Professional Documents
Culture Documents
Ryan Solnosky1, Ph.D., E.I.T, M.ASCE, Brian Quinn2, PE, Lisa Willard3, PE
1
Research Associate, Dept. of Architectural Engineering, The Pennsylvania State University, Email:
rls5008@engr.psu.edu
Downloaded from ascelibrary.org by Nanyang Technological University- Library on 11/11/15. Copyright ASCE. For personal use only; all rights reserved.
2
President, SE Solutions, LLC, Holland, MI, Email: Brian.Quinn@FindYourEngineer.com
3
Vice President, SE Solutions, LLC, Holland, MI, Email: Lisa.Willard@LearnWithSEU.com
INTRODUCTION
and the adequacy of models needs to be appraised with respect to these goals (Alvi
2013). Adding to this, computational analysis should be used as a means of validating
the behavior only after the engineer has developed a sufficiently comprehensive and
detailed vision of the structural system to permit analytical work (Luth 2011).
Engineers often face modeling and analysis issues when designing irregular and
complex shaped buildings as the simplistic notions and interactions are not enough
Downloaded from ascelibrary.org by Nanyang Technological University- Library on 11/11/15. Copyright ASCE. For personal use only; all rights reserved.
(Shin et al. 2008). Modeling and the associated interpretation of the results from each
analysis and design are not simple tasks. They can be complex even for structures
that have static loads and with linear response, even more so when dynamic and
nonlinear concepts are introduced (Powell 2008a). A major advantage in computer
utilization is the repetitive evaluation of parameters and variations used to narrow
down feasible systems (Carpenter 2005).
Knowing the structural domain scope that software encompasses, it is quite
clear that software can easily become too complex to effectively manage and utilize.
Beyond complexity, software has a growing variety of programs that are available.
With each type and vender updates are frequently made available for these programs.
Updates are critical as they provide solutions to internal glitches in software and
update code provisions. As a result, it is a precarious balance between increasing
productivity by using these tools while limiting the risk of errors when using the
programs.
This paper provides greater insight into errors and ultimately how to reduce
them across structural sectors. The focus will be twofold, the first is with analysis and
design software and the second is on Building Information Modeling (BIM) software.
Commonalities between the two sectors will first be discussed followed by specifics
on how they are applied to each.
To understand errors and how to reduce them, one must understand the
software and the nature behind errors before they can be reduced or eliminated. Many
current practitioners and practitioners turned academics have provided strong
justifications as to how we should approach structural computing. They also have
provided warnings on what makes for these types of errors. To summarize Emkin
(1998), Krawinkler (2006), Luth (2011), and Powell (2008b) to name a few, the
following notions are paramount to being able to accurately represent a physical
project in a digital setting.
Knowledge of Buildings:
1) Modelers need to understand the process of how a building is constructed and
to realize how their actions affect the bigger picture.
2) The engineer can no longer rely on someone else to figure out his or her own
incomplete knowledge at hand when the information is generated.
Knowledge of Structures:
1) A deep understanding of structure principles based on material behavior,
system behavior, modeling techniques, analysis methods, design procedures
and codes, and error assessment are essential prerequisites of leveraging the
power of automation.
2) Sound engineering judgment and intuition around building principles, codes
and standards are necessary for safe and efficient results.
So, to prevent problems such as operator error and lack of knowledge about
software, a method developed by Willard and Quinn (2012a-b; 2013a-b) can be
adopted for structural based programs. This method intends firms and practitioners
develop a Software Error Reduction Plan (SERP). The SERP was first introduced in
2009-2010 through a series of articles were published in Structural Engineer
Magazine (SE University 2012). The documents provided key steps that could be
used to create a tool that would help in reducing the risk of errors while using
structural software.
When formulated correctly, the SERP can help significantly reduce the
chance of errors being made while also increasing the chance of catching any
problem areas that could arise. The four steps for a SERP include:
(National BIM Standard), agencies (e.g. GSA and USACE), and research
organizations (Penn State and Stanford) are developing guides and best practice
recommendations on how to implement BIM. Two of the more recent guides that
provide excellent reference into BIM usage and how to obtain accurate results are
from the Penn State Computer Integrated Construction Research Group.
The first guide is the BIM Project Execution Planning (PxP) Guide. This
document focuses a structured procedure for planning the process of how BIM will
be executed on a project (CIC 2009). The main steps are:
1. Identify high value BIM uses during project planning, design, construction
and operational phases.
