The meter’s running

If you want to get a handle on the energy your building is using it’s probably best not to rely on the billing information provided by the utilities companies.

For the majority of UK buildings the frequency of the information gathered by the electricity, gas and water supply companies rarely provides the level of detail necessary to identify specific faults while many utility bills contain an unacceptably high number of errors. A recent survey showed over 3% of bills to be incorrect and this does not include the large number of estimated bills, of which approximately 70% relate to gas. Such paucity of information does not allow energy managers to effectively get on with the job of saving energy.

But all is not lost. Technology that can collect and transmit high frequency (half hourly) utility meter information is now readily available. And as it is automatically read from the meter and transmitted for storage and analysis, advanced meter reading removes the issue of estimated bills altogether and can reliably provide precise billing periods. What’s more, armed with such high frequency data, it becomes possible for building managers to identify changes in energy consumption patterns throughout the day. When used in conjunction with building occupancy and usage patterns, this information can be used to identify wastage and quantify potential cost and carbon savings.

However, such analysis is not without its difficulties. The shear volume of data makes it difficult to interrogate and clearly identify specific patterns. What is needed are ways to automatically analyse the data as it is collected in order to feed back specific and useful information about a building’s on-going consumption.

To investigate the potential benefits offered by automatic meter reading, a team from London South Bank University and energy management company TEAM (EAA Ltd) undertook a research study - part of a larger study where automatic meter reading was installed in 90 buildings – on a live building to assess the typical overall savings that can be expected from its use as an energy management tool.

Testing the theory

The building chosen was an open-plan, 11 000 m2 museum with close control air conditioning to meet the environmental conditions necessary to preserve the exhibits. This is provided by four rotary screw chillers with a total notional compressor power of 156 kW, which are allowed to operate 24 hours.

The building was served by a single electricity meter (without any sub-metering), and half hourly data was recorded using a dedicated logger and retrieved for analysis using a traditional public switched telephone network modem.

Figure 1 (linked above) shows the half hourly consumption pattern for 18-24th September 2005 and the 9-15 October 2005.

Straightaway, analysis of the first of these periods showed a particularly high night-time consumption of around 140kWh per half hour (280 kWh). The chillers together with associated pumps and fans, excluding the air distribution fans, constituted the largest single load in the building and the consumption pattern suggested they were operating regularly at high load at night.

A site inspection confirmed this to be true, as other night loads (security lighting etc) could not account for such a demand. Environmental monitoring revealed that internal temperature and humidity were controlled to extremely tight tolerances over the entire 24-hour period. Consultations with the staff suggested these conditions could be relaxed at night without compromising the requirements of the exhibits. Recommendations were made to make these changes during unoccupied hours in order to reduce chiller demand - in fact 24 hour relaxation was put into practice.

The week commencing 9th October 2005 in Figure 2 shows the half hourly consumption pattern following these changes to the controls. What is noticeable is the sharp reduction in overnight consumption, together with a reduction in peaks. Variations in peaks (for cooling in particular) can be due to a number of factors: outdoor air temperatures, solar gains or number of visitors.

More detailed analysis is needed to establish savings in this area and the normalisation of short term energy consumption is both problematic and subject to large uncertainties. Variations in base load, however, are less ambiguous. In this case overnight cooling loads (due to casual gains), should be more or less constant day to day and if there are strong seasonal variations base load reductions can be interpreted as pure savings. This shows the value of the timing of energy data received, and how half hourly metering can identify quick wins, and highlight areas for further investigation.

Table 1 (linked above) shows the comparison in weekly energy consumption for the same two weekly periods. This shows savings of around 14,800 kWh, which equates to around £740 per week 6.3 tonnes of CO2 – although in the light of the discussion on peak loads this may not be wholly attributable to the savings measures. Note that to extrapolate these savings for an entire year would require weather correction both at the high resolution (half-hourly) level, and to account for seasonal effects. The application of weather normalisation to high frequency automatic meter reading is a further objective of this study.

A particular issue with the building used for the case study is the nature of the contractual relationship between the building owner and the contract facilities management company. The contract states that rigid conditions are required in the building, even though the building owner can agree to the relaxation of the space conditions. The contractor, on the other hand, is reluctant to break the terms of the contract and to operate a more complex regime for the building. In order for the building owner and operator to immediately benefit from such techniques, a more suitable contractual arrangement is needed to ensure a more appropriate running of the building.

Since this initial trial the evidence of potential savings, together with analysis of internal conditions in this museum, have been presented to the conservation officers of the museum. They are now convinced of the need to alter the operation of the building systems to provide a more appropriate internal environment, coupled to significant energy and cost savings. This approach is to be adopted across the museum estate.

Lessons learned

The acquisition of half hourly data reveals consumption patterns that would otherwise be unknown to the building operator, and from which anomalies, trends and specific faults can be identified. Traditional data collection of weekly or monthly meter readings can show overall trends in energy use, and is indeed particularly useful for aspects such as weather normalisation. High frequency data can be used to compliment these traditional methods in order to rapidly identify and rectify specific faults.

The case study shows how examining half hourly demand profiles can provide fault diagnosis, identify wastage, or be used to demonstrate savings from control or behavioural changes.

The investment cost of automatic meter reading in the case study was around £9500. In general the cost of installation is dependent on the type of meters fitted in the building, the type (or absence of) building management systems, and the contractual relationship with the energy supplier or shipper. Costs are normally less than those for the case study building because this required additional equipment for monitoring chiller consumption, but even here payback looks attractive (probably between 12 and 18 months).

Trends in energy demand can be readily observed which, in conjunction with a reasonable knowledge of the building, can be used to target investigations of waste, whether through plant failure or operational behaviour. This potential for rapid response is a significant advantage of automatic meter reading. It is also becoming apparent how the data can be used to analyse specific system operational issues such as plant start times (and control functions), and enable specific questions to be asked about the operational needs of a building. Techniques are currently being developed that will provide such analysis and exception reporting at a distance and highlight changes and quantify waste and savings.

There are some significant issues that are apparent for the installation of automatic meter reading, particularly in retrofit situations, and they need to be considered. There can be significant barriers to installation from activities of the utility companies, justification of investment costs, and engagement of the building operators (particularly where a contract facilities management system operates). In addition it is important that once the systems have been installed the information is used properly. A danger with large quantities of high resolution data is that it gets ignored or overlooked. What is required is a set of data analysis tools, such as dedicated monitoring and targetting software, that can automatically identify typical faults and produce useful and reliable exception reports. The techniques and systems currently being developed will enable building operators to have better quality of information available to them in order to improve the energy performance of their buildings. For a wide uptake to occur building owners and operators must have confidence that this information is reliable and worth the time, trouble and effort needed to establish these advanced systems.