An hourly hybrid multi-variate change-point inverse model using short-term monitored data for annual prediction of building energy performance, part II: Methodology (1404-RP)

In this article, the second in a three-article sequence, the methodology of investigating the capabilities and the limits of hybrid inverse models developed from the shortest monitoring periods possible for a reliable and accurate long-term energy performance prediction in large commercial buildings is presented. Hundreds of thousands of test regression models were created using different lengths of short-term hourly data, ranging from 2 weeks to the entire year. The hourly regression models were created using synthetic data from five different building types and eight different climate zones along with actual monitored data from two buildings. Color area plots enabled detailed analysis of the predictive capability of the hourly hybrid inverse models developed from short-term data sets and conclusions on the necessary length and timing of short-term monitoring were able to be drawn from them. Three application areas for research of this nature were identified: detailed audits, green building performance verification, and verification of post-retrofit claims using pre–post monitored data.