By Colin Grenville*

This example considers a real project to deliver an energy performance contract across a number of schools using a range of ECMs at each school. Whilst savings at most schools were approximately as expected, one seemed to perform adequately in the first measurement period with savings worsening dramatically in subsequent years. Unfortunately, an absence of detailed monitoring and regular customer dialogue following ECM implementation meant that the underlying issues were not identified until the second year-end reconciliation was calculated.

At this point retrospective investigations were undertaken to identify why savings had not been achieved. These were completed using a range of techniques including data analysis, site survey, end-user dialogue and indeed the use of web-searches to explore whether publicly available information about the school could yield any clues. Whilst this was in fact a fairly small project in terms of expected cost savings, the challenges of accurate measurement became large - illustrating issues that can arise from following a whole facility measurement approach.

Project background

A secondary school facility had multiple and interactive ECMs installed during a single implementation phase including: lighting upgrades, building management system optimisation, thermostatic radiator valves, new boiler etc.

Anticipated savings were >10% of baseline therefore the M&V plan agreed was based on Option C for the whole school facility using existing utility metering for gas and electricity. Whilst alternative approaches were considered at design stage, the cost of additional metering required would have been prohibitive relative to the anticipated savings and the interactive nature of some of the ECMs suggested that only a whole facility approach would correctly capture the performance.   The absolute value of anticipated cost savings is not massive therefore only a modest M&V budget existed. The measurement period was set to align with the academic year, starting in September following the installation of ECMs during the summer holiday period. Utility metering provided half-hourly electricity data but only monthly gas meter data from customer invoices or manual meter readings. Savings achieved reduced year on year.

Investigations identified a significant increase in student numbers at the start of the second academic year resulting in various changes to buildings and energy demand. Initial small increases in student numbers were absorbed into existing classrooms, then “temporary” additional modular classrooms were added and internal space was re-configured to create more teaching space within the main school buildings. Whilst the M&V plan had captured some out of hours’ activities this baseline data was not adequately detailed and additional activities were found to occur in the reporting period that had not been captured in the original baseline data.

Whilst the M&V plan recognised that changes in occupancy numbers and times could occur, it did not propose an adjustment methodology for large changes in pupil numbers.

The number of students was identified for each academic year, including for the baseline period.

(Figures in the table above are illustrative in the interests of client confidentiality)

Questions for consideration:

1. What additional data might you need to obtain in order to estimate/calculate a robust adjustment to account for changes that have occurred?
2. What a non-routine adjustments may be appropriate in this situation?
3. Is ongoing measurement likely to be reliable given the significant change in site occupancy, floor area and building utilisation?

Firstly, whilst we have identified data for the year-on-year change in student numbers there are a number of other data items that need consideration relating to the change in building fabric, floor area and utilisation. The following list provides some suggestions and is not intended to be exhaustive.

• Change in total conditioned floor area - what has been added or modified and to what extent?
• What changes to existing installed plant or to the installed ECMs as a result of building modifications?
• When did the building works take place? This is probably a series of dates which may enable changes to be spotted in the energy consumption data, and provides start dates for any baseline adjustments.
• Has any new energy consuming plant been added? In fact, the modular buildings added had their own electric heating, lighting etc. Loads for these need to be identified and calculated using reasonable seasonal profiles and an understanding of the control strategy deployed – and perhaps using inexpensive temporary temperature logging to support analysis.
• Has increased occupancy caused changed demand on central services – e.g. catering gas load, space heating (or indeed cooling)? In fact, some areas of the main school building had been converted from circulation areas or infrequently used meeting places into permanently heated class rooms resulting in higher temperature set points in those spaces.
• To what extent are these changes permanent and how will these be captured in an updated M&V plan – e.g. through permanent baseline change or through a specific annually calculated adjustment?
• Are there any external benchmarks of energy use that could be used as a fair and cost effective alternative to a detailed engineering calculation and adjustment?

Secondly, in consideration of what may be appropriate non-routine adjustments, it is important to ensure one does not apply multiple non-routine adjustments which would account for the same issue twice. For example, if you apply an adjustment to account for increased occupancy numbers then would a further adjustment to reflect additional classrooms and floor area also be valid or could this cause an over-adjustment?

The answer is informed by data availability. If the new modular classrooms had been provided with their own electricity sub-meter then electricity use therein could be deducted from the whole facility and increased student numbers associated with those specific additional classrooms could be excluded from adjustment (whilst still needing to recognise that they may impact on central services in the main buildings like catering gas load.) In fact, no such sub-metering was installed by the school leaving a risk of double-adjustment for both student numbers and floor area. The IMPVP expects reported savings to be conservative therefore this isn’t a reliable solution. The uncertainty here would be challenging to accurately quantify.

There is also the issue that the school had increased its out of hours’ utilisation for community events, training courses, fitness classes etc. Whilst the additional hours of each event was identified from publicly advertised sources, the specific energy impact of each could not be determined without significant additional investigation. A fitness class immediately following the school day might have minimal impact – a few lights left on for an hour and no changes to heating strategy. However, a weekend course in one particular classroom in the middle of winter might require the central boiler plant to operate and heat whole zones of the school in order to heat only a small actual area required. Such detailed investigation and calculation is probably disproportionate to the M&V budget available and would require significant input from both the ESCO and their client. The cost of such a detailed approach would swiftly cannibalise the anticipated project cost savings.

Please take the time to go through this exercice alone or in group within your company. Come back on June 25 for the solution.

NOTE: To see the solution, you must have an EVO website subscription. Please login to your account to see the solution and access M&V Focus additional interactive features. Follow this link to know more about EVO web subscriptions.

 (*) Colin Grenville is the President of Erebus Environment.