energia e serviços públicos

Detecção de anomalias de consumo

Decision Flow para detecção de anomalias no aquecimento residencial com base no consumo anormal ou em problemas relatados pelo cliente.

DecisionRules

David Škarka

Autor do modelo

This Decision Flow detects home heating anomalies based on abnormal consumption or customer-reported issues. It evaluates monthly consumption data against historical baselines and temperature adjustments to identify specific issues such as HVAC malfunctions, underheating, or tampering.

Solution Components:
The solution utilizes specific Decision Tables and global variables to process the input data:
  • Global Variables (Temperature Factors): A global variable set named temperatureFactorAndThreshold defines the tempAdjustmentFactor (set to $0.02$ or 2% per degree) and the thresholdDeviation (set to $0.15$ or 15%) used to normalize calculations.
  • Reference Numbers (Decision Table): A Decision Table named Reference Numbers functions as a lookup table. It accepts the month as an input and outputs the historical AvgTempBaseLine (Average Temperature) and AvgConsumptionBaseline (Average Consumption) for that specific month.
  • Outcome Decisions (Decision Table): A Decision Table named Outcome Decisions determines the final alert status. It evaluates the calculated temperature difference, percentage deviation, and customer flags to output an alert (Yes/No) and a specific nextAction (e.g., "Possible HVAC malfunction" or "No action required").
  • Consumption Anomaly Detection (Decision Flow) executes the Reference Numbers table to retrieve the baseline temperature and consumption for the given month, then calculates anomaly parameters, and then the Outcome Decisions table categorizes the anomaly.
Check iconA checkmark inside a circle signifying "yes"Minus iconA minus inside a circle signifying "no"PROS IconA plus symbol representing positive aspects or benefits.CONS IconA minus symbol representing negative aspects or drawbacks.

Mais modelos

Ver outros modelos