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

Welcome to Scribd! Start your free trial and access books, documents and more.Find out more

“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 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. mode r.Modal Assurance Criterion (MAC) 5 1. The orthogonality check. degree-of-freedom q. State of the art. which can be very powerful when used correctly but very misleading when used incorrectly. = Modal vector for reference p. Antecedents. coupled errors in the analytical model development. was the orthogonality check. 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. The standard of the late 1970s. = Number of outputs (assumed to be larger than Ni). typical abuses will be identified. Before starting with the review of the methods. = Modal coefficient for reference p. not always the best approach. = Transpose of {φ}. The modal assurance criterion is a statistical indicator. Typical uses of the modal assurance criterion will be discussed and finally. This first section will first review the historical development of the modal assurance criterion. when the modal assurance criterion was developed. = Modal coefficient for degree-of-freedom q. = Modal vector for mode r. just like ordinary coherence. the reduction of the analytical model and the estimated modal vectors into a single indicator and was. mode r.1. = Synthesized frequency response function. = Measured frequency response function. linear regression correlation coefficient. mode r. 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. = Complex conjugate of A. . therefore. Other similar assurance criteria will then be identified although the list is not intended to be comprehensive. = Number of analytical modal vectors. however. 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. an analytical model was not available which renders the orthogonality check impractical. = Number of inputs. = Complex conjugate transpose (Hermitian) of {φ}. 1. Many times. = Number of experimental modal vectors.

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. each modal vector of a system will be orthogonal to all other modal vectors of that system when weighted by the mass. In the pseudo-orthogonality check.1. the off-diagonal values in the modal mass matrix are expected to be less than 0. a big modal assurance criterion (Big MAC) and a multiple modal assurance criterion (MultiMAC) were formulated as part of the original development. to evaluate orthogonality between the experimental and analytical modal vectors. In the original thought process. The experimental and analytical modal vectors are scaled so that the diagonal terms of the modal mass matrix are unity. For this reason. MAC is the designation that has persisted. 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). for the case of proportional damping. Theoretically. normally derived from the same finite element model.1 (10 percent of the diagonal terms). Since the “Big Mac” acronym was already in use at that time. normally derived from a finite element model. normally derived from a finite element model. Variations of this process include using analytical modal vectors together with experimental modal vectors and the appropriately sized mass or stiffness matrix. With this form of scaling. a little modal assurance criterion (Little MAC). Little MAC and Multi-MAC were primarily testing methods and are not discussed further here. stiffness or damping matrix. The modal assurance criterion that survives today is what was originally identified as Big MAC. This latter comparison is normally referred to as a pseudo-orthogonality check (POC). the experimental modal vectors are used together with a mass matrix. In the traditional orthogonality check. 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. In practice. to evaluate orthogonality of the experimental modal vectors. any further discussion of . these matrices are made available by way of a finite element analysis and normally the mass matrix is considered to be the most accurate. 1.1. In the internal development of the modal assurance criterion at the University of Cincinnati. Structural Dynamics Research Lab (UCSDRL).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. and the analytical modal vectors. the experimental modal vectors are used together with a mass matrix.

all three situations contribute to the failure of orthogonality or pseudoorthogonality criteria on occasion. part of the problem. . When the orthogonality conditions are not satisfied. experimental modal vectors are not always real-valued and Eqs. the reduction of the mass matrix can be invalid. As a result. in general. In such a situation. 1 is not sufficiently satisfied. such as the center of an unsupported panel.Modal Assurance Criterion (MAC) 7 orthogonality will be made with respect to mass matrix weighting. 1 and 2 are developed based upon normal or real-valued modal vectors. the result of zero for the cross orthogonality calculations (r ≠ s. but may represent only incidental motion of the overall modal vector. Note that. Third. This normally involves some sort of reduction algorithm as well as interpolation of grid locations to match the measurement situation. 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. First. 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. The accepted criterion in the aerospace industry. This can certainly be a realistic problem and cause severe errors. Since the mass matrix does not always represent the actual physical properties of the system when it is built or assembled. 1 and 2. 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. the modal vectors can be invalid. even under this criterion. This complication has to be resolved by a process of real normalization of the measured modal vectors prior to utilizing Eqs. it is important to try to develop methods that indicate confidence that the modal vector is. 2. 1 and 2 or by applying an equivalent procedure involving the state-space form of the weighted orthogonality relationship. In Eqs. the mass matrix must be an No×No matrix corresponding to the measurement locations on the structure. an attempt is made to adjust the modal vectors so that the cross orthogonality conditions are satisfied. 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. one (or more) of three situations may exist. This can be due to measurement errors or problems with the modal parameter estimation algorithms. where this confidence check is made most often. This is a very common assumption and many times contributes to the problem. Often. or is not. 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. From an experimental point of view. is for all of the generalized mass terms to be unity and all cross orthogonality terms to be less than 0. In all probability. the measurement location is weighted very heavily in the orthogonality calculation of Eq. the mass matrix can be invalid. When Eq. Second.1. Eq. this result does not indicate where the problem originates.

