Are we “boiled” for choice?

August 3, 2017

A new analysis of historical boiler repair and replace decisions has revealed which make of boiler costs the least to repair.  The study used machine learning processes to come to its conclusion.

Collaborating with the University of Surrey through a Knowledge Transfer Partnership (KTP), is allowing Sustainable Homes to draw upon knowledge and expertise that we would not otherwise have access to.  The KTP includes Dr Lilian Tang from the Department of Computer Science and Dr Wolfgang Garn from Surrey Business School.  The findings showed that a simple change to Hastoe’s approach to boilers could save up to £45,000 per year.

Before explaining the detail of this finding, it is useful to understand the machine learning process.  In this study there were three crucial steps:

  1. Gather data – this included gaining repairs costs for 17,000 repairs over 14 years. It also included finding out other features about the properties.  For example, data such as boiler age, previous year’s repairs costs, size of home and energy efficiency were gained for the study.
  2. Prepare data for analysis – this is arguably the most time consuming part of the process. All the data has to be assembled into one database that use the property UPRN as a single point to link the data together
  3. Use machine learning techniques – in this case multivariate linear regression was used to generate a predictive model to predict the repairs costs in the next coming year. The intent was to use this repairs prediction cost to help decide whether or not it was cheaper to repair or replace the boiler.

For each home the repairs costs were predicted for the next 6 years.  If the total of these costs added up to £1500 or more (our assumed cost of boiler replacement) then, we deemed it cheaper to replace the boiler in the event of a breakdown, instead of repair.  In the event, this analysis found that not many (39) homes would have benefitted from this decision. Even adding the maintenance teams decision making time, would bring the savings to around £2,000 a year.  Perhaps this is not a big enough gain to take the decision to “trust” the predictive capability of decision making.

The real benefit in this study, came from step 2 of the process which was to prepare the data into an analysable form.  This made it easy to see which model of boiler cost the least to repair in the last 14 years.  This enabled us to put together a league table of boilers.  The cost per boiler includes, those homes where no boiler repairs were reported.  In other words, the savings come from a) lower cost per actual repair and b) lower incidents of repair.

Boiler league table (figure 1) shows that switching all boilers to “G” could save £45,000 per year.

Figure 1

“This analysis will certainly give us something to consider when we replace boilers in future. We will also inform our new build colleagues“. Paul Nicholson, Head of Business Performance at Hastoe

The next step for the analysis will be to see if the predictive model can be used to extend the life of a boiler i.e. see if the likely repair costs are suitably low so that life can be extended beyond the usual 15 years.  An example curve for one of the models of boiler looks like the below (Figure 2), suggesting it could be extended to over 20 years.  This would also involve a whole other level of analysis.  This is because replacing old boilers improves the energy efficiency of homes, which in turn reduces management costs.

Figure 2

If any of the below machine learning applications would support your business, please get in touch:

  • Predictive maintenance – when is it cost effective to replace components
  • Data filling – essential if you want to make strategic decisions without having to wait for lengthy and costly stock condition surveys
  • Condensation and mould – what are the main factors in predicting mould problems
  • Voids and arrears – what factors (environmental or otherwise) contribute most to arrears and voids and how can they be managed

Alternatively, please ask us about any other ideas.  Nothing is off-limits – predicting resident satisfaction, general wellbeing, risk management of contractors… the possibilities are endless. Please contact Richard for more details.

Read Richard’s blog on how machine learning can lead to cost savings and environmental improvements The rise of the machines – learning for environmental good and cost savings

Richard Lupo

Richard Lupo

Areas of expertise: developing and instigating stream-lined processes to ensure environmental effectiveness