. users who do not recognize this issue are often led astray in subsequent applications involving complex-valued vectors. multiple reference modal parameter estimation algorithms estimate the stability or consistency diagram. This innocent error often occurs when the author is utilizing real-valued vectors and notices no problem. If few or no sensors are placed on the engine and a large number are placed on the surface of the automobile body. The modal assurance criterion is sensitive to large values (wild points?) and insensitive to small values: The modal assurance criterion is based upon the minimization of the squared error between two vector spaces.Modal Assurance Criterion (MAC) 19 The modal analysis criterion is not an orthogonality check: It is important to recognize that the modal assurance criterion effectively weights the computation based upon the spatial distribution of the degrees-of-freedom included in the modal vectors. Therefore. The number of elements in the modal vectors (space) is small: Since the modal assurance criterion is essentially a statistical computation where the number of averages comes from the number of elements in the modal vectors. where some of the vectors are complex-valued. However. in the general case. However. This error causes no problem as long as analytical vectors or real-valued experimental vectors are involved in the calculation. This also means that. typically two to five. . modal parameter estimation mistakes. The wrong mathematical formulation for the modal assurance criterion is used: Frequently. this will skew the meaning of the numerical MAC value. particularly if the modal vector is not well excited from one or more of the reference locations. etc. nodal information (small modal coefficients) will generally not have much effect on the MAC calculation and large modal coefficients will potentially have the greatest effect. This example indicates to the user that an incomplete modal vector was measured and the user has violated one of the primary assumptions of experimental modal analysis (observability). The modal assurance criterion does not weight the modal vectors with a mass or stiffness matrix and. using a vector transpose in the numerator and denominator calculations rather than an Hermitian (conjugate transpose). Vectors with many elements reduce the sensitivity of MAC to this problem. Modal vector stability or consistency is identified using a MAC computation where the vectors include only the degrees-of-freedom at the reference locations. or a related similar computation. The original mathematical formulation assumes the general case but has been reported incorrectly in some literature. cannot compensate for situations where a very limited number of degrees-of-freedom (sensors) have been placed on a massive substructure of a mechanical system. This frequently happens when high order. if there have been erroneous data included in the modal vectors due to calibration errors.

1. This approach extends to the modal assurance criterion as well. When the modal assurance criterion is calculated for this case. this is the case since there is no designation for not measuring the information. many users are utilizing more statistical approaches to understand the meaning and bounds of experimental modal parameters. in an experimental situation. when modal vectors are exported from one computational environment to another. the modal vectors include zero values when no value was ever measured.5.Y) of translational response may be measured at some degrees-offreedom rather than three dimensions (X. for that degree-of-freedom. but is rarely done. In this way. with nonzero information at these degrees-of-freedom. . In the commonly used Universal File Format for modal vectors (File Format 55). there will be a problem if some other vectors. are included in the computation. jackknife uses sequential elimination) to evaluate the bounds or limits on the MAC values.Y.Modal Assurance Criterion (MAC) 20 The modal vectors have been zero padded. or computed.Z). the sensitivity of the MAC computation can be more effectively evaluated than with the current single number indicating the degree of linearity between two modal vectors that are being compared. This can be avoided if information is dropped from the computation when either vector includes a perfect zero (within computational precision) at a degree-of-freedom. These approaches remove and/or replace portions of the computation (bootstrap uses replicative resampling. Frequently. For example. Examples are the bootstrap and jackknife approaches to the evaluation of the mean and standard deviation of discrete sets of experimental data. Current Developments Currently. one (X) or two dimensions (X.

Software: One of the most used computer program. LMS Virtual. quickly assess multiple design alternatives and optimize designs before prototype construction.Lab. time and frequency functions and much more.Lab Desktop also offers a complete visualization environment for part and assembly models. With LMS Virtual. users have seamless access to models and load data. LMS Virtual.Lab Desktop.Lab Desktop.Lab offers an integrated software suite to simulate and optimize the performance of mechanical systems for structural integrity. system dynamics and durability.1. functional performance engineering data.LMS Virtual.Lab.Modal Assurance Criterion (MAC) 21 2. Figure 2.Lab Desktop: LMS Virtual. and recommended by the Head of Dynamics Programs of the CTA. 2. noise and vibration.Lab Overview LMS Virtual.Lab Desktop provides a common environment for multiple functional performance applications. .. geometry and simulation models from industry-standard CAD and CAE tools as well as test data. engineering teams can build accurate simulation models. simulate their real-life performance. Using LMS Virtual. LMS Virtual. is the LMS Virtual. LMS Virtual.Lab covers all the process steps and required technologies to perform an end-to-end design assessment in each key discipline.

LMS Virtual. Figure 3. LMS Virtual.Lab Structures: LMS Virtual. The resulting loads can also be used for structural analysis.Lab Structures.Lab Structures offers full meshing capabilities and captures the complete modeling and analysis process from CAD drawing to multiattribute simulation results. The resulting simulation is excellent input to optimize the design‟s dynamic performance. CATIA CAE and Nastran (MD. Users can easily create a complete and accurate system model from scratch or import geometry models from any industry-standard CAD system. completely integrated solution to build multibody models that simulate the full-motion behavior of complex mechanical system designs. It offers multi-solver support for Abaqus. durability.Lab Motion offers a highly efficient.. integrating advanced model creation and manipulation tools to efficiently generate component. LMS Virtual.LMS Virtual.. Figure 4.Lab Motion: LMS Virtual. and noise and vibration studies.Lab Motion.Lab Motion applies forces and motion to simulate the actual operational behavior of the new design. Ansys.Lab Structures offers a scalable solution for structural modeling and analysis. subsystem and full-system models. .LMS Virtual. NEi). NX.Modal Assurance Criterion (MAC) 22 LMS Virtual. MSC.