Once two different modal vector estimates are scaled similarly. In terms of modern. 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). 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 these cases. taking into account magnitude and phase differences. methods can be compared and contrasted in order to evaluate the mutual consistency of different procedures rather than estimating the modal vectors specifically. If an analytical and an experimental vector are deemed consistent or similar. the modal scale factor is a normalized estimate of the modal participation factor between two references for a specific mode of vibration. 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 will contain potential variance errors. the estimation of the modal vector will contain potential bias and variance errors. Frequently. If the modal vectors are not well represented in the frequency response function matrix. 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. In this approach. 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. In any case. different subsets of the frequency response function matrix and/or different modal parameter estimation algorithms are utilized to estimate separate.Modal Assurance Criterion (MAC) 8 1. multiple reference modal parameter estimation algorithms. together with the modal scale factor. can be used to complete the experimental modal vector if some degrees of freedom could not be measured. The modal scale factor and the modal assurance criterion also provide a method of easily comparing estimates of modal vectors originating from different sources. . 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. redundant modal vectors for comparison purposes. The function of the modal scale factor (MSF) is to provide a means of normalizing all estimates of the same modal vector. This provides an additional confidence factor in the evaluation of a modal vector from different excitation (reference) locations or different modal parameter estimation algorithms.2. the analytical modal vector.1. 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. if different estimates of the same modal vectors are generated.

according to this approach. modal vector c can be considered to be made of two parts. The first part is the part correlated with modal vector d.Modal Assurance Criterion (MAC) 9 The modal scale factor is defined. 3 implies that the modal vector d is the reference to which the modal vector c is compared. In the general case. 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 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. 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 error vector is considered to be noise. 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. this is also equivalent to: {𝜑 }𝐻 {𝜑 ∗ } 𝑑𝑟 𝑐𝑟 𝑀𝑆𝐹𝑐𝑑𝑟 = {𝜑 𝑑𝑟 }𝐻 {𝜑 ∗ } 𝑑𝑟 (3c) Eq.

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

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 first four reasons can be eliminated. It could simply be a random noise vector or a vector reflecting the bias in the modal parameter estimation algorithm. Hopefully. if the first three reasons can be eliminated. This does not necessarily mean that they are correct. this is an indication that the modal vectors are consistent.4. a rotating piece of equipment with an unbalance is present in the system being tested. This situation can occur whenever too few response stations have been included in the experimental determination of the modal vector. during the measurement of the frequency response function. the two modal vectors should differ only by the complex valued scale factor. difference or sort the modal vectors. The modal vectors are the result of a forced excitation other than the desired input. Since the reference modal vector may be arbitrarily chosen. 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. this measure of inconsistency will imply that the modal vectors are orthogonal. This process results in a Ne×Na rectangular modal assurance criterion matrix with values that . this modal vector may not be one of the true modal vectors of the system. The modal vectors can be consistent for any of the following reasons: The modal vectors have been incompletely measured. the modal assurance criterion can be applied in many different ways. The modal vectors represent the same modal vector with different arbitrary scaling. Modal Assurance Criterion (MAC) Unity If the modal assurance criterion has a value near unity. In any case. If the two modal vectors being compared have the same expected value when normalized. the modal assurance criterion will only reflect a consistent (linear) relationship to the reference modal vector. The modal assurance criterion can be used to verify or correlate an experimental modal vector with respect to a theoretical modal vector (eigenvector). Under the constraints mentioned previously. The modal vectors are primarily coherent noise. 1.1. which is a function of the common modal coefficients between the rows or columns. since the different modal vector estimates are from different excitation positions.Modal Assurance Criterion (MAC) 11 The modal vectors are from linearly unrelated mode shape vectors. This would be the situation if. Therefore. the modal assurance criterion can be interpreted in a similar way as an orthogonality calculation.

