Proceedings of BSO Conference 2022: 6th Conference of IBPSA-England
Predicting Operative Temperature with Machine Learning (ML)
Aritro De, Amanda Thounaojam, Prasad Vaidya, Divij Sinha, Sooraj M RaveendranIndian Institute for Human SettlementsAbstract: With climate change, low carbon space-cooling approaches are becoming more important. Cooling energy demand can be reduced through new interventions, low energy systems, and optimised operation. The adaptive comfort model for mixed mode operation can be a promising approach to the cooling energy challenge. However, adaptive models use indoor operative temperature, which requires the measurement of air temperature, air velocity, and globe temperature in a space. Collecting real-time and long-term data for these is difficult.
This paper summarises a study on an affordable cooling approach to develop a machine learning algorithm to predict OT. Field measurements and Energy Plus simulation were used to create large datasets, 75 % of which were used to train the machine learning algorithm to predict operating temperature, and the remaining 25% were used for testing the algorithm.
The testing of the OT predicted with the random forest model shows an RMSE of 0.34%. In terms of classification of the thermal environment as being in/out of the adaptive comfort band, 0.88% of values were misclassified. When the predicted OT values were compared with the one-week measured OT values, the RMSE was found to be 3%.
The results demonstrate that our algorithm that uses indoor air temperature readings in a space and outdoor weather station data can reliably predict OT. This enables a scalable and affordable approach for accurate and long- term prediction of OT to determine the comfort condition.
This will enable control systems to use OT to determine thermal comfort in a space using adaptive comfort models and to account for ceiling fan usage to reduce or eliminate air-conditioning (AC). Keywords: Thermal comfort, Machine learning, Operative temperature, PredictionPaper:bso2022_63