Normative decision theory increases trust and efficiency in building energy retrofits

Significance Statement

Retrofitting reduces a building’s energy consumption which leads to savings for many years to come. Realizing such savings requires an initial investment for the promise of an attractive rate of return, which has to be backed up by rigorous analysis and an investment valuation method. According to Dr. Qinpeng Wang and his PhD advisor Godfried Augenbroe, this is exactly where “Conventional Investment methods have fallen short. They provide potentially unrealistic savings predictions when they do not consider the performance uncertainty in the predictions, thereby hiding a risk for the investor”.

The broader context of this research is that even though many studies have shown that deploying energy efficient technologies across various sectors has the lowest cost per unit of saved energy, our industry suffers from the “energy efficiency gap” between potential and actually performed energy efficiency projects/savings. This is created by a variety of market barriers, with lack of information being the most cited and widespread one. It might be surprising for some readers to realize that transparent information is difficult to acquire: the uncertainties and risks involved with energy efficiency projects are usually difficult and costly to quantify, so many Energy Service Companies (ESCOs) choose to ignore them and simply present the best (deterministic) engineering estimate to clients. Consequently, clients could charge a high risk premium on energy efficiency projects, because of a mistrust towards estimated future returns. In addition, not all clients are sophisticated enough to process information given its quantity and complexity, and make decisions that maximize their well-being.

The research team from the Georgia Institute of Technology in Atlanta proposed to address prevailing drawbacks by the application of axiomatic utility theory in the decision-making process. Incorporating the risk preference of the decision maker can manage the uncertainty of predicted energy savings. Construction projects will have a bigger appeal on investors provided uncertainty in future energy performance is taken into account. “This will effectively provide the risk–transparent information to all parties that could lead to retrofit contracts that maximises the well-being of all parties” says Wang. This contracting process can in fact be automated, increasing the realization rate and thus benefiting the whole construction market. The study is published in peer-reviewed journal, Automation in Construction.

Transparent uncertainty analysis brings increased trust in the outcome of a retrofit project and smooths the contract development, by considering two prevailing sources of uncertainty in the prediction of future energy savings. The first and obvious one is the uncertainty in the future conditions and use of the building and the physical properties of its many components. The other source is the accuracy of the building energy model used for prediction. The two sources combine into prediction uncertainties that a decision maker needs to consider. In addition, the individual preference towards risk cannot be ignored in this context. This preference manifests itself when a decision maker either weighs the risk of future loss heavier than potential gain or weighs potential gain heavier than future loss. This subjective aspect has to be taken into consideration in the valuation, hence knowing the preference of the decision maker is essential to verify, given a certain quantified performance risk, whether the retrofit is a beneficial proposition.

Normative utility theory is well suited for decision making under uncertainty. With such theory, various alternatives are analyzed and ranked with the decision maker’s preference of risk. Wang and Augenbroe add that “understanding how the decision is actually made is quite complex, and the best we can usually do is to develop a descriptive approach”. One of the elements in this approach is that, as axiomatic theory states, the decision makers should be assumed rational based on their risk perception.

The theory is applied to a special contracting method, called energy performance contracting, which involves the ESCO, the building owner or client, and (in some cases) a separate lending institution, A performance contract has a couple of unique traits, for instance, it “guarantees” a certain performance and is meant to spread the risk and responsibility over various entities in the retrofit contract. However, performance contracts are often quite complex and come with pre-specified stipulations, the risks associated with which often elude investors. Furthermore, most investors may not understand the important distinction between four different savings clarified in the paper, namely projected saving, proposed saving, verified saving and actual saving. The research team has used game theory to inspect how ESCO and client will arrive at the best mutually beneficial contract. Using the backward induction, the “best choice” is found upon the premise that each party identifies the most beneficial action, given an expectation of the actions the other decision maker may take.

Wang and Augenbroe offer the conclusion that “Our energy performance contracting methodology ensures a certain level of performance and thereby increases the trust between clients and Energy Service Companies”. The fact that the process can be automated reduces cost and increases efficiency in the energy retrofit business as a whole.

About The Author

Godfried Augenbroe, Professor, School of Architecture, Georgia Institute of Technology

Godfried Augenbroe has a 35 year track record of teaching and research in the building science. His main expertise is in modeling and simulation of buildings at various scales. He is internationally recognized in promoting professional use of building simulation and has served on the board of the International Building Performance Simulation Association (IBPSA).

Augenbroe has been main advisor of 30+ PhD graduates in Europe and the USA. He teaches graduate courses and conducts research in the fields of building performance, computational building simulation, indoor air quality, intelligent building systems, uncertainty and risk, system monitoring and diagnostics.  He serves on the scientific board of five international journals and has published over 200 refereed papers and three books.

Augenbroe has obtained major research grants in Europe and the US. Most recently he received the NSF EFRI-SEED award “Risk conscious design and retrofit of buildings for low energy”, $2.0 Million, 2010-2014. He has given many keynote lectures, most recently at the 2015 Building Simulation Conference in Hyderabad where he also received the distinguished achievement award from IBPSA.


About The Author

Qinpeng Wang, Research Assistant, School of Architecture, Georgia Institute of Technology

Qinpeng Wang received his B.S. in building science from Tsinghua University in Beijing, China in 2011 and then came to the United States to pursue graduate study. Since entering the Ph.D. program in the College of Architecture at Georgia Tech, he has been involved in a research project funded by a $2 million grant from the Emerging Frontiers in Research Innovation (EFRI) program of the National Science Foundation. During his doctoral studies, he has investigated advanced energy management options for buildings, data-driven approaches to stochastic occupancy behavior, the uncertainty and risk quantification of building performance, and innovative retrofit financing models with transparently quantified risks.

Wang recently started his career as a senior engineer/data scientist at Siemens Building Technology, working on fault detection and the next generation of building data analytics.


Journal Reference

Qinpeng Wang, Benjamin D. Lee, Godfried Augenbroe, Christiaan J.J. Paredis, An application of normative decision theory to the valuation of energy efficiency investments under uncertainty, Automation in Construction, Volume 73, January 2017, Pages 78 – 87.

Georgia Institute of Technology, 247 4th St. NW, Atlanta, GA 30332, USA