Saving energy by anticipating hot water production: identification of key points for an efficient statistical model integration

Abstract

This work aims to evaluate the energy savings that can be achieved in domestic hot water (DHW) production using consumption forecasting through statistical modeling. It uses our forecast algorithm and aims at investigating how it can improve energy efficiency depending on the system configuration. Especially, the influence of the DHW production type used is evaluated as well as the water tank insulation. To that end, real consumption measurements are used for model training. Then simulations are run on using TRNSYS software to compute the total energy consumption of DHW production systems over 1 year. Simulations are also based on real consumption measurements for realistic results. To appraise the energy savings, we compared simulations that consider either no forecast (reactive control), perfect forecast (to estimate the control ability to consider forecast), or the forecast provided by our algorithm. The measurements and simulations are run on 26 different but real dwellings to assess results variability. Several system configurations are also compared with varying thermal insulation indices for a complete benchmark of the approach so that an overall performance of the system and the anticipation strategy could be evaluated.

Electric Water Boiler Energy Prediction: State-of-the-Art Review of Influencing Factors, Techniques, and Future Directions

“…Artificial intelligence (AI) and machine learning (ML) algorithms thrive on data, and the quality and size of the dataset significantly impact the performance and applicability of these models in realworld scenarios. Researchers emphasize the importance of large, high-quality datasets to train algorithms that can make accurate predictions and decisions across different situations [100,138,139]. It is evident that data privacy is essential in the realm of electrical water boilers' energy consumption prediction, surpassing the relevance of data quality, availability, and privacy.…”

Sustainability matchmaking: Linking renewable sources to electric water heating through machine learning

“…The authors recorded 97% of the DHW usage profiles and 92% of the electricity usage profiles as a correlation between the simulated and calculated results. Using consumption forecasting through statistical modelling, Denis et al [28] evaluated the energy savings that can be achieved in DHW output. The authors reported the energy savings of 3.6% to 12.8%.…”

“…ML evolved from Artificial Intelligence (AI). ML methods are used to characterize algorithms that learn from existing data and these algorithms use a large amount of data for the learning process with a very limited number of input features [28]. There are three major ML methods available for behavioural analysis.…”