5. Additionally. into a complete estimate of each modal vector at all measurement stations. 1. If the error appears to be biased or skewed. These characteristics. 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. 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. Historically. 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. provide a useful tool in the processing of experimental modal vectors. as well as others. Since the modal scale factor is a complex scalar that allows two vectors to be phased the same and to the same mean value. Since the modal scale factor and modal assurance criterion are computed analogous to the frequency response function and coherence function. differenced or sorted to determine the best single estimate or the potential source of contamination using the modal scale factor. agreement between existing methods can be established and new modal parameter estimation methods can be checked for characteristics that are consistent with accepted procedures. 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. 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. assuming the modal assurance function indicates consistency. Once the modal assurance criterion establishes that two vectors represent the same information. modal vectors can be sorted. In this way. both the advantages and limitations of the computation procedure are well understood.Modal Assurance Criterion (MAC) 12 approach unity whenever an experimental modal vector and an analytical modal vector are consistently related. the vectors can be averaged. .1. these vectors can be subtracted to evaluate whether the error is random or biased. a table of numbers was usually presented as shown in Table 1. the modal vectors can be averaged (using the modal scale factor) to improve the estimate of a modal vector. Based upon partial but overlapping measurement of two columns of the frequency response function matrix. If the error appears to be random and the modal assurance criterion is high. 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.

It is important to remember. however..Modal Assurance Criterion (MAC) 13 Table 1.2-D and 3-D presentation of MAC values. Today. . 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. that MAC is a discrete calculation and what appears as a color contour plot really only represents the discrete mode to mode comparison.Numerical presentation of MAC values. a color plot does allow for more data to be presented in an understandable form in a minimum space.. Nevertheless. Figure 1.

just sensor distribution. 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. 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. This is required since the modal assurance criterion minimizes the squared error and is dominated by the larger values. . Y and/or Z) or only the degrees-offreedom from a component of the complete modal vector. 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.2.Modal Assurance Criterion (MAC) 14 1. 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. In this case. The purpose of the weighting matrix is to recognize that MAC is not sensitive to mass or stiffness distribution. WMAC is proposed for these cases. Essentially this approach utilizes the square root of the MAC calculation. The subset is chosen based upon the user‟s interest and may reflect only a certain dominant sensor direction (X. Other Similar Assurance Criteria The following brief discussion highlights assurance criteria that utilize the same linear. least squares computation approach to the analysis (projection) of two vector spaces as the modal assurance criterion. 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. 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. which tends to highlight the cross terms (off diagonal) that are generally very small MAC values. The equations for each assurance criterion are not repeated unless there is a significant computational difference that needs to be clarified or highlighted. 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. and to adjust the modal assurance criterion to weight the degrees-of-freedom in the modal vectors accordingly. 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.

but the set of mode pairs represents all modes of interest in a given frequency range. assessment of parameter variation. therefore.Modal Assurance Criterion (MAC) 15 Modal Assurance Criterion Using Reciprocal Vectors (MACRV): A reciprocal modal vector is defined as the mathematical vector that. 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. etc. there will be a value of COMAC computed for each (measurement) degree-of-freedom. 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 . 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. 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.” 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.). 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. analytical versus analytical. Only those modes that match between the two sets are included in the computation. the result is zero. experimentally based. The COMAC is calculated over a set of mode pairs. This extension of MAC “offers a means of displaying simultaneously the mode shape correlation. 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). Coordinate Modal Assurance Criterion (COMAC): An extension of the modal assurance criterion is the coordinate modal assurance criterion (COMAC). when transposed and premultiplied by a specific modal vector. experimental versus experimental or experimental versus analytical. The COMAC attempts to identify which measurement degrees-of-freedom contribute negatively to a low value of MAC. For two sets of modes that are to be compared. 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. yields unity. 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. the degree of spatial aliasing and the frequency comparison in a single plot. When the same computation is performed with this reciprocal modal vector and any other modal vector or any other reciprocal modal vector.

The equation in this formulation utilizes a transpose and will only correctly apply to real valued vectors. 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. 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. 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). 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. 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. etc. 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 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 approach uses essentially the same computational scheme as MAC but utilizes the inverse of the modal coefficients. in the vector comparisons. Therefore. The modal correlation coefficient (MCC) is a modification of MAC that attempts to provide a more sensitive indicator.). this computation suffers from the possibility that a modal coefficient could be numerically zero. 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). 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.Modal Assurance Criterion (MAC) 16 defining modal vectors such as sensor scaling mistakes and sensor orientation (plus or minus sign) errors. Naturally. small modal coefficients become significant in the least squares based correlation coefficient computation. The simplest example is a validation procedure that compares the FRF data synthesized from the modal model with the measured FRF data. synthesis correlation coefficient and response vector assurance criterion (RVAC)). pressure. Naturally. This serves as an independent check of the modal parameter estimation process. position by position. intensity. This approach has been utilized in the modal parameter estimation process for a number of years under various designations (parameter estimation correlation coefficient.

