ecpEnergyTwinAnalytics

Generates demo data (site/equip/point + historical data + weather station) and a model. The model was identified using historical data generated by etDemogen function. All records generated by etDemogen function have tag etDemogen tag (useful for demo data removal). Note that etDemogen function can be used without a valid Energy Twin license.

Parameters:

No parameters

Examples:

etDemogen()

This function fixes corrupted models by setting to default missing or invalid tags. Note that not all corrupted models can be fixed. The function does not change model parameters from model identification.

Parameters:

• models
  • list of models to be fixed, all models will be fixed if this parameter is omitted

Examples:

etFixModels(readAll(etModel and mySpecialTag))

Returns an Energy Twin model that is linked to the point provided as a function parameter.

Parameters:

• point
  • function will return model linked to this point

Examples:

etGetModelForPoint(read(point))

This function can be used as hisfunc. It returns the difference between measurement and prediction (residuals). Tag measurementPointRef on hisfunc point is required. This tag references point with measured data and linked Energy twin model.

Parameters:

Standard hisFunc paramteres

Examples:

adding the point tag

hisFunc : etHisFuncDifference

This function can be used as hisfunc. It returns Energy Twin model prediction. Tag measurementPointRef on hisfunc point is required. This tag references point with measured data and linked Energy twin model.

Parameters:

Standard hisFunc paramteres

Examples:

adding the point tag

hisFunc : etHisFuncPrediction

Rule-ready KPI function that calculates the difference between measured and predicted values.

Parameters:

Standard rule paramters

Examples:

Use the Rule app to crate this KPI.

Rule-ready KPI function that focuses on the difference between Saturday and Sunday energy consumption profiles. See the ET KPI sim. application for details.

Parameters:

Standard rule paramteres

Examples:

KPI can be easily created via the user interface in the ET KPI sim. application.

Rule-ready KPI function that focuses on avoidable energy during weekend setback. See the ET KPI sim. application for details.

Parameters:

Standard rule paramteres

Examples:

KPI can be easily created via the user interface in the ET KPI sim. application.

Rule-ready KPI function that focuses on setback savings potential. See the ET KPI sim. application for details.

Parameters:

Standard rule paramteres

Examples:

KPI can be easily created via the user interface in the ET KPI sim. application. KPI parameters can be interactively adjusted there before creating KPI.

Rule-ready KPI function that focuses on weather-dependent load. See the ET KPI sim. application for details. KPI can be easily created via the user interface in the ET KPI sim. application.

Parameters:

Standard rule paramteres

Examples:

KPI can be easily created via the user interface in the ET KPI sim. application. KPI parameters can be interactively adjusted there before creating KPI.

Rule-ready KPI function that focuses on weather-dependent load. See the ET KPI sim. application for details. KPI can be easily created via the user interface in the ET KPI sim. application.

Parameters:

Standard rule paramteres

Examples:

KPI can be easily created via the user interface in the ET KPI sim. application. KPI parameters can be interactively adjusted there before creating KPI.

This function will create new model.

Parameters:

• dis
  • Model name
• pointRef
  • Measurement point for training
• trainingPeriod
  • Training period
• opts
  • Dictionary with options
  • Options:
    • additionalTrainingPeriods
      • Additional training periods, list of spans
    • samplingPeriodPreferences
      • Data rollup period
      • Default: "auto"
      • Options:
        • "auto"
          • Auto period detection
        • "ud15min"
          • 15min
        • "ud60min"
          • 60min
        • "ud240min"
          • 240min
    • scheduleType
      • Schedule type used in training
      • Default: "occupiedWeekdays"
      • Options:
        • "defByScheduleRefTag"
          • This will use scheduleRef tag to read schedule, this requires scheduleRef tag
        • "occupiedWeekdays"
          • Occupied Monday-Friday 8am-4pm
        • "defByOccupiedPoint"
          • This will use toOccupied() on pointRef function to get schedule
        • "occupiedEveryDay"
          • Always Occupied 24/7
    • scheduleRef
      • Schedule Ref required when scheduleType is defByScheduleRefTag
    • allowNegativeValues
      • Determines if negative values should be filtered out before training
      • Default: false
      • Options: true, false
    • preprocessingFunc
      • Custom preprocessing function