transmission loss through panels.LMS Virtual. refine and optimize the vibro-acoustic behavior of a design. aero-acoustic phenomena and much more. Convenient modeling capabilities combined with efficient solvers and easy-to-interpret visualization tools enable users to quickly gain insight to the acoustic performance of their product..Lab Acoustics.Lab Acoustics offers an integrated solution to minimize radiated noise or optimize the sound quality in new designs. It offers all the required tools to create system-level models. LMS Virtual.Lab Acoustics: LMS Virtual. .LMS Virtual. LMS Virtual.Modal Assurance Criterion (MAC) 23 LMS Virtual.Lab Noise and Vibrations: LMS Virtual.Lab Noise and Vibration is developed to efficiently analyze. It includes a wide range of visualization and analysis tools to analyze noise and vibration performance and accurately pinpoint the most critical contributors to noise and vibration issues. Figure 6. Convenient tools enable engineers to quickly perform design modifications and assess the noise and vibration performance of a design variant in minutes.. Figure 5. engine acoustics. build realistic load cases and simulate noise and vibration responses.Lab Acoustics simulates both internal and external acoustic radiation and offers dedicated applications for structural noise radiation.Lab Noise and Vibrations.

LMS Virtual. This immediate insight enables engineering teams to validate more design variants for fatigue life within ever-shorter development cycles.Lab Durability allows engineers to design reliable products right from the start.Modal Assurance Criterion (MAC) 24 LMS Virtual. Figure 8..Lab Correlation: LMS Virtual. It offers direct access to standard FE and test data formats and a unique export to LMS Test.Lab. It predicts fatigue hotspots and system-level fatigue life by combining dynamic component loads with stress results and fatigue material parameters.Lab Durability. LMS Virtual.Lab quickly compares and validates FE models to test data and identifies possible modeling errors to systematically improve existing simulation models.LMS Virtual.LMS Virtual.Lab Correlation. LMS Virtual.Lab Correlation allows users to combine test-based and virtual component models into system-level models for more productive simulation.. .Lab Durability provides direct feedback regarding critical fatigue areas and the root cause of fatigue problems.Lab Durability: LMS Virtual. Figure 7.

A comprehensive tool set significantly facilitates simulated and measured mode shape comparison and operational deflection shapes and response functions. Correlating structural characteristics: Although static physical tests serve many design purposes. 2. connections. models used for vibro-acoustic simulations usually require systematic test-based validation of dynamic properties. subsystem and full-system models to be compared with experimental data.LMS Virtual. and the required number of excitation and response points helps avoid testing errors and redundancy.Lab Optimization automatically explores a multitude of design alternatives using design of experiment and response surface modeling techniques. Building and validating system models from the bottom up is the only way to prevent accumulating inaccuracies. LMS Virtual.Lab Correlation helps correlate physical test results and prepare structural tests. it is essential that simulation models meet stringent accuracy standards. LMS Virtual. validated models provide a better understanding of assumptions made regarding material properties. Besides more reliable what-if analyses. LMS Virtual..Modal Assurance Criterion (MAC) 25 LMS Virtual. Ensuring reliable simulation results requires component.Lab Optimization lets design and engineering teams automatically select the optimal design while accounting for multiple performance targets. or alternatively validated models of similar structures.2. Using the original FE model as a basis to provide optimal comparison positions. Users can easily identify the key variables that have the most influence on the functional performance of a mechanical system. .Lab Optimization: LMS Virtual. Figure 9. It also analyzes design robustness and reliability according to Design for Six Sigma criteria.Lab Optimization.Lab Correlation To guarantee realistic high fidelity simulations. joints and boundary conditions.

. it runs sensitivity analyses that efficiently retain the most influential specified parameters. Using Sol200 sensitivities.2.Systematic Validation from the bottom up. Pre-Test: When preparing measurements for physical structures. this means defining a set of measuring points and excitation points. Several correlation metrics.Lab allows users to easily setup and drive Nastran Sol200 to obtain the sensitivity of FE dynamic properties towards a set of design parameters to decide which parameters to change to obtain better correlation results. For a modal test setup.Lab Correlation offers specialized features to identify specific locations that need improvement. like MAC and FRAC. are available to study (mode) shape or frequency response function correlation interactively. Correlation: Once good test data for the physical model is available. LMS Virtual.Lab Correlation provides tools to quickly carry out this pre-test analysis in a user-friendly way. Systematic validation from the bottom up Figure 10. such as Nastran Solution 200. one can use modal information of preliminary Finite Element models to define the optimal measurement set-up. MAC and frequency difference . LMS Virtual.1. which focuses on tuning modal frequencies and response functions. LMS Virtual. LMS Virtual. To facilitate model updating driven by validations.Lab Correlation allows its users to quantify the geometrical and dynamic (FRF and Modal) resemblance between the test model and its FE equivalent model. Users can also automatically update models using internal and external algorithms. Sensitivity and updating: After the dynamic correlation between two models has been quantified. 2. For example.Modal Assurance Criterion (MAC) 26 Validation-driven model updating: Deducing model improvements using validation output is not always obvious. The objective is to obtain a measurement set-up that guarantees high quality measurement data. Specialized algorithms and post-processing tools allow to localize the problem locations of bad shape correlation and give insight in stiffness differences between the two models.