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

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

users implement the modal assurance criterion. these wild points may dominate the MAC calculation. 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. Modal vector stability or consistency is identified using a MAC computation where the vectors include only the degrees-of-freedom at the reference locations. particularly if the modal vector is not well excited from one or more of the reference locations. However. if the modal vectors have only a limited number of degrees-offreedom. Vectors with many elements reduce the sensitivity of MAC to this problem. modal parameter estimation mistakes. if there have been erroneous data included in the modal vectors due to calibration errors. This also means that. However. using a vector transpose in the numerator and denominator calculations rather than an Hermitian (conjugate transpose). typically two to five. multiple reference modal parameter estimation algorithms estimate the stability or consistency diagram. In these situations. This innocent error often occurs when the author is utilizing real-valued vectors and notices no problem. This error causes no problem as long as analytical vectors or real-valued experimental vectors are involved in the calculation. this will skew the meaning of the numerical MAC value. . users who do not recognize this issue are often led astray in subsequent applications involving complex-valued vectors. several modal vectors at different modal frequencies will have very high MAC numbers indicating that the modal vectors are the same.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. 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). etc.. This frequently happens when high order. or a related similar computation. Therefore. 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. The original mathematical formulation assumes the general case but has been reported incorrectly in some literature. 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 does not weight the modal vectors with a mass or stiffness matrix and. this does not give the correct result. where some of the vectors are complex-valued. 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. The typical example involves the engine of an automobile. 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. The wrong mathematical formulation for the modal assurance criterion is used: Frequently. 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. there may be great variability in the MAC computation. in the general case.

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

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

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

Modal Assurance Criterion (MAC) 23 LMS Virtual.Lab Noise and Vibration is developed to efficiently analyze. engine acoustics.Lab Acoustics.Lab Acoustics simulates both internal and external acoustic radiation and offers dedicated applications for structural noise radiation.. 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 Noise and Vibrations: LMS Virtual.LMS Virtual. aero-acoustic phenomena and much more.Lab Acoustics: LMS Virtual. transmission loss through panels. Figure 6. LMS Virtual. refine and optimize the vibro-acoustic behavior of a design. Convenient tools enable engineers to quickly perform design modifications and assess the noise and vibration performance of a design variant in minutes. . It offers all the required tools to create system-level models. 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.Lab Acoustics offers an integrated solution to minimize radiated noise or optimize the sound quality in new designs.. LMS Virtual. build realistic load cases and simulate noise and vibration responses. Figure 5.Lab Noise and Vibrations.LMS Virtual.

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

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

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

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

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

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

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

the use of the MAC is essential to validate and/or check the tests performed and the predicted FEA results. Furthermore.1. In both cases. and comparing them with the results from FEA. 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”. observes the actual behavior of the structure under controlled laboratory (ground vibration test) or real operating conditions (in-flight testing). the orthogonality matrix of the test modes with respect to the analytical reduced mass matrix is used to assess the quality of thereof. analyzing the experimentally obtained mode shapes. It is an easy criterion and has been used primarily to check the independence of two modes. several modal based assessment criteria are used to validate the analytical model. based on modal surveys on the prototype. commonly used tools therefore are the Modal Assurance Criteria (MAC) and the Cross-Orthogonality Criteria. The test data is acceptable if the off-diagonal terms of this orthogonality matrix are less than 0. The modal assurance criterion (MAC) is used to evaluate the correlation between two modes ignoring the effects of the system mass. A generally accepted requirement for the crossorthogonality is to have all diagonal terms larger than 0. The cross-orthogonality is used to identify the corresponding test mode that associates with an analytical mode. . The experimental approach. Uses for aircraft design and testing/certificating companies In the following pages two specific examples. is critical in assessing the value of the analytical model and its interpretation. 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.0. After the difficult geometry mapping (geometrical correlation) that aligns both topologies.9 and all the off-diagonal terms less than 0. including the effects of system mass. In aerospace.Modal Assurance Criterion (MAC) 31 3. where the modal assurance criterion is used. 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. Since the outcome of the cross-orthogonality calculation is also dependent on the quality of the measured test modes.1 when the diagonal terms are normalized to 1.1. The analytical approach is predictive and can be used for predicting the flight loads and assessing the structural integrity prior to the prototype production. 3.