Examples:

etModelCreate("my model 1", read(point)->id, 2019)
etModelCreate("my model 2", read(point)->id, 2019, {scheduleType:"occupiedEveryDay"})
etModelCreate("my model 3", read(point)->id, 2019, {allowNegativeValues:true, preprocessingFunc:"myPreprocessingFunc", additionalTrainingPeriods:[2020-01, 2020-03]})

Returns measurement of model with correct sampling interval. Can be used to hisJoin with etModelPredict.

This function calculates the prediction of the given model. Occupancy and weather data are loaded using weather corresponding to a given point and its schedule. Prediction is composed of mean value and std. Output grid columns are named accordingly "vPred" and "vStd".

Parameters:

• model
  • record of chosen model ( see function metGetModelFromPoint)
• span
  • prediction span
• opts
  • options
    • includeMeasurement: adds measurement ,eg: {includeMeasurement}
    • interval: based on value aggregates data, eg: {interval: 1day}

Examples:

etModelPredict(read(etModel),2019-01)

etModelPredict(read(etModel),2019-01, {includeMeasurement, interval: 1day})

This function calculates the prediction of the given model. Prediction is given separately for weather-dependent load and time-dependent load. Occupancy and weather data are loaded using weather corresponding to a given point and its schedule. Prediction is composed:

• alphaM
  • Time dependant load for the full model.
• alphaSM
  • Time dependant load for the simple model.
• betaS
  • Weather dependant load for the full model.
• betaSM
  • Weather dependant load for the simple model.
• alphaSTD
  • Time dependant STD for the full model.
• alphaSSTD
  • Time dependant STD for the simple model.
• betaSTD
  • Weather dependant STD for the full model.
• betaSSTD
  • Weather dependant STD for the simple model.

In the current version of EnergyTwin STD is modelled only based on time, thus betaSTD values are expected to be zero.

Parameters:

• model
  • record of chosen model ( see function metGetModelFromPoint)
• span
  • prediction span

Examples:

etModelPredictStacked(read(etModel),2019-01)

This function will add models to training queue.

Parameters:

• models
  • List of etModel ids or single etModel id

Examples:

etModelTrain([@modelId1, @modelId2, readAll(etModel)[3])
etModelTrain(@modelId)

Rule-ready spark function that detects a deviation of measured values from values predicted by a model. The threshold value is defined as an absolute value (same unit as a point).

Parameters:

Standard rule parameters. For customization use the standard Tuning framework (parameter threshold). Default parameters can be adjusted in the ET Spark Sim. application before creating a spark.

Examples:

Spark can be easily created via the user interface in the ET Spark Sim. application.

Rule-ready spark function that detects a deviation of measured values from values predicted by a model. The threshold value is compared with the integrated difference between measured values and predicted values.

Parameters:

Standard rule parameters. For customization use the standard Tuning framework (parameter threshold). Default parameters can be adjusted in the ET Spark Sim. application before creating a spark.

Examples:

Spark can be easily created via the user interface in the ET Spark Sim. application.

Rule-ready spark function that detects a deviation of measured values from values predicted by a model. The threshold value is defined as an relative value (percent).

Parameters:

Standard rule parameters. For customization use the standard Tuning framework (parameter threshold). Default parameters can be adjusted in the ET Spark Sim. application before creating a spark.

Examples:

Spark can be easily created via the user interface in the ET Spark Sim. application.

Rule-ready spark function that detects a deviation of measured values from values predicted by a model. This function make use of statistical properties. The threshold is defined as mutliple of standard deviation. For example, with threshold set to 3, the spark condition will be as follows: measured > predicted + 3 * standardDeviation.

Parameters:

Standard rule parameters. For customization use the standard Tuning framework (parameter threshold). Default parameters can be adjusted in the ET Spark Sim. application before creating a spark.

Examples:

Spark can be easily created via the user interface in the ET Spark Sim. application.