For this.Lab offers the possibilities to carry out Design of Experiments. LMS Virtual. An orthogonality check between two models adds a degree of correlation accuracy by using the mass matrix to compare system dynamics. The DPR (Driving Point Residue) criterion is used for the excitation point set. Visual Shape correlation for side-by-side model animation (FE or Test). LMS Virtual.Lab Correlation offers tools to ensure that high-quality FE models are used in a CAE environment and that correct sensor and excitation locations are employed in a dynamic physical structure test environment. modes and frequency spectra. Features and Benefits LMS Virtual. Features: Universal access to test and FE data for models.Lab Correlation. 2. the MAC Contribution (MACCo) criterion points out the differences to be examined. If the MAC values are too low to subjectively correlate the modes.2. A broad range of sensitivities can also be computed using LMS Virtual. LMS Virtual. It helps users to quantitatively articulate the degree of shape correlation using a MAC matrix. LMS Virtual. users can verify different modeling assumptions by comparing reference or measurement data. users can create an optimal test geometry from an existing FE model. LMS Virtual.2.Lab Correlation interactively creates a test wireframe on top of the FE mesh and directly quantifies its quality according to relevant mode capturing and modal excitation. The FRAC (Frequency Response Assurance Criterion) compares transfer functions between two models and provides information about global stiffness and mass modeling errors. Users can easily change the test geometry and directly assess new quality levels using the MAC (Modal Assurance Criterion).Lab Correlation also lets users easily and quickly compare the dynamic behavior of two models and deal with incompatible meshes (test and/or FE). Response Surface Modeling and Updating with several local and global optimization algorithms. This improves model and simulation reliability.Lab Correlation provides an easy way to analyze why the model was off-target. In case of poor sensor location set quality. Once the set-up of design parameters (inputs) and correlation metrics (outputs) is in place.Modal Assurance Criterion (MAC) 27 sensitivities are derived to use for modal updating. LMS Virtual. Modal Assurance Criterion (MAC) and MAC Contribution (MACCo) support error localization.Lab Correlation sets up the Nastran DMIG Solution to obtain reduced system mass matrices required for orthogonality checks between test and FE modes. In this way. For pre-test analysis. .Lab Optimization: the user can define a variety of dynamic properties to optimize for a very broad range of design parameters.

Confirm FE simulation model validity using measurements. Export data to LMS Test. Figure 11. Orthogonality check for better dynamic correlation. Improve simulation model reliability. Benefits: Maximum test information with minimized excitation and measurement locations.Lab Correlation uses.Lab integration. Increase measurement productivity with direct LMS Test.. Identify modeling errors or evaluate modeling strategies.LMS Virtual. Driving Point Residue (DPR) for shaker location identification. .Lab or a universal file format.Modal Assurance Criterion (MAC) 28 Frequency Response Assurance Criterion (FRAC).

directly driving the Nastran Guyan reduction. Features: Input design parameters for material and element properties. LMS Virtual. This ensures that the correct FE shape is used in correlation with the reference model during the automated updating process.Lab Desktop.Lab Model Updating can deal adequately with mode switching during the updating process. FRAC (Frequency Response Assurance Criterion) as well as tools to check orthogonality between two models. This can be done by inserting a Nastran Sol200 case from LMS Virtual. Features and Benefits LMS Virtual. FE models are first correlated with reference models.Lab Model Updating. . such as MAC (Modal Assurance Criterion). but can also be FE models. Sensitivity information is then used to update or optimize the Nastran model to match real-life condition better. Response Surface Modeling and several optimization algorithms. LMS Virtual.3. a specific eigenfrequency that is poorly correlated.Lab Optimization. For non-Nastran users.Lab Model Updating. In this way. LMS Virtual. modal frequencies and vibration levels. which are typically test models. Modal Assurance Criterion (MAC). The MAC combined with Nastran Sol200 sensitivities for mode shapes and eigenfrequencies helps users to compute and study MAC and frequency difference sensitivity for mode pair sets. LMS Virtual. users can easily define element group properties for sensitivity analysis.Modal Assurance Criterion (MAC) 29 2. the FE model can still be updated or optimized. using LMS Virtual. Frequency difference sensitivity and MAC sensitivity.Lab Model Updating is a model correlation and updating tool that improves simulation model quality based on reference data. With LMS Virtual. Frequency Response Assurance Criterion (FRAC). These sensitivities help obtain the best dynamic match between two models. The next step is to compute dynamic property sensitivity with respect to design parameter uncertainties.Lab Model Updating easily handles incompatible geometries that typically occur when comparing test and FE models. DOE.2. Targets for mass. including material and property data. Models can be correlated geometrically through alignment. sizing and mapping procedures. Mode Pair Table and MAC Contribution (MACCo). analysts can make models that match reality more closely. Dynamic targets can be the total system mass.Lab Model Updating provides numerical tools. vibration levels for unit load conditions or mode shapes.

..LMS Virtual. Identify modeling errors or evaluate modeling strategies.Modal Assurance Criterion (MAC) 30 Benefits: Confirm FE simulation model validity using measurements.Lab Model Updating Applications. Improve Nastran model reliability with integrated optimization capabilities. Figure 12.

0. After the difficult geometry mapping (geometrical correlation) that aligns both topologies. The cross-orthogonality is used to identify the corresponding test mode that associates with an analytical mode. In aerospace. Uses for aircraft design and testing/certificating companies In the following pages two specific examples. Since the outcome of the cross-orthogonality calculation is also dependent on the quality of the measured test modes. In both cases. including the effects of system mass. the use of the MAC is essential to validate and/or check the tests performed and the predicted FEA results. . based on modal surveys on the prototype. is critical in assessing the value of the analytical model and its interpretation. 3. observes the actual behavior of the structure under controlled laboratory (ground vibration test) or real operating conditions (in-flight testing). and comparing them with the results from FEA. A generally accepted requirement for the crossorthogonality is to have all diagonal terms larger than 0. Furthermore. The analytical approach is predictive and can be used for predicting the flight loads and assessing the structural integrity prior to the prototype production. will be presented: “Using MSC/NASTRAN and LMS/PRETEST to find an optimal sensor placement for modal identification and correlation of aerospace structures” and “Modal Test of L-610G Aeroplane”. The test data is acceptable if the off-diagonal terms of this orthogonality matrix are less than 0.1 when the diagonal terms are normalized to 1. several modal based assessment criteria are used to validate the analytical model. where the modal assurance criterion is used. It is an easy criterion and has been used primarily to check the independence of two modes. analyzing the experimentally obtained mode shapes. The modal assurance criterion (MAC) is used to evaluate the correlation between two modes ignoring the effects of the system mass.9 and all the off-diagonal terms less than 0.1.Modal Assurance Criterion (MAC) 31 3. The experimental approach.1. commonly used tools therefore are the Modal Assurance Criteria (MAC) and the Cross-Orthogonality Criteria. The benefits of such a combined approach are that: Testing provides reliable information to cross-check predicted FEA results (Correlation Analysis): Testing can provide reliable estimates for system damping and resonance frequencies. Using MSC/NASTRAN and LMS/PRETEST to find an optimal sensor placement for modal identification and correlation of aerospace structures The objective of an effective integration of finite element analysis with structural testing is to combine the advantages of both approaches in a more valuable synergistic approach. the orthogonality matrix of the test modes with respect to the analytical reduced mass matrix is used to assess the quality of thereof.

