With Arup’s total engineering domain expertise, we have developed different types of machine learning models to address engineering challenges in different industries. Historical data from sensors, Building Management System (BMS) and people counting system, etc. can be used as inputs for the machine learning model’s training, with the model improving itself through continual self-learning.
Through Analyzing historical data, ML models uncover hidden patterns and give estimation on energy usage for better planning in advance.
Arup has developed and applied an integrated platform based on AI machine learning for the energy forecasting and optimization of HVAC system operation in existing developments with the aim to reduce the total building energy consumption. The platform was developed using thousands of existing HVAC plants’ operating data points as well as hourly regional weather data for evaluation. The digital platform learnt the system’s operating patterns with respect to the building’s occupancy profile in a typical day of operation for predictions on the cooling demand.