2. Design the BIM execution process by creating process maps.
3. Define the BIM deliverables in the form of information exchanges.
4. Develop the infrastructure in the form of contracts, communication
procedures, technology and quality control to support the implementation.
The second guide is the BIM Planning Guide for Facility Owners. This
document helps in planning and integrating BIM throughout an organization and for
the lifecycle of the building facility through a BIM organizational strategic planning
and BIM project procurement planning (CIC 2012; Chunduri et al. 2013). Each of
these are discussed more specifically related to structural engineering later but they
encompass:
These basic components to the SERP, BIM PxP and BIM Owners Plan are
broadly similar. Each have the capacity to meet the needs of reducing errors within
both domain sectors through a common platform of identify the situation and
providing measures to reduce error.
Analysis and design can be argued as the most critical topics to model within
Downloaded from ascelibrary.org by Nanyang Technological University- Library on 11/11/15. Copyright ASCE. For personal use only; all rights reserved.
structures as they determine 1) the structural system, 2) the behavior to the manmade
and environmental conditions it will see during the lifecycle and 3) identify the
capacity of the designs. These models are mainly based on the application of
structural mechanics (sometimes idealized), as well as models of structural elements
calibrated by lab testing (Alvi 2013).
Mistakes here can often be less obvious and have a greater chance of being
overlooked. Analytical methods often, whether directly or indirectly, are driving the
design to select systems and elements based on a wide variety of criteria. Early
methods in the lifecycle look at more holistic provisions and best practices to narrow
system possibilities. Many of these do not have strong data support in authenticity,
instead they focus on the designer experience. Some of the most common errors
observed by the authors in modeling analysis and design are listed in Table 1. As it
can be seen, the errors are quite varied.
the fundamentals.
3. Limited users should have the skills to understand of how to check and verify
design criteria and results.
4. In new versions of software, one individual should be responsible to
investigate and fully understand changes.
Checks and reviews are the focus of the third and fourth steps in the SERP
(Review and Create Processes and Procedures and Perform Internal Reviews of
Models). By establishing processes on a project for creating models, engineers have a
framework to start building their models. This can also be paired with checklists,
which would allow for the model to be reviewed ahead of critical steps to confirm
that certain calculations and requirements have been made or were accounted for.
While performing internal reviews of models, engineers have a chance, not only to
have another set of eyes review their model, but also have the opportunity to work
collaboratively, and learn new information that can be used on subsequent projects.
Areas and topics that can be checked and verified within models and results are
summarized in Tables 2 and 3.
Focused within reviews is the ability to interoperate the results given by the
software as blind agreement with software should never be trusted. These checklists
and internal reviews boil down to that engineers need to be able to rapidly assess the
overall performance of a building structure subjected to multiple actions. These
actions look at the net overall effect and localized effects. Tables 2 and 3 clearly
show the wide variability of items that should be looked at to provide a
Downloaded from ascelibrary.org by Nanyang Technological University- Library on 11/11/15. Copyright ASCE. For personal use only; all rights reserved.
• Explicitly identify:
o The assumptions underlying a model.
o Identify what a model leaves out.
• Conduct independent peer reviews and checks during and after model
development.
• Perform sensitivity studies of its parameters to check a model’s robustness.
• Perform equilibrium checks and test models with simplified load cases.
• Investigate discrepancies between results from different models of the same
structure until they can be explained.
cross, most of which are error related such as: technical content, accuracy, and
inclusion and exclusion of material and also confidence in the modeler and the firm.
Recently, suggestions for successful BIM adoption have been presented in
books, reports, on the Internet, and at conferences and webinars. Often limitedly
discussed, however, is the need for good modeling practices in how to reduce errors.
This is particularly true with companies that have limited manpower, time, and
Downloaded from ascelibrary.org by Nanyang Technological University- Library on 11/11/15. Copyright ASCE. For personal use only; all rights reserved.
financial resources often do not know how to approach the complexity of BIM
modeling (Won et al. 2013). Companies that adopt BIM often find a lack of
understanding of the software and process that naturally leads directly to inherent
errors in the process and models. Typical errors observed can be found in Table 4.
the opportunities, limitations, and implications of what the models will be used for
need to be defined so that everyone knows what their efforts can alter. At each
different phase, the models may change in type and complexity which leads itself to
more errors. The difference though is not significant. Specifically looking towards
educating staff, two areas should always be considered in developing a training plan:
1) the type of experience needed by all members and 2) how the technology can
Downloaded from ascelibrary.org by Nanyang Technological University- Library on 11/11/15. Copyright ASCE. For personal use only; all rights reserved.
