Power Generation

Power generation/power consumption forecasting and power plant anomaly detection

SmallTrain provides power generation/power consumption forecasting and power plant anomaly detection.

SmallTrain forecasts the amount of power generation and consumption to create an efficient energy cycle.
Forecasting supply and demand values ​​with SmallTrain makes it easier for you to formulate power sales plans for your power generation business.
In addition, the deviation between the predicted value and the measured value of the power generation amount leads to the discovery of defects in the power plant, which is also useful for the O & M business of the power plant.
SmallTrain’s power generation prediction is highly accurate with a mean square error rate of less than 1%, so it can be used for energy management EMS in the actual power generation business.

  • Case 1

    • Providing power generation forecasting that are useful for the solar power generation business with an accuracy of less than 1% mean square error rate
    • Contributing to increased revenue from power sales in the new power business

  • Case 2

    • Achieved a mean square error rate of 2% that exceeds the accuracy of other companies in forecasting air conditioning load heat
    • PoC requested by a major company

  • Case 3

    • Detect anomalies in photovoltaic power plants.
    • Contributing to increased power generation revenues in the electric power business by determining power plant panel failures, power conditioner failures, and defects, and reductions in power generation due to the effects of shadows from trees and weeds

  • Research

    • Power consumption prediction
    • Achieved a mean square error RMSE of 0.667
    • Go over the mean squared error RMSE 0.677 from previous studies of power consumption prediction by CNN (initial 2018)
    • Expected to spread in applications that require extremely high prediction accuracy