A novel method of determining events in combination gas boilers: Assessing the feasibility of a passive acoustic sensor
Thomas Neeld, UCL Energy Institute
For researchers, having access to tools that monitor energy systems is imperative to understanding the behaviours of energy consuming devices in the home. Such tools may help to analyse the impact of energy-saving interventions, whether they be behavioural or environmental (insulation improvement for example), or to determine if energy systems are fit-for-purpose.
In the UK over 70% of energy consumed in the domestic sector is done so in gas-fired boilers (2013-2014); with the combination gas-fired boiler becoming more and more prevalent (over 60% of all heating systems in 2012). Thus the combination gas-fired boiler is arguably the most important single technology type to consider in regard to monitoring energy systems in the domestic sector. Researchers have however found it challenging to accurately determine consumption of combination gas-fired boilers for space heating use and often relied on estimates from vast datasets or using temperature sensors.
This project looked to develop a method to ‘spy’ on combination gas-fired boilers in order to determine accurate estimates of energy consumption and demand type (hot water or central heating). The method by which to do this ‘spying’ was selected to be acoustics: Noises produced by boilers are available externally to the boiler and can in theory provide a significant amount of ‘free’ data if one can ‘understand’ the language of the system; similar to how recent developments in speech recognition software can ‘understand’ human speech. Due to the complexity of such analysis, among other techniques, application of machine learning methods are required – this is a growing area of computer science which has found a multitude of applications.
The aims for the project can thus be summarised as follows:
Based on the acoustic data received from combination gas-fired boilers:
- Develop and test techniques which determine the demand type of operation (hot water or central heating)
- Develop and test techniques which determine the time of burner ignition
- Develop and test techniques which determines the intensity of a burn
Images of the setup:
Spectrogram showing frequency distribution of sounds produced by a boiler under investigation. Here we can see the noise produced at the point of ignition (infrasound pressure pulsation):
During the MRes year the project was successful in meeting the first two aims for the boiler under investigation: acoustically determining demand type and the time of ignition to a high levels of accuracy (100% and 97% respectively). Additionally it was deemed that significant potential existed for more work in this area, thus the project was taken on by the researcher onto a PhD – see W5UP10.
- Setting up the start of a PhD project for the researcher with more ambitious aims – see W5UP10.
- Collaboration developed between Imperial College London, Dept. of Electrical and Electronic Engineering.
- Interest developed with industry bodies to potentially develop a collaboration.
Publication in Building and Environment based on work done during MRes:
Neeld, T., Eaton, J., Naylor, P. A., & Shipworth, D. (2016). A novel method of determining events in combination gas boilers: Assessing the feasibility of a passive acoustic sensor. Building and Environment, 100, 1-9.