To ensure that the information is input properly, the modeling of the data rich
attributes is critical. To grasp what information needs to be modeled, it is important
for the author and receiver of each information exchange transaction to clearly
understand the information content. The best way to ensure information is properly
modeled, it is recommended to step through each element type of the facility and
determine if 1) a visualization of that element would be beneficial and sufficient or if
2) additional data rich properties are required.
These decisions need made in each BIM Use at each lifecycle phase. An
example is if in Schematic Design (SD), material properties are not needed but in
Construction Documentation (CD) they are needed or if sequencing software doesn’t
require moments but analysis software requires them. This is associated with
developing a Level of Development (LOD) for each element type for each use and
each phase. AIA E202, AIA E203, and BIM Forum’s LOD Specification is a good
start for what is typical for structural systems (AIA 2008, 2013; BIM FORUM 2013).
These however only look at visual based information and not the rich data attributes
currently. Table 5 lists common areas to identify before generating a BIM model.
To make these decisions effectively, the purpose of the information and how
it will be utilized should be understood. Most activities in the lifecycle either
consume and/or produce single or multiple information items. This information can
best be classified into three main categories to differentiate them. The first is Product
information where items that related to the building’s physical elements and
properties, the second is Feedback information where items related to the control and
decision making characteristics on the building performance, and the third is Process
Control information where items related to the restriction and control over how the
sequence of a set of activities is done. More information into this can be found in
Solnosky (2013).
Beyond understanding what the program and process can do, it is also
important to make sure that checks and balances are completed throughout the
phases. This is the last main critical step needing to be conducted on each model
frequently. The primary items to perform an internal review on are varied but they
have common themes among the different BIM uses. Common themes are listed in
Table 6.
• Use the software as it was designed to be used and don't use shortcuts purely
for the production of documentation.
• Embed all the data you can into your model, then use it.
• Check your model for consistency before issuing.
• Provide documentation to others on how your model is structured.
• Define the degree of precision (LOD and measurements) to your model.
• Insist on a robust coordinate base point.
CONCLUSION
tools as absolute correctness nor should they rely on them without verification. It is
imperative that we comprehend that the objective of software is not to predict ‘exact’
behavior for a structure as all models are simplifications and abstract representations
of actual structures. Consequently, errors could always be present in the models
depending on the project and the modeler.
Errors, as it was mentioned, can always be present where software is used.
Methods nevertheless exist to limit them or isolate and correct them. These errors
span all domains of structural software from analysis to design to BIM. While the
details may be different, the overall goal and concept is relatively similar. Four
paramount steps exist within a Software Error Reduction Plan (SERP) to help identify
or limit errors on a project. They are: select a champion, create a training plan, create
software usage written procedures in your office, and perform internal reviews.
No matter who you are in a firm or your role as a structural engineer there are
embedded notions within SERPs that bring to light just what an engineer is and what
software is and most importantly how to manage it. It is always important to have
engineering intuition skills and software specific skills to conceive, check, and
understand the model. Information can then be taken and applied to where real
engineering comes into play in making specific selections for that project.
References
Alvi, I.A. (2013). “Engineers Need to Get Real, But Can’t: The Role of Models”,
2013 ASCE Structures Congress, May 2-4 Pittsburgh, Pa, 916-927
American Institute of Architects (AIA) (2008) “Document E202-2008 Building
Information Protocol”, American Institute of Architect
American Institute of Architects (AIA) (2013) “Document E203-2013 Building
Information Modeling and Digital Data Exhibit”, American Institute of Architect
BIM Forum (2013). “2013 Level of Development Specification for Building
Information Models,” AGC BIM Forum, 8/22/2012, pages 125
Carpenter, L.D. (2005). “Influences on structural engineering”, Struct. Design Tall
Spec. Build, 14, 419-425
Chunduri, S., Kreider, R., Messner, J.I. (2013). “A Case Study on Implementation of
the BIM Planning Procedures for Facility Owners”, AEI 2013, University Park,
PA, 691-701
Computer Integrated Construction Research Program (CIC) (2009). BIM Project
Execution Planning Guide – Version 1.1”, October 8, The Pennsylvania State
University, University Park, PA
Computer Integrated Construction Research Program (CIC) (2012). "BIM Planning
Guide for Facility Owners", 1.02 ed. University Park, PA, USA: The
Pennsylvania State University, University Park, PA