An enhanced weather normalization method for identifying changes in the building condition

• Published in: Journal of Building Engineering 2021-08, Vol.40, p.102354, Article 102354
• Main Authors: Tam, Claire, Liao, Zaiyi, Poh, Paul S.H.
• Format: Article
• Language: English
• Subjects: Building condition assessments; Energy assessment; Energy-weather relationship; Multiple year building energy simulation; Weather normalization
• ISSN: 2352-7102; 2352-7102
• DOI: 10.1016/j.jobe.2021.102354

Besides physical testing on buildings to detect the performance degradation of building components during Building Condition Assessments, energy consumption data should also be indicative of these defects. However, the methods to accurately analyze the changes in energy consumption over time have rarely been sufficiently evaluated. This paper compares the effectiveness of conventional weather normalization methods in deriving an accurate indication of the impacts on energy consumption from the changes in building condition. A proposed method based on the degree-days method is presented which aim to increase the accuracy of weather normalization by improving the representation of the energy-weather relationship through (1) excluding the energy consumption during the “shoulder seasons” from the dataset; and (2) applying a weighting exponent to the degree-days ratio. Multi-year building energy simulation was conducted to generate reference values of Typical Year energy consumption without influences from weather variation and the energy consumption affected by Heating, Ventilation, and Air Conditioning system degradation. Normalized annual space heating and cooling energy consumption data were compared to the reference values. Proposed method normalized values from the simulation study are within 2% Coefficient of Variance of the Root Mean Squared Error of the reference values, which is an improvement from the conventional degree-days ratio-based method by 59% for heating energy consumption and 9.7% for cooling energy consumption. Measured data of a case study were normalized to test the proposed method which confirmed minimal changes in energy consumption of the house over the years. The proposed method will be further evaluated when the data for more case studies with known building condition become available. This improved methodology can potentially be used to analyze how building condition degrades over time and its corresponding impacts on energy consumption, as well as to compare building condition across climate regions. •Building performance degradation revealed by weather normalized energy consumption.•Improved degree-days method is effective as linear regression weather normalization.•Weather normalization should consider building specific energy-weather relationship.•Energy consumption uncorrelated with degree-days should be excluded in normalization.•Simulated energy consumption data show building sensitivity to weather variation.