Contributed By: Alisdair McDougall, Commercial Product Manager, Schneider Electric Energy & Sustainability Services
Internal benchmarking, reporting, conducting analysis, and reporting to external third parties requires complete and accurate data. For energy and sustainability teams, managing gaps in data is a common headache and can be a timely, frustrating, and costly task.
For some time, EcoStruxure™ Resource Advisor, Schneider Electric’s enterprise-level energy management software platform, has allowed data to be marked as “estimated,” allowing clients to manage and report forecasted data alongside actual data. After observing user estimation procedures, we believed we could further assist. Users were exporting historical, “actual” data from Resource Advisor, calculating estimates, and uploading data into Resource Advisor, as estimated, for reporting. Teams ran estimates more frequently as focus shifted from annual reporting to quarterly and even monthly performance management reporting.
“From Bradesco’s perspective this new functionality has made the whole reporting process a lot more efficient, transparent and reliable. We are very pleased with the results.”
~ Diego Nunes, Ecoefficiency Leader, Bradesco
Schneider Electric experts worked with a selection of companies on a project to automate the entire gap-filling process. Four themes emerged from these conversations:
- Flexible: It is important to have a selection of estimation methods available
- Configurable: Clients need to be able to configure estimation routines down to a specific site and data stream
- Traceable: It needs to be clear what data has been estimated and how
- Replicable: Replicability of estimation calculations is sometimes just as important, if not more important, than accuracy for those undergoing third-party audits
Born out of a response to these needs, the team launched a new functionality within Resource Advisor to help companies improve data completeness by automating their manual gap-filling processes. Companies can now configure and run automated estimation routines within Resource Advisor. The system identifies all gaps and fills them with estimates, calculated per the preferred method of the specific company.
Resource Advisor users and internal subject matter experts widely apply four standard estimation methods. To push the boundaries beyond the current best practice, a Schneider Electric Data Scientist suggested a method that uses Gaussian Process Regression, a machine learning technique that we could use to model and predict cost and usage. Thus, in total, a company can select from five different estimation methods.
The applicability of certain methods is dependent on the amount of historical data available. Limiting selection to only one method would result in a significant number of gaps being unfilled, which is not ideal when the sole purpose of the project was to fill as many gaps in data as possible.
Rather than having to pick one method per routine, companies can rank estimation methods according to their preference. Resource Advisor will always use the most preferred method. If the data is not available, it will opt to the second method and so forth. Under this dynamic method, the system tries everything it can to fill data gaps, using methods and parameters that the company deems reasonable.
With the automated estimation functionality in Resource Advisor, companies can now automate what was previously a manual process. What used to take hours, if not days, now takes minutes. As of April 2020, almost 100 companies have successfully run more than 1,300 estimation routines.
“Before deploying Automated Gap-Filling, the Schneider Electric client team used to spend hours supporting this process for Bradesco. The automated process now takes minutes, freeing up my team to look for other ways we can support Bradesco meet its goals.”
~Beatriz Yabiku, Bradesco Client Manager, Schneider Electric
Contact us to learn more about how Resource Advisor can help support your energy and sustainability strategy.