are met.. cross-orthogonality and frequency discrepancy. using MSC/PATRAN and MSC/NASTRAN 2. Moreover. Pretest Analysis using MSC/NASTRAN and LMS/PRETEST 3. This synergistic approach consists thus of the following steps. see Figure 13: 1. and will teach the designer how to improve his design. spring stiffnesses and such. Testing results can be used to enhance the Analytical Model (FEA model Updating): The outcome of the correlation analysis will decide if it is necessary to modify the analytical model so that it better describes the results observed from testing. Correlation Analysis using LMS/Correlation and MSC/NastranForLink 5. Such a structural optimization (updating) can be performed using the MSC/NASTRAN Sol200 capability and thus are the changeable parameters shell thickness. beam cross-sections.Linking test and FE. FEA Modeling and Analysis. the analytical model is said to be test-verified. it will make the geometry mapping of both topologies trivial since the experimental geometry was originally created from the FEA model.Modal Assurance Criterion (MAC) 32 The requirement for modal frequencies of corresponding experimental and analytical modes is to have a discrepancy within 5%. The FEA results can be used to better design the Test (Pretest Analysis): FEA information can complement the Test Engineer‟s expertise in selecting optimal ways of stimulating and measuring the dynamic behavior of the test structure. . An improved analytical model is obtained by changing analytical model parameters such that the discrepancy between test and FE resonance frequencies is minimized. Modal Testing & Analysis using LMS CADA-X Modal 4. If both criteria. This in-depth correlation analysis will provide understanding of the discrepancies between the analytical results and the test results. FEM Model Updating using LMS/Updating and MSC/NASTRAN Sol200 Figure 13.

In addition to that. and their selection is critical for the generation of a validated analytical model.1. Using both MSC/NASTRAN and LMS/PRETEST in combination offers the structural dynamicist an additional surplus because the outcome of most of his MSC/NASTRAN dynamic calculations becomes available for interpretation in nice displays at the same time. Starting from a CAD model. especially since the modal density of launch vehicles and other aerospace structures within the frequency range of interest is usually very high. Out of all these modes. because only some of them will contribute significantly to the critical component responses. a limited set of target modes has to be selected and sensors and shakers have to be placed such that they efficiently capture and excite all of these target modes. LMS/Pretest offers some additional tools. an analytical model is created and the dynamic behavior is calculated in terms of resonance frequencies. see Figure 14. very important step in the pre-test analysis is the selection of the target modes. 3. These important structural modes are called „target‟ modes. Figure 14. mode shapes and system‟s mass and stiffness matrices. MSC/NASTRAN and LMS/PRETEST A typical pre-test analysis will usually consist of different steps. by definition.2.the different steps in a pre-test analysis.Modal Assurance Criterion (MAC) 33 3. Lots of techniques and methodologies have been developed already and are still being developed and most of them are implemented by means of user programming (DMAP) in MSC/NASTRAN. Target Mode Selection A first.1. A test-verified model will. but not necessarily for . have a good correlation between the test target modes and the analytical target modes.. It is however not necessary to „capture‟ all these closely spaced modes during a modal survey test. These critical responses are usually located in the areas of hardware concern.1.

It follows that a poor selection of the target modes could result in an analytical model. These criteria are in other words used to find the important system modes. If the modes are again mass normalized the formula for the kinetic energy fraction becomes: 𝐷𝑖𝑎𝑔 𝜙𝑐 𝐾𝐸𝐹 = 𝐷𝑖𝑎𝑔 𝜙𝑠 𝑇 𝑇 𝑀𝑐 𝜙𝑐 𝑀𝑠 𝜙𝑠 = 𝐷𝑖𝑎𝑔 𝜙𝑐 𝑇 𝑀𝑐 𝜙𝑐 . Generally. but the constraint modes are used instead of the rigid body modes. payload) is over constrained. Another method. the kinetic and/or strain energy fraction of that subsystem is calculated. the formula becomes: 𝑀𝑒𝑓𝑓 = 𝜙𝑑 𝑇 𝑀𝑠 𝜙𝑟𝑏 2 Note that this is the same as the root of the mass orthogonality between the deformation modes and the rigid body modes. These are the rigid body modal effective mass. Besides these. Before any criteria are used to determine the target modes. also other techniques are reported already. Rigid Body Modal and Constraint Modal Effective Mass: The rigid body effective mass associated with each deformation mode represents the amount of system mass participating in that mode. the constraint modal effective mass. This formula makes more sense if the component (e. A typical requirement for the selection of target modes is that modes with a translational effective mass equal to or greater than 2 percent of the total mass are target modes. If the modes are calculated using mass normalization. they are usually able to identify the important system modes but they are less useful for the determination of important local modes. is the use of modal participation factors.g.Modal Assurance Criterion (MAC) 34 the non-target modes. The selection criterion to consider a component mode as target mode is an energy content of 50% of the total system energy. This inspection gives the fundamental insight in the modal behavior of the structure and will also serve to interpret all used target mode selection criteria. Therefore. To include the significant local modes of a subsystem for improving the response prediction. there are four methods or combinations thereof widely used in the aerospace industry. which would not accurately predict the structural responses and member loads. a mode with a large effective mass is usually a significant contributor to the system‟s response. Kinetic/Strain Energy and Kinetic/Strain Energy Fraction: Since the modal effective mass criteria look at the structure‟s dynamic behavior on a global basis. The constraint modal effective mass is similar to the rigid body modal effective mass. The kinetic energy fraction is defined as the amount of kinetic energy in that subsystem relative to that of the whole system. which uses a somewhat different approach. all modes coming from an analytical modal analysis should be described in detail by a „simple‟ visual inspection. These target modes will be added to the target mode set if not yet been selected by the previous criteria. the modal kinetic energy fraction and the modal strain energy fraction.

