This is an epoch-making program that allows you to easily and lively check the heat and power load predictions.
<Features of power / heat load prediction system>
- Predict the energy load of the day with high accuracy by the cluster analysis method from the past time series data.
- Eliminate the phase lag caused by mathematical methods Can be widely applied to the energy field (electric power, cooling, heating, hot water supply, etc.)
- Compatible with all facilities that use electricity, such as buildings, commercial facilities, hospitals, department stores, factories, small and medium-sized factories, etc.
- Achieve highly accurate forecasts without weather information.
- Prediction starts on the 5th day without long-term accumulation of actual measurement data.
- Short-term forecast (30 to 60 minutes) and predicts the load pattern of the day from 1:00 am on the day, and realizes accurate long-term forecast while sequentially correcting.
- Achieves short-time prediction and actual measurement EEP error within 2 to 5%
* EEP is the ratio of the root mean square error square root of the predicted heat load value to the maximum load during the period expressed as a percentage.
<Details of functions and performance>
- The energy load of the next day of the heat source equipment can be predicted accurately by the Kalman filter, and the amount of heat storage required for the next day, the operation schedule of the equipment, and the operation evaluation value can be instantly displayed from the optimum operation data.
- Data is read every 10 minutes from the central monitoring, and except for the operation of the prediction start button, the load prediction, operation schedule, and operation evaluation value are automatically displayed, so there is no operation that burdens the operator.
- Load is predicted at 8 o’clock on the day, the optimum operation pattern is read, and the operation schedule and operation evaluation value are displayed.
- It takes about 1 minute from the start of the next day / same day forecast to the display of the operation evaluation value by simulation, which is extremely short.
- The operation schedule displays the operation schedule of the equipment in consideration of the leveling of the operation time of the equipment, maintenance, and failure.
<Novelty / Excellence>
- The manual correction function allows the operator to arbitrarily change the operation schedule, compare the operation evaluation values, and quantitatively evaluate whether or not the change is appropriate.
- We have succeeded in developing a kind of “expert system” method that can call up the optimum operation plan.
- Even if the load is changed or reset manually, the optimum operation plan is displayed instantly. This makes it easy to study various operation planning cases.
- All-day forecasting in the load forecasting program makes load forecasting for each building based on actual data for each building, so more accurate load forecasting can be achieved. (Power load prediction: 3% accuracy in EEP evaluation, heat load prediction: 5% accuracy)
“Visualization” of driving support
- After loading the load, the operation plan is visually presented with a single click operation.
- The operation schedule displays the calorific value load pattern and the flow rate load pattern.
- You can easily change the operation schedule of the device manually while looking at the screen.
- The operation schedule after the change is also displayed instantly.
- The learning effect of the operator is also great.
- You can visually recognize and respond to emergencies and emergency responses.
<Forecasting and driving support>
- Create a 24-hour heat load pattern for the next day using the Kalman filter from the weather data and actual data of the next day.
- The operation schedule of the optimum equipment is displayed, and at the same time, the numerical value for evaluating the operation of the next day is also displayed instantly.
- If the operator wants to change the operation schedule of the equipment, he / she can change it using the manual allocation function of the equipment, compare the operation evaluation values, and quantitatively evaluate which operation is the most suitable.
The forecast load on the previous day is affected by various factors such as weather conditions, building usage conditions (number of people enrolled, operating status of installed equipment), etc. The measured values on the day are time-series data that reflect all fluctuation factors in the environment and operating conditions up to the measurement time. Therefore, “From the actual load data from the forecast date to about one month ago to the time to start the forecast of the day, similar load patterns are extracted for 5 to 6 days by the cluster analysis method, and the load of the next 10 minutes ahead. And predict the load pattern throughout the day (24:00) by repeating this method.
In this way, a novel load prediction on the day is performed, and the deviation from the load pattern predicted the day before is confirmed. If it almost matches the predicted load of the previous day, the operation schedule will not be changed.
If there is a discrepancy, or if the operation guideline is changed in a hurry (“minimum operating cost” → “maximum power saving”), the operation schedule will be revised. The forecast is made all day every time the actual value is entered every 10 minutes, so you can check the forecast all day at any time.