This in-depth correlation analysis will provide understanding of the discrepancies between the analytical results and the test results. beam cross-sections. are met. . FEM Model Updating using LMS/Updating and MSC/NASTRAN Sol200 Figure 13. cross-orthogonality and frequency discrepancy. An improved analytical model is obtained by changing analytical model parameters such that the discrepancy between test and FE resonance frequencies is minimized. spring stiffnesses and such.Modal Assurance Criterion (MAC) 32 The requirement for modal frequencies of corresponding experimental and analytical modes is to have a discrepancy within 5%. and will teach the designer how to improve his design. 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. FEA Modeling and Analysis. 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. 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. see Figure 13: 1. Pretest Analysis using MSC/NASTRAN and LMS/PRETEST 3. Such a structural optimization (updating) can be performed using the MSC/NASTRAN Sol200 capability and thus are the changeable parameters shell thickness. Moreover. using MSC/PATRAN and MSC/NASTRAN 2. If both criteria.. This synergistic approach consists thus of the following steps.Linking test and FE. Modal Testing & Analysis using LMS CADA-X Modal 4.

2.1. MSC/NASTRAN and LMS/PRETEST A typical pre-test analysis will usually consist of different steps. 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. A test-verified model will. 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. Figure 14.. very important step in the pre-test analysis is the selection of the target modes. Target Mode Selection A first. 3. an analytical model is created and the dynamic behavior is calculated in terms of resonance frequencies.the different steps in a pre-test analysis. Out of all these modes. It is however not necessary to „capture‟ all these closely spaced modes during a modal survey test. by definition. Starting from a CAD model. see Figure 14. but not necessarily for . LMS/Pretest offers some additional tools. and their selection is critical for the generation of a validated analytical model. especially since the modal density of launch vehicles and other aerospace structures within the frequency range of interest is usually very high.1.Modal Assurance Criterion (MAC) 33 3.1. have a good correlation between the test target modes and the analytical target modes. mode shapes and system‟s mass and stiffness matrices. These critical responses are usually located in the areas of hardware concern. In addition to that. These important structural modes are called „target‟ modes. 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.

which would not accurately predict the structural responses and member loads.Modal Assurance Criterion (MAC) 34 the non-target modes. there are four methods or combinations thereof widely used in the aerospace industry. Therefore. These are the rigid body modal effective mass. To include the significant local modes of a subsystem for improving the response prediction. If the modes are calculated using mass normalization. the kinetic and/or strain energy fraction of that subsystem is calculated. all modes coming from an analytical modal analysis should be described in detail by a „simple‟ visual inspection. Before any criteria are used to determine the target modes. a mode with a large effective mass is usually a significant contributor to the system‟s response. Another method. The selection criterion to consider a component mode as target mode is an energy content of 50% of the total system energy. These criteria are in other words used to find the important system modes. the constraint modal effective mass. which uses a somewhat different approach. 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. 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. The constraint modal effective mass is similar to the rigid body modal effective mass. 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. is the use of modal participation factors.g. also other techniques are reported already. Generally. payload) is over constrained. The kinetic energy fraction is defined as the amount of kinetic energy in that subsystem relative to that of the whole system. the modal kinetic energy fraction and the modal strain energy fraction. but the constraint modes are used instead of the rigid body modes. This formula makes more sense if the component (e. Besides these. 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. These target modes will be added to the target mode set if not yet been selected by the previous criteria. 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. they are usually able to identify the important system modes but they are less useful for the determination of important local 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 . It follows that a poor selection of the target modes could result in an analytical model.

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

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.The PF for each mode (each vertical line is a PF) in a colormap display.. Figure 17.The first five PF. as can be seen in Figure 17. on the y-axis the amplitude of the PF. Plotting the participation factors of all modes for a certain frequency band of interest results in the colormap diagram of Figure 18. for the PF of the first five modes. Figure 18. It is now easy to investigate if some modes are still being missed in the target mode set.. 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). . On the x-axis the frequency bandwidth.