Figure 15. but also on the frequency characteristics of the excitations.Modal Assurance Criterion (MAC) 35 The kinetic energy fraction for the first deformation mode of a scale model of a Boeing 747 is shows in Figure 15 on top of the geometry.. some relevant modes critical to the payload or component responses may not be selected because none of them takes the excitation into account. The structural integrity depends not only on the structure‟s resonance frequencies.The modal kinetic energy for the first deformation mode for several parts of Boeing 747. see Figure 16.The modal kinetic energy for the first deformation mode for each element separately. a tool that includes the excitation characteristics in the target mode selection process will ensure the completeness of the target mode set. Ultimately. Therefore. The lower plane is the undeformed mode shape and each color represents a different group. Useful in this context are the mode participation factors.. The fuselage. which are calculated during the dynamic solution sequences. Mode Participation Factors: Although the previous methods may identify the most of the target modes. defined as (if mass normalization is used): . the mode shapes and the damping. it is possible to visualize the kinetic energy of each element in model separately. both wings including engines and the tail wings are clearly visible as being the different components. Figure 16.

Figure 17. as can be seen in Figure 17. . The participation factor is frequency dependant and its amplitude is determined by the structure‟s resonance behavior (for w close to ls) and by the excitation spectra (for w far from ls).Modal Assurance Criterion (MAC) 36 𝑃𝐹𝑖 = 𝜙 𝑇 𝐹𝑖 𝜆2 − 𝑤 2 𝑠 The output is related to these participation factors by: 𝑉 = 𝑁 𝑖=1 𝑃𝐹𝑖 𝜙𝑖 Important is that these participation factors are independent of the output. On the x-axis the frequency bandwidth. Plotting the participation factors of all modes for a certain frequency band of interest results in the colormap diagram of Figure 18. It is now easy to investigate if some modes are still being missed in the target mode set.. Figure 18.The first five PF.The PF for each mode (each vertical line is a PF) in a colormap display. for the PF of the first five modes. on the y-axis the amplitude of the PF..

a phenomenon that is called „Spatial Aliasing‟. in order to sufficiently define the spatial resolution of all the target modes. see Figure 19. the choice of the sensor set is also extremely important for the outcome and the interpretation of the dynamic correlation tools. To asses the correlation of the mathematical model predictions which in general do not have dynamic degrees of freedom uniquely one to one with the modal test measurements. Sensor Placement Once the set of target modes has been defined. This set is then eventually modified to obtain a qualitative TAM model. the challenge is to use a minimal number of sensors.3. Figure 19. First a sensor set is searched to meet the observability criterion. a reduction (usually Guyan) to the test-analysis model (TAM) is required.Systematic approach to find the optimal sensor (and shaker) set. .Modal Assurance Criterion (MAC) 37 3. must thus be used. a systematic approach in which the test engineer‟s experience is central. the measurement locations and their corresponding degrees of freedom have to be chosen such that all target modes can be observed by the modal survey test.1. Since the analytical model sizes of complete assemblies are way too big for a manual selection of the sensor locations.. Since this dynamic reduction is done towards the measured degrees of freedom. especially for in-flight testing. This can be extremely though for large space structures where the target modes can be closely spaced. Since it is not practical to instrument the test article in all degrees of freedom corresponding to those of the analytical model. An erroneous or too limited subset of sensor locations will lead to an incomplete geometric definition of the mode shapes.

Figure 21 shows a spread of nodes on the internal LO2 tanks of the X-33. the additional group and the target modes are the only input to this algorithm. shows offdiagonal MAC values below a given threshold. a maximum offdiagonality MAC (MODMAC) can be launched. the results and especially the final amount of sensors still depend on the quality of the selection of initial and additional set of possible measurement points.50 nodes (triax) spread of the whole structure of the X-33: Reusable Launch Vehicle. Different tools may assist the test engineer‟s experience in the selection of those groups. Although MAC and MODMAC calculations are straightforward and powerful.2. The spreading can be performed either on all nodes of the structure.. as is illustrated in Figure 20 where 50 points are spread out over the outer shell of the X-33 Advanced Technology Demonstrator. Master DOF Selection – Geometrical Spread: This tool constructs a group with a userspecified number of nodes that are maximally spread out over the structure. If the initial group of points is not able to discriminate all target modes. usually all three translational degrees of freedom. If a lot of component target modes were selected. To avoid the risk of clustering. Figure 20. the crosscorrelation between the target modes is sufficiently low and the chosen set of measurement point will be able to observe all target modes. The initial group.1 or 0. If the off-diagonal terms of this MAC matrix are smaller than 0. given a set of target modes. This algorithm aims at the completion of the initial subset with extra points/degrees of freedom that are chosen out of an additional subset such that a resulting group of points/degrees of freedom is kept that. is a Modal Assurance Criterion (MAC) calculation. .Modal Assurance Criterion (MAC) 38 Sensor set definition to meet the observability criterion: The methodology used to evaluate the quality of a possible subset of the available analytical nodes and their corresponding degrees of freedom. it is possible to ask for a minimal distance between the chosen locations. it is also necessary to have a sensor distribution on these components.

