Subscribe / Renew
|Return to the Building Green Index|
February 18, 2010
KPG is improving its design engineering decision-making for municipal-industrial facilities using techniques borrowed from financial services consulting.
These are the folks who seek stability in an uncertain financial world. In wealth management, clients need to understand the statistical probability that their investment strategies will obtain desirable results. We made the connection it’s equally important our clients know the best value for their dollar with respect to sustainable project decisions.
A holistic concept
KPG is using “predictive modeling” techniques and tools in two areas: to assist design decision-making, and to identify project-management risks and analyze sustainable-design projects.
“Predictive modeling” is a holistic concept, with a goal of quantifying and reproducing many examples or iterations of a project, each example having similar attributes within a defined range of probable outcomes.
Most of us are familiar with “energy modeling” techniques for comparing a proposed building’s energy performance against a baseline design. Predictive modeling is a different approach, and does not attempt to predict or quantify energy performance.
We’ve taken this on ourselves after recognizing the need for better information than our previous experience would provide. Each of our buildings to this point have been prototypes for the building we wish we had built. We were looking to move beyond our intuition and into early analysis.
KPG found early analytical approaches generally weren’t accessible in our industry apart from a few specialized groups that are using “Monte Carlo” simulations in project risk-analysis applications. Most midsize regional design engineering firms like ours do not have the market reach to justify our employing doctoral graduates in advanced mathematics, so we naturally looked for more user-friendly approaches.
We researched the available software programs used at the university level. Most of these had primitive unfamiliar user interfaces. We ultimately decided to employ an enterprise software package from Oracle.
We were in the process of adopting database-driven scheduling software, and using products from a single provider made sense. The key features include ease of input and analysis in a spreadsheet format.
The software allows simulation of thousands of rapid iterations for a modeled condition. For example, we can estimate the relative benefit of a design decision given a number of factors assigned by the designer.
Each predictive model requires us to define assumptions, define forecasts, run simulations and interpret the results. Each simulation can include thousands of iterations, which is roughly the same as if we constructed the project many times and could keep track of the results achieved for each version.
What we learned
We found it is important to begin simply, and we learned to define simple distributions from the onset. For example, a normal distribution follows a simple bell curve, and it is a simple matter to define a statistical range, often described as a standard deviation. A uniform distribution is another common distribution assumption suited for the kinds of analysis we perform.
Sustainable design decisions help project leaders determine with a degree of certainty the best combination of design strategies and the likelihood the strategies will bear out within a range of performance. For example, KPG wants to know whether it is best to use a window option in concert with an insulation option, or instead use a different structural system in a substantially different building configuration. This early analysis requires multiple simulations employed with a degree of confidence based in prior experience. So we cannot completely remove intuition from the decision-making process, nor would we feel comfortable doing so in most cases.
Carolyn Forbes, a project manager at KPG, manages sustainable design for a LEED platinum project, and is adopting a working predictive model.
Asked about the cost, Forbes said, “We think the process of identifying the range of probable outcomes is a worthwhile activity. But the real benefit will come when we capture the synergistic effects for better implementations in completed projects. And those benefits ultimately belong to our clients.”
Use by project teams
KPG uses the same software for project-risk assessment, in concert with risk-management principles.
We’ve found early analysis of risk management and sustainable design decision-making are two profoundly different activities, requiring unique staff. We can share the software among workgroups, but these are isolated activities.
We’ve not yet found a way to achieve the best results in either application within a project team. We think this experience points to different directions in our industry taking shape in the future.
Some teams will be good at risk-management applications, and others skilled at sustainable decisions modeling. Also, we have concerns with the relative complexity and possible misinterpretation of data. Our implementation strategy includes quality-control measures we believe will help manage outcomes and clarify expectations.
Early analysis is challenging our thinking, and we expect the results will be borne out over time. We are looking for confirmation of energy and resource savings through commissioning, including measurement and verification. Ultimately our goal for projects is that facilities will function correctly and efficiently by design, in concert with users’ needs.
We hope that in 10 years these techniques are employed on all capital projects. We’ve found it is a blend of many disciplines, and that high-level mathematics is a difficult fit for most design firms. But there is an important place in our industry for the reliable prediction of performance benefits associated with green building practices.
Doug Brinley is the principal architect of KPG, an interdisciplinary design firm with offices in Seattle and Tacoma.