a reduction (usually Guyan) to the test-analysis model (TAM) is required. a systematic approach in which the test engineer‟s experience is central. This can be extremely though for large space structures where the target modes can be closely spaced. in order to sufficiently define the spatial resolution of all the target modes. the choice of the sensor set is also extremely important for the outcome and the interpretation of the dynamic correlation tools. see Figure 19. a phenomenon that is called „Spatial Aliasing‟.Modal Assurance Criterion (MAC) 37 3. 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. 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. especially for in-flight testing. 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. This set is then eventually modified to obtain a qualitative TAM model.1. Sensor Placement Once the set of target modes has been defined. must thus be used. the challenge is to use a minimal number of sensors.. First a sensor set is searched to meet the observability criterion.Systematic approach to find the optimal sensor (and shaker) set. . 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. Figure 19.3.

the additional group and the target modes are the only input to this algorithm. it is also necessary to have a sensor distribution on these components. the crosscorrelation between the target modes is sufficiently low and the chosen set of measurement point will be able to observe all target modes.. is a Modal Assurance Criterion (MAC) calculation. If the initial group of points is not able to discriminate all target modes. as is illustrated in Figure 20 where 50 points are spread out over the outer shell of the X-33 Advanced Technology Demonstrator. To avoid the risk of clustering. . given a set of target modes. Figure 21 shows a spread of nodes on the internal LO2 tanks of the X-33. it is possible to ask for a minimal distance between the chosen locations.2. a maximum offdiagonality MAC (MODMAC) can be launched. If the off-diagonal terms of this MAC matrix are smaller than 0. Although MAC and MODMAC calculations are straightforward and powerful. shows offdiagonal MAC values below a given threshold. Figure 20. 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. usually all three translational degrees of freedom.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. The spreading can be performed either on all nodes of the structure. The initial group. 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 a lot of component target modes were selected. 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.50 nodes (triax) spread of the whole structure of the X-33: Reusable Launch Vehicle.1 or 0. Different tools may assist the test engineer‟s experience in the selection of those groups.

creation of groups by clicking points in the geometry. Two points were asked as output. An example for a part of a satellite is shown in Figure 22.Modal Assurance Criterion (MAC) 39 Figure 21. The summed mode shape is shown together with the two most moving points. 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 user-specified number of nodes with the highest (summed) deformation will be grouped..50 Points spread of an internal tank of the X-33. If the target threshold cannot be reached. for instance because there are local component modes amongst the . Group definition tools: Of course also different tools are available such as manual group creation and editing. A character line wireframe gives the position of the tank in the whole model.The summed mode shape (in color and deformation) together with the undeformed mesh. 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.

and the off-diagonal terms are smaller than 0. . It is obvious that the current set of measurement points is only valid for the first 7 target modes.Modal Assurance Criterion (MAC) 40 target modes. Figure 23. 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. The target put forward for this orthogonality matrix is that the diagonal terms are larger than 0.. although it is possible that more target modes can be observed using this set. as can be seen also in the example of Figure 24. by performing the actual reduction in MSC/NASTRAN. If we suppose Guyan reduction. one still has to check if this set of points can be used to obtain a high quality TAM model. It is also necessary to check if the modes from the TAM model are similar to the original target modes. is only valid for the first 13 modes (including the 6 rigid body modes). the TAM model produced by the chosen set of measurement points in the example.Cross-Orthogonality using original modes and Guyan reduced mass matrix. The correlation between the original target modes and the reduced TAM modes and the resonance frequency discrepancy can be investigated using a MAC calculation.9. one can check if the mass distribution by the calculation of the orthogonality between the spatially reduced modes and the Guyan reduced mass matrix.1 Therefore. Typical for Guyan reduction is that the reduction deteriorates for higher frequencies. 𝑋𝑂𝑅 = 𝜙 𝑇 𝑀𝑇𝐴𝑀 𝜙 An example is given in Figure 23.

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

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

the response was measured in 280 points. Technical level of experimental facilities . access to vibrating structural parts. if any. A total of 19 exciters were used for excitation. time for experiment available. 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. were eliminated by additional masses of 20 kg suspended on the trailing edges by soft rubber bundles.2. Maximum take-off mass with forty passengers is 15 000 kg. The additional mass natural frequency was less than 1 Hz and so it did not have to be considered in the aeroplane total weight. proper use of particular circuits. transducers and suspension. experimenter's skill. 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. One problem of every modal test is evaluating the quality of received natural frequencies. which cause disturbing shocks and deteriorate measurement results. method used. The flap plays. The initial part of the test consisted in identification of the aeroplane natural frequencies. Relative damping and generalised mass were measured by two techniques. 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.modal density. At the L-610G test several criteria were used for verification of linearity. The aeroplane modal parameters were investigated by the method of sinusoidal excitation of isolated normal modes (method of appropriated forces). influence of moving parts of exciters. position and magnitude of exciting forces. Factors influencing modal test results fall into three categories: Properties of the structure tested . linearity. . 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. Effect of the experiment . the nose fuselage was elastically suspended. The disturbing effect of the factors mentioned above must be minimalised. the complex power method and the method of forces in quadrature. After measuring each mode the result must be evaluated immediately so as to clear up uncertainties.Modal Assurance Criterion (MAC) 43 3. Illustration of L-610G model test arrangement is shown in Figure on envelope. number.suspension of the structure tested. damping and modal vectors.calibration. During the test the aeroplane was standing on under-inflated tyres of the main landing gear. generalised masses. damping intensity and distribution. or repeat the measurement.