The summed mode shape is shown together with the two most moving points. creation of groups by clicking points in the geometry.Modal Assurance Criterion (MAC) 39 Figure 21. a user-specified number of nodes with the highest (summed) deformation will be grouped.. An example for a part of a satellite is shown in Figure 22. Mode Shape Summation: This tools calculates the sum of a set or subset of (target) modes and for this set of modes and within the selected nodes (assembly or component).. A character line wireframe gives the position of the tank in the whole model. for instance because there are local component modes amongst the .The summed mode shape (in color and deformation) together with the undeformed mesh. Two points were asked as output.50 Points spread of an internal tank of the X-33. grouping all nodes that correspond with a certain element type… A possible strategy to find an optimal set of measurement points can be starting with a relative small number of a-priori know response locations and launching a MODMAC with as additional group a spread of points over the structure. Group definition tools: Of course also different tools are available such as manual group creation and editing. If the target threshold cannot be reached. Figure 22.

one still has to check if this set of points can be used to obtain a high quality TAM model. a second MODMAC can be launched with an additional group that contains a spread of points only of that component… Sensor set definition to meet also the cross-orthogonality criterion: Once a set of possible measurement locations is found that meets the observability criterion. 𝑋𝑂𝑅 = 𝜙 𝑇 𝑀𝑇𝐴𝑀 𝜙 An example is given in Figure 23.Modal Assurance Criterion (MAC) 40 target modes. The correlation between the original target modes and the reduced TAM modes and the resonance frequency discrepancy can be investigated using a MAC calculation. It is obvious that the current set of measurement points is only valid for the first 7 target modes. as can be seen also in the example of Figure 24. Figure 23. by performing the actual reduction in MSC/NASTRAN. It is also necessary to check if the modes from the TAM model are similar to the original target modes. one can check if the mass distribution by the calculation of the orthogonality between the spatially reduced modes and the Guyan reduced mass matrix. and the off-diagonal terms are smaller than 0.1 Therefore.9.. The target put forward for this orthogonality matrix is that the diagonal terms are larger than 0.Cross-Orthogonality using original modes and Guyan reduced mass matrix. is only valid for the first 13 modes (including the 6 rigid body modes). . Typical for Guyan reduction is that the reduction deteriorates for higher frequencies. although it is possible that more target modes can be observed using this set. the TAM model produced by the chosen set of measurement points in the example. If we suppose Guyan reduction.

or participated in the overall response. This process is more or less trial and error. . If the cross-orthogonality and modal assurance criteria are not met. it is possible to add some extra measurement points to the A-set. The definition of the driving point residue. and the experimental modal model would be hard to identify. As such also the modal participation factors in all possible measurement points can be used. is: 𝐷𝑃𝑅𝑘 𝑖 = 2 𝜙𝑖𝑘 2𝑚𝑘 𝑤𝑘 The degrees of freedom with maximum average DPR over all mode shapes are considered to be the best excitation dofs for the specific set of target modes. DPR‟s are stated to be equivalent to modal participation factors. the corresponding resonance would be difficult to observe in the measurement data. for mode k and node i. The tool that is used in LMS/Pretest for the selection of excitation locations is the calculation of the „driving point residues‟ (DPR‟s).Modal Assurance Criterion (MAC) 41 Figure 24. it is however possible to use the Master DOF selection – Ratio M/K tool..4. Shaker Positioning The third stage in the pre-test analysis is the selection of the exciter locations out of the resulting group of measurement points in order to optimally stimulate all the modes of interest. at the driving point. and are a measure of how much each mode is excited.Correlation between the original target modes and the reduced TAM modes and the resonance frequency discrepancy. and the amplitude and the direction of the red arrow show the best position and direction to place the shaker.1. An example is given for a tail boom problem of a helicopter. If using Guyan reduction. The averaged DPR for all target modes is given. If the structure were to be excited close to a node of a particular mode. 3.

.Averaged DPR for all possible excitation points.A typical tail boom mode together with the undeformed mesh. Figure 26..Modal Assurance Criterion (MAC) 42 Figure 25. .

method used. the nose fuselage was elastically suspended. experimenter's skill. Maximum take-off mass with forty passengers is 15 000 kg. Modal Test of L-610G Aeroplane The L-610G is a high-wing monoplane with a T-shaped tail and a pressurised cabin powered by two General Electric turboprop engines.calibration. One problem of every modal test is evaluating the quality of received natural frequencies. the complex power method and the method of forces in quadrature.2.suspension of the structure tested. Effect of the experiment . generalised masses. damping intensity and distribution. The test was intended to determine the effect of significant structural changes on aeroplane's modal characteristics and to obtain data to tune up an analytical finite-element model and also to carry out detailed investigation into flight control circuits in all failure-present conditions tolerated by the rules. During the test the aeroplane was standing on under-inflated tyres of the main landing gear. which cause disturbing shocks and deteriorate measurement results. number and position of transducers. Some important parts of the test are FRFs from all 280 transducers measured as recorded in different configurations of swept sinusoidal excitation. Relative damping and generalised mass were measured by two techniques. proper use of particular circuits. A total of 19 exciters were used for excitation. time for experiment available. damping and modal vectors. The aeroplane modal parameters were investigated by the method of sinusoidal excitation of isolated normal modes (method of appropriated forces). After measuring each mode the result must be evaluated immediately so as to clear up uncertainties. or repeat the measurement. Factors influencing modal test results fall into three categories: Properties of the structure tested . The initial part of the test consisted in identification of the aeroplane natural frequencies. The rubber bundle stiffness was taken into account in making calculation corrections for additional masses and stiffness of exciters and transducers and for aeroplane suspension. access to vibrating structural parts. The flap plays. the response was measured in 280 points. number. influence of moving parts of exciters. were eliminated by additional masses of 20 kg suspended on the trailing edges by soft rubber bundles. linearity. Technical level of experimental facilities . if any. position and magnitude of exciting forces. The disturbing effect of the factors mentioned above must be minimalised. At the L-610G test several criteria were used for verification of linearity. transducers and suspension.Modal Assurance Criterion (MAC) 43 3.modal density. . Illustration of L-610G model test arrangement is shown in Figure on envelope. The additional mass natural frequency was less than 1 Hz and so it did not have to be considered in the aeroplane total weight.

