Sigma Spotlight – Monitoring & Targeting

The role Sigma can play in your M&T

We are all aware of the versatile energy management technique Monitoring and Targeting (M&T), which is based on the principle of “you can’t manage what you don’t measure”. M&T enables you to simply and continually measure, understand and optimise your energy consumption.

The technique supports:

  1. The detection of avoidable energy waste or excessive energy use, for example caused by poor control or equipment issues
  2. Quantification of savings achieved, which can be calculated across all energy projects even with different driving factors to allow like-for-like comparison
  3. The identification of qualified issues to investigate
  4. Stakeholder engagement to provide effective feedback for staff awareness
  5. Data for the development of performance targets.

 
M&T is about understanding energy information by using data to show trends and anomalies in your energy consumption. It helps organisations identify signs of avoidable waste and make operational and cost saving efficiencies. It can also help you achieve increased resource efficiency, improved budgeting and reduce greenhouse gas (GHG) emissions.

M&T is particularly useful for measuring energy data against factors that can impact your consumption, for example the weather, building occupancy and production output. This allows you to draw meaningful conclusions about your consumption across your organisation and highlight periods of unusual usage.

The latest enhancements to Sigma provide an enriched and flexible M&T tool allowing users to explore data in granular detail. This enables the tracking of the expected performance of an energy estate against operational or budgeted usage and reduces the need to manually and slowly piece data together in spreadsheets.

An effective M&T scheme should not be time-consuming or complex. Sigma provides the tools to analyse and explore data and manages M&T activities at scale. This supports an organisation’s energy strategy and enables access to timely, relevant information on energy use, indicators for action needed and energy reports to support accountability.

Regression analysis

Regression analysis can be used to monitor and target energy usage providing insight into the relationship between two data sets. By exploring how the data sets interact we can determine the expected performance, for example if one increases, so does the other.

If the energy consumption is less than expected, it indicates good energy performance, however if it is greater than expected, it indicates poor energy performance. For example, when plotting outside temperatures against gas usage, we’d expect to see consumption increase as the temperature cools due to an increase in gas consumption (central heating). If the data fit is poor and not in-line with the expected performance, it indicates a poor level of control and a need for further investigation.

Sigma has a dynamic and interactive tool that helps simplify the regression analysis process, supporting the ability to model the relationship between different factors over configurable periods of time. The variable data sets can then be plotted against the consumption.

As mentioned, buildings can be influenced by varying factors including temperature, occupancy and opening hours. By building a relationship between these different influences, we can paint a more accurate picture of their impact on energy usage and provide a basis for other activities such as performance tracking and forecasting.

This example chart shows a relationship between electricity consumption and the daily occupancy for a 15-month period. The line that goes through the centre of the data points is known as the ‘line of best fit’ and highlights the expected performance based on the established relationship.

In this example the correlation between the data sets is good, a determination reinforced by the R2 value of 0.957. This value is known as the ‘coefficient of determination’ and is a statistical measure of how close the data points are to the regression line. Typically, a value of 0.9 or above represents a good correlation and indicates that the two data sets are related, i.e. a 90% correlation between the variation in consumption and the influencing data set.

Regression Correlation

Diagram 1.1

Sigma’s interactive tool allows data points to be selected and excluded, date ranges to be updated, and different regression lines to be created to model different scenarios. Control lines can be added based on a fixed value deviation, a percentage or the number of standard deviations from the expected performance.

CuSUM

Once the regression has been determined, it can then be extended to show a Cumulative Sum (CuSUM) representation of the data. This will highlight the step-change in performance over time and show periods where there is significant performance degradation (line goes up), or improvement (line goes down). This visual indication may otherwise be hidden in a large scatter of the data, like in the example above.

A CuSUM technique is a simple but powerful statistical method for highlighting small differences in energy efficiency performances over time. A typical CuSUM graph follows a trend, demonstrating the random fluctuation of energy consumption and should oscillate around zero (standard or expected consumption). This trend will continue until the pattern of consumption changes because of the implementation of an energy saving measure or degradation in energy efficiency. When looking at a CuSUM chart, a change in the direction of the line indicates events that have relevance to the energy consumption pattern.

The example chart here represents the above example. We can clearly see there was a significant change in performance from February 2019 (annotated in the chart below) which isn’t immediately obvious in the general correlation (see Diagram 1.1).

CuSUM Poor Performance

Diagram 1.2

Results like this could be investigated to assess the change in performance and fix any issues. The regression method would then be used to reassess the performance and validate any corrective action.

In this scenario, corrective action was taken to resolve the poor performance seen after the 1 February, by putting new measures in place from 1 April. After re-running the regression (from before the change in performance to after the resolution was implemented) we can see a new performance line for the date 1 January to 25 April with a good correlation.

Revised Regression Correlation

Diagram 1.3

And when looking at the CuSUM analysis, we can see a very different picture to before.

CuSUM Bad Good Performance

Diagram 1.4

The red line indicates the start of February where the poor behaviour started, the green line then clearly shows a step-change in performance after the corrective action was taken. It indicates that the changes achieved the desired effect in restoring a more typical and expected performance.

This technique should then be used on an ongoing basis in the continuous improvements life-cycle to help identify and resolve issues effectively.

As you can see, M&T can be of great value to an organisation in delivering an effective energy management programme. Sigma’s latest significant upgrade delivers impressive changes to transform your data into meaningful visualisations and narratives that enhance an M&T strategy. The new tool is quick and easy to use and delivers a variety of outputs that will boost stakeholder engagement, measure expected performance and act as an early warning system that can spot abnormal consumption trends.

We have a suite of additional changes planned for future releases to help with M&T related activities and effective management of energy data which you can see in our product roadmap. Exciting future changes include trend analysis and exception notification which will proactively notify you of abnormal events.

Be the first to know about our regular roadmap updates by subscribing to our newsletter.

Posted by TEAM on 30 April 2019
  • Wholesale Market

    • 2019 (19)
    • +2018 (17)
  • Blog & News Archive

    • 2019 (35)
    • +2018 (59)
    • +2017 (32)
    • +2016 (33)
    • +2015 (57)
    • +2014 (26)
    • +2013 (13)
    • +2012 (8)
    • +2010 (1)