5 3.5 3.9 0.5 -4.6 -13.5 4.1 -7.7 9.0 0.4 -0.3 -6.2 1.0 100 27.83 5 3.4 100 46.1 -4.9 4.2 -2.4 -4.0 -4.5 -0.8 100 11.7 -4.9 4.0 -2.4 0.34 4 6.8 -5.4 12.3 13.2 -1.1 -1.1 9.4 0.7 -3.7 7.7 10.81 12 6.1 100 SHTB2 60. Table 3 contains Auto-MAC values calculated for the same modes for all 280 measured points.78 13 -7.2 -7.4 100 10.4 -0.1 1.9 3.57 6 4..47 10 2.2 1.4 0.5 0.4 -0.3 3.9 19.8 -4.5 -1.6 1.6 -1.8 -2.3 -3. 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 4.2 -0.8 11 -0. Table 2 shows the matrix of generalised masses verifying orthogonality of the L-610G symmetric modes.4 100 SWHB3 46.0 0.9 2.0 -1.3 5.2 100 SHTT1 56.5 -1.8 100 SEMVB 6.5 12.88 9 -0.3 100 SHTB1 15.2 0.0 25.7 7.6 -8. .0 -4.5 4.8 2.2 100 Table 2.3 1.6 12.8 14.3 -7.4 3.3 21.9 -1.6 0.3 -3.7 -1.9 3.2 -5.5 13.0 -2.8 0.3 3.1 15.05 7 -2.41 3 0.5 -4.1 18 0.8 18.5 -1.9 25.3 10.0 -4.5 -2.8 0.9 -9.7 7.6 6.6 -0.4 5.4 -8.5 100 SEMHB 15.8 0.5 1.0 16 -4.2 6.9 3.0 -6.9 1.17 1 100 8.2 -2.5 1.1 -1.Modal Assurance Criterion (MAC) 44 effects of moving parts of the test equipment and quality of natural modes isolation.69 2 4.46 17 0.8 -5.9 5.5 -1.9 -3.1 14 -2.0 -5.0 25.1 0.6 100 36.2 4.8 -1. Mode shapes were checked by their graphic representation and also by verifying their orthogonality and by Modal Assurance Criteria (MAC).1 23.7 2.3 0.8 -3.6 3.4 12.9 -7.4 -4.9 0.8 2.3 -2.1 7.2 4.0 -0.7 -0.2 5.5 -1.5 -0.2 -3.6 15 4.5 100 SWB3 SWT1 SWB4 SWT2 22.Orthogonality of L-610G aircraft symmetric mode shapes.7 -2.9 100 14.7 100 SWHB1 9.9 100 SHTHB1 54.0 -4.1 100 SWHB2 21.0 -1.1 -2.0 0.13 8 5.5 4.