17 1 100 8.2 100 SHTT1 56..2 -5.8 -3.8 100 11.83 5 3.1 14 -2.69 2 4.3 10.5 -1. Table 3 contains Auto-MAC values calculated for the same modes for all 280 measured points.13 8 5.1 0.5 -1.3 -6.1 100 SHTB2 60.46 17 0.8 -4.4 12.6 6.5 -1.1 -1.4 0.0 -6.7 -1.1 -1.3 3.3 -3.7 -0.5 3.3 100 SHTB1 15.8 0.5 -0.3 21.2 -1.5 12.88 9 -0.9 100 SHTHB1 54.2 -2.8 0.8 -1.0 -4.2 4.4 -0.Orthogonality of L-610G aircraft symmetric mode shapes.0 -4.9 1.0 25.0 -1. .5 4.4 0.1 -2.34 4 6.2 -3.0 25.8 11 -0.6 0.5 100 SEMHB 15.4 -4. In the two matrices there are following modes: Mode SWB1 FVB1 SWB2 SEY SEP f[Hz] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 4.0 -5.7 9.3 -7.9 4.6 -0.5 3.1 18 0.3 13.5 4.2 -2.47 10 2.4 -0.4 0.9 19.6 15 4.0 -2.9 2.1 7.9 3.5 100 SWB3 SWT1 SWB4 SWT2 22.8 -5.0 100 27.9 4.0 16 -4.9 -7.7 -2.1 -4.5 0. Mode shapes were checked by their graphic representation and also by verifying their orthogonality and by Modal Assurance Criteria (MAC).5 -1.4 100 46.1 1.9 5.7 7.2 100 Table 2.5 -2.1 -7. Table 2 shows the matrix of generalised masses verifying orthogonality of the L-610G symmetric modes.9 0.1 9.7 2.5 4.4 5.0 -4.5 -4.4 100 SWHB3 46.81 12 6.9 3.0 0.6 3.5 13.6 -1.2 1.4 -4.0 -2.3 -3.6 1.9 -9.8 2.6 -8.9 -1.3 3.7 -3.0 -0.9 3.8 -2.5 -0.9 100 14.2 -7.9 -3.7 10.4 100 10.1 15.0 -4.41 3 0.2 5.6 -13.4 -8.8 100 SEMVB 6.3 -2.2 6.2 4.8 -5.5 -1.9 25.4 3.5 1.5 -4.0 0.3 5.2 -0.2 1.05 7 -2.8 2.7 -4.3 1.2 0.5 1.4 12.7 7.78 13 -7.3 0.0 0.8 0.8 18.7 100 SWHB1 9.4 -0.Modal Assurance Criterion (MAC) 44 effects of moving parts of the test equipment and quality of natural modes isolation.0 -1.8 14.1 23.9 0.57 6 4.1 100 SWHB2 21.6 100 36.6 12.7 7.

3 0.05 7 19.2 1.5 34.9 0. the elements over 10 per cent have again "horizontal" mode shapes and in addition.81 12 22.0 0.0 100 SWHB3 46.4 4.1 0.3 0. Most non-diagonal elements are very small under 10 per cent.8 100 14.78 13 3.1 100 SWB3 SWT1 SWB4 SWT2 22.41 3 0. It is assumed that this is the reason why these values are rather higher.8 100 SHTHB1 54. mode shapes of engines.3 100 SHTB2 60.2 0.7 0.0 100 SHTT1 56.4 100 SEMVB 6.9 0.2 0.1 0.0 0.0 0.8 3.4 2.0 0.4 2.0 0.2 1.0 0. Most of them are "horizontal" modes with movement in longitudinal direction.1 11.0 0.1 1.9 0.3 100 SWHB1 9.0 15.0 19.0 0.0 0.0 0. With the Auto-MAC values the situation is similar.Auto MAC of L-610G aircraft symmetric mode shapes.8 19.3 1.5 22.0 0.0 9.1 0.69 2 24.1 0.3 0.46 17 0.0 1.0 0.4 0.8 1.0 30.0 16 0.5 100 46.0 0.3 11.0 0.3 2.57 6 8.1 16.1 0.5 2.8 100 SEMHB 15.0 0.6 0.3 5.5 11.1 5.4 0.4 2.3 1.0 0.0 1.6 0.1 18 0.3 0.5 1.4 4.7 0.34 4 13.88 9 0.0 0.2 0.6 2.1 14 0.2 0.6 0.1 0.17 1 100 8.6 4.4 9.5 0.9 2.8 15.1 0.3 1.2 2. MAC depends to some extent on choice of the points (and directions) on the structure whose measured displacements were .0 1.1 2. In this direction there was limited number of measured points on fuselage and engines.3 100 11.6 0.0 0.9 11.6 0.8 11 6.1 4.4 0.2 5. SWB1 symmetric first wing bending SEMVB symmetric engine mounting vertical bending FVB1 first fuselage vertical bending SWHB1 symmetric first wing horizontal bending SWB2 symmetric second wing bending SEY symmetric engines yaw SEP symmetric engines pitch SEMHB symmetric engine mounting horizontal bending SHTB1 symmetric first horizontal tailplane bending SWHB2 symmetric second wing horizontal bending SWB3 symmetric third wing bending SWT1 symmetric first wing torsion SWB4 symmetric forth wing bending SWHB3 symmetric third wing horizontal bending SWT2 symmetric second wing torsion SHTHB1 symmetric first horizontal tailplane horizontal bending SHTT1 symmetric first horizontal tailplane torsion SHTB2 symmetric second horizontal tailplane bending In the two matrices the values are given in per cent.2 0.1 0.2 10.9 0.1 0.1 13.5 0.1 4.4 0.4 6.7 4.1 0.4 6.6 0.2 5.1 16.0 1.0 0.0 2.0 11.4 100 10.0 0.Modal Assurance Criterion (MAC) 45 Mode SWB1 FVB1 SWB2 SEY SEP f[Hz] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 4.5 16.5 12. Only a few values are slightly greater.5 0.47 10 0.0 0.2 3.1 2.13 8 1.4 4.2 1.8 2.2 14.0 0.1 3.7 0.1 0.2 0. But it may be seen that higher values may also be found with some shapes featuring good orthogonality.2 0.2 100 27.3 3.0 0.5 18.0 0.1 0.8 7.2 100 Table 3.0 100 SHTB1 15.83 5 23.6 15 0.7 100 SWHB2 21.1 0.7 20..0 15.1 100 36.6 0.

