This widget is available under Site details. Normally, the energy consumption of a site follows a consistent pattern based on the usual weekly schedule of five workdays and two weekend days. The anomaly detections function enables users to identify and analyze any deviations from the normal state that may be caused by hardware malfunction, human error, or wrong equipment configuration. The information is displayed on a widget in the Site details page and also reported via email. To learn how to configure the function, see Configuring anomaly detections.

It is important to note, when a historic value is overwritten for values older than one day, the displayed chart may no longer accurately reflect anomalies. This is because the anomaly detection typically runs nightly and only evaluates the last day’s data, storing results as detected. Therefore, any changes to historic values used in previous anomaly detections won’t be flagged in the chart.

If you use the widget to detect anomalies, the chart displays data in 15-minute intervals (bins). Each bin shows the expected or measured consumption at the end of the 15-minute period. However, as you zoom into the chart and the bins become more focused, the interpretation of anomalies may no longer be as obvious. This can look strange, especially if the previous bin's value was within the thresholds and the current bin's value is only slightly outside the thresholds, leading to the impression that most of the bin should not have been flagged as an anomaly.

Additionally, the accuracy of anomaly detection improved following a system update on June 14th, 2024. Rendering detections prior to this date less reliable. Consider these limitations when reviewing historical data and rely on post-update results for accurate anomaly identification.

Widget description

126543
  1. Anomaly configuration: Use this dropdown list to choose the configuration to be displayed in the widget. Use the arrows to choose the previous/next configuration.
  2. More: For more information, see More button.
  3. Navigation: Use the arrows to navigate to the previous/next detected anomaly.
  4. Anomaly diagram: This diagram displays the expected (black line) and actual (purple line) consumption on a time axis. Individual anomalies are displayed as vertical stripes with dark stripes indicating positive and light stripes negative deviations from the expected value. The light grey area around the expected consumption line represents the allowed tolerance. Select a point in time in the diagram to see the values.
  5. Filters: Use the filters to show/hide individual elements on the diagram.
  6. Time frame: Set the period to be displayed. The whole period is analyzed but only the period specified here is displayed in the diagram. The current display limit is one year.
    You can control it as follows:
    • Use the left and right sliders to adjust the displayed period.
    • Press the top blue ribbon and drag the frame left and right to apply the same period to the preceding/succeeding days.
    • Draw a rectangle inside or outside of the frame to select another period.