6 0.0 0.5 0.3 5.0 0.0 16 0. But it may be seen that higher values may also be found with some shapes featuring good orthogonality.4 0.4 9.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 4.4 0.0 1.0 0.34 4 13.1 0. the elements over 10 per cent have again "horizontal" mode shapes and in addition.1 0.1 11.0 0.1 100 SWB3 SWT1 SWB4 SWT2 22.5 34.6 4.3 100 11.2 0.3 0.5 22.8 19.1 2.0 0.83 5 23.9 0. mode shapes of engines.1 0.8 100 SHTHB1 54.2 0.0 0.3 0.9 2. Most of them are "horizontal" modes with movement in longitudinal direction.1 0.2 100 Table 3.2 0.8 11 6.0 0.2 0.1 16..0 1.0 2.7 0.7 0.3 2.1 3.78 13 3. 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.9 0.1 0.2 2.47 10 0.69 2 24.1 100 36.81 12 22.0 0.1 0.5 16.1 0.6 0.8 100 14.6 0. In this direction there was limited number of measured points on fuselage and engines.2 14.3 1.1 0.9 11.1 4.0 0.57 6 8.0 100 SHTT1 56.6 0.0 0.0 0.4 2.05 7 19.1 0.0 0.0 30.3 0.4 100 SEMVB 6.0 0.1 0.0 0.0 0.3 0.9 0.0 0.3 100 SWHB1 9.4 4. With the Auto-MAC values the situation is similar.88 9 0.2 1. MAC depends to some extent on choice of the points (and directions) on the structure whose measured displacements were .0 11.8 100 SEMHB 15.3 100 SHTB2 60.2 10.1 2.5 11.0 100 SWHB3 46.4 6.6 15 0. It is assumed that this is the reason why these values are rather higher.5 0.8 7.4 0.2 0.17 1 100 8.8 3.5 100 46.0 19.0 0.2 5.1 5.1 4.9 0.4 0.5 1.2 100 27.41 3 0.2 0.3 1.2 5.4 6.4 100 10.1 14 0.6 0.4 2.1 0.13 8 1.0 0.5 2.7 4.0 0.0 1.4 2.8 1.6 0.1 1.0 9.4 4.8 2.Auto MAC of L-610G aircraft symmetric mode shapes.0 15. Most non-diagonal elements are very small under 10 per cent.2 3.0 1.5 0.7 0.1 0.3 11.0 15.46 17 0.6 2.6 0.1 13.1 0.4 4.0 100 SHTB1 15.2 1.0 0.0 0.2 0.2 1.3 3.7 100 SWHB2 21.5 18.0 0.3 1.1 16.0 0.8 15.1 18 0.7 20.5 12.0 0. Only a few values are slightly greater.

In analysing orthogonality and particularly the Auto-MAC values. Figure 27..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.Modal Assurance Criterion (MAC) 46 incorporated into the calculation. .

will continue to be developed with strengths and weaknesses. Simplistic tools like the modal assurance criterion are limited in their meaningful application. . 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. The development of related assurance criteria has been initiated by shortcomings. The synergy of MSC/NASTRAN and LMS/Pretest gives the engineer the additional benefit that. removed from the actual development or unaware of application limitations in subsequent implementations. An overview is given of the commonly used techniques to address the target mode selection. of the original modal assurance criterion. it will always be important to recognize the origins and limitations of tools like the modal assurance criterion to avoid misuse of the methodology. that the interpretation of the calculations can be visualized and that both program communicate directly with each other. just like the modal assurance criterion. 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. Even so. In regarding to the uses for aircraft design and testing/certificating companies. the sensor location placement and the positioning of the exciters. 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. Certainly in the next few years. a unique environment becomes available that guides the engineer through the complete process. real or perceived. 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. has been remarkable and is most likely due to the overall simplicity of the concept. Dissatisfaction often has resulted from the misuse of these tools by users. 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. 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.Modal Assurance Criterion (MAC) 47 Conclusions Over the last twenty years. from pre-test analysis over correlation to end up eventually at the model updating step. It is clear that users will continue to need more feedback concerning quality assurance information relative to experimental modal parameters and that new techniques. 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.

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

- Theoretical and Experimental Modal Analysis
- D. J. Ewins - Modal Testing
- PPt on Mimo Channel Modeling
- Confusion
- Modal analysis of a robot arm using the finite element analysis a.pdf
- Linear Regression Project
- Using Radial Basis Probabilistic Neural Network for Speech Recognition
- Bootstrap-After-Bootstrap Prediction Intervals for Auto Regressive Models
- Case Study Report
- AUT for Pipeline Girth Welds 2nd Edition Sample
- 06 Paper 23110936 IJCSIS Camera Ready Pp041-045
- live-1648-2403-jair
- Fang Rfmcae Nn Review 1999
- MPM Wong Parker
- ajustes
- Journal 2007 Jcej Jp
- OPTIMIZED WAVELETS FOR FINGERPRINT COMPRESSION
- IC-L3
- 2 SimulationXpress Study 1
- 1744859
- Ethesis Finally
- Fusion of Multi-spectral Palmprint Images for Improved Identification Performance
- base-Study 1-1
- Day 05 Session
- Eric Lanoix and Jing Yuan Final Project
- Www.math.Uni-hamburg.de Home Iske Papers Blpunq
- Planning the Research Study_ Part 2 - Ouellette
- Fft Calculation
- Anomaly Detection
- Final Presentation of Tools and Techniques of Structure Analysis (1) (1)

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 listening from where you left off, or restart the preview.

scribd