Data analysis à la ISO 50006 - on the fly!

How would you recognize anomalous behaviour of a technical component? For example, because it is misconfigured or broken, or because the meter is faulty? In the deZem Software, this is easy to recognize: If consumption depends on external variables (e.g. outside temperature, output manufactured, flow, air pressure etc.), the target behavior of the component is determined on the basis of a statistical model and compared with reality. 

The modelling method is also known as “Multivariate Regression Analysis”. 

The energy management standards ISO 50001: 2018 and 50006 require the use of statistical modelling or similar methods to "normalize" energy performance indicators. This typically requires the energy manager to spend several days experimenting with Excel. With deZemVis, on the other hand, it is a matter of seconds!

The image shows a heat meter (red curve) that failed between January and March. On the basis of statistical modelling, the deZem software determined the “target curve” (pale corridor) as a function of the outside temperature. The real consumption could be modelled rather precisely! Check the quality of the modeling on the left and right of the diagram, when the thermal measurement is operational.

A final note: our modelling is an excellent tool to determine the general dependencies of a consumption behavior on the individual input variables.

Please contact us for a short live demo via a web meeting!