Modal Assurance Criterion (MAC) 46 incorporated into the calculation. In analysing orthogonality and particularly the Auto-MAC values.Test of L-610G Aeroplane. it was useful to assess the effect on the resulting value of motion components in the co-ordinate axes of particular aeroplane parts. Figure 27.. .

just like the modal assurance criterion. removed from the actual development or unaware of application limitations in subsequent implementations. Simplistic tools like the modal assurance criterion are limited in their meaningful application. Certainly in the next few years. of the original modal assurance criterion. that the interpretation of the calculations can be visualized and that both program communicate directly with each other. . a unique environment becomes available that guides the engineer through the complete process.Modal Assurance Criterion (MAC) 47 Conclusions Over the last twenty years. Even so. from pre-test analysis over correlation to end up eventually at the model updating step. New uses of the modal assurance criterion and new criteria will be developed over the next years as users more fully understand the limitations of the current criteria. It may be clear that the use of the available tools and the user programming capabilities of MSC/NASTRAN form a crucial aspect for the calculation of all described tools. It is clear that users will continue to need more feedback concerning quality assurance information relative to experimental modal parameters and that new techniques. An overview is given of the commonly used techniques to address the target mode selection. the sensor location placement and the positioning of the exciters. it will always be important to recognize the origins and limitations of tools like the modal assurance criterion to avoid misuse of the methodology. the modal assurance criterion has demonstrated how a simple statistical concept can become an extremely useful tool in the field of experimental modal analysis and structural dynamics. this report has tried to explain why a carefully performed pre-test analysis really is necessary to end-up with meaningful modal survey test results. has been remarkable and is most likely due to the overall simplicity of the concept. will continue to be developed with strengths and weaknesses. In regarding to the uses for aircraft design and testing/certificating companies. real or perceived. particularly statistical methods that utilize the redundant information present in the measured data. The use of the modal assurance criterion and the development and use of a significant number of related criteria. The synergy of MSC/NASTRAN and LMS/Pretest gives the engineer the additional benefit that. Good test results are really a sine qua non for the interpretation of all dynamic correlation tools that are used in the analytical model verification and validation. Dissatisfaction often has resulted from the misuse of these tools by users. the increased use of other statistical methods as well as further development of singular value/vector methods are related areas that will generate useful tools in this area. The development of related assurance criteria has been initiated by shortcomings.

E. 1999.: Ajuste y validación de modelos teóricos mediante ensayos de vibración sobre el componente.A. Spain.A.Modal Assurance Criterion (MAC) 48 Bibliography [1] [2] [3] [4] [5] [6] [7] [8] Allemang. V. R. Universidad Pública de Navarra. August 2003.. O. . Braun. 1999. R. Report VZLÚ 3/98. LMS International. LMS International (Belgium). C. J. de Ingeniería Mecánica. M. 2006. and Cuerno-Rejado.: The Modal Assurance Criterion – Twenty Years of Use and Abuse.: Modal Test Correlation and Error Localization for Finite Elements Models of Spacecraft structures.: Primera Modal Analysis Systems and Software/Aircraft Modal Testing. Ingeniería Mecánica Energética y Materiales.. M. Universidad de Chile. Marco-Gómez.Lab Correlation. Chile. Spain. Sound and Vibration. LMS: LMS Virtual. Madsen. 14-21. Dpto. LMS: LMS Virtual.Lab Introduction. Van Langenhove and T. Dpto.S. LMS International. Facultad de Ciencias Físicas y Matemáticas. Julio de 2003. Garbayo. pp.. J. M. and Pintor. V. López-Díez.: Using MSC/NASTRAN and LMS/PRETEST to find an optimal sensor placement for modal identification and correlation of aerospace structures. C. Brughmans.: Análisis Modal Experimental de una Estructura Aeronáutica. Cerný. N. Belgium. División Espacio. and Meruane. J.

Sign up to vote on this title

UsefulNot useful- Lecture Notes 897by Azza Sundus Antartika
- Jacquet-Richardet Et Al-1997-Communications in Numerical Methods in Engineeringby Victor Hugo Borges
- Mariano Giaquinta, Giuseppe Modica-Mathematical Analysis _ Linear and Metric Structures and Continuity-Birkhauser (2007)by greenblueleaf
- Melody Secrets 2by Jean Pierre Vz

- Theoretical and Experimental Modal Analysis
- D. J. Ewins - Modal Testing
- ma1201_kapittel1
- 021007CommsI
- r Manul Port
- Problem Set 2
- Mnemonics Matrix
- syllabus_MAT1339N
- Brief MatLab Commands
- me471s03_lec04
- Top 8 Physics Homework Mistakes People Make
- CalcIII_ThreeDimensions
- me471_hwk05ans
- Data Consistency Check
- capitulo 15
- Chapter 12
- MIT18 06S10 Final Exam
- EM Tute_1
- Igor Evgenyevich Irodov-Fundamental Laws of Mechanics -Mir Publishers Moscow (1980).pdf
- Statics-Unit-05.pdf
- Electrodynamics
- Schaum Continuum Mechanics Mase
- Introduction
- Lecture 1 Nd 2 Review of the Three Laws of Motion and Vector Algebra
- Lecture Notes 897
- Jacquet-Richardet Et Al-1997-Communications in Numerical Methods in Engineering
- Mariano Giaquinta, Giuseppe Modica-Mathematical Analysis _ Linear and Metric Structures and Continuity-Birkhauser (2007)
- Melody Secrets 2
- Voicing Modes eBook SAMPLE
- Face Component Extraction Using Segmentation
- Modal Assurance Criterion - Álvaro Rivero

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue reading from where you left off, or restart the preview.