Electricity Demand Prediction Using Device-Specific Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional electricity demand prediction systems in industrial plants lack accuracy, especially when production schedules change or unexpected device actions occur, leading to significant differences between predicted and actual values, making it difficult to effectively adjust non-utility generation and contract electricity demand.
Innovation Solution
An electricity demand prediction system that collects device-specific electricity usage data, creates models based on past production schedules, and computes future demand using a combination of short-term and long-term prediction methods, allowing for precise adjustments to match changing conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If electricity demand prediction is performed on the basis of the amount of electricity used by a whole plant, then the prediction process is simple, but the prediction accuracy is low and cannot adapt to changes in production schedule or device actions
Solution Approach 1:
The patent segments the electricity demand prediction by dividing the plant into multiple devices and further dividing each device's prediction by product type. Instead of predicting total plant electricity usage as a single aggregate value, the system creates separate electricity amount calculation models for each device-product combination, then sums these segmented predictions to obtain the total predicted electricity demand. This segmentation enables the prediction to adapt to changes in specific devices or product mixes while maintaining manageable model complexity.
2Measurement precision
If the accuracy of electricity demand prediction is increased by considering device-specific data and production schedules, then the prediction adapts to changes, but the system complexity increases
Solution Approach 1:
The patent creates a universal electricity amount calculation model template that can be applied to any device-product combination. The model structure remains consistent across all devices, using the same mathematical framework that relates electricity consumption to production volume and product type characteristics. This universality allows the system to handle multiple devices and product types without proportionally increasing system complexity, as the same model logic is reused across different contexts.
Solution Approach 2:
The system manages complexity by changing parameters rather than structure. The electricity amount calculation model uses variable parameters such as production volume, product type identifiers, and device-specific coefficients that can be adjusted without altering the fundamental model architecture. When production schedules change or new devices are added, the system updates parameter values rather than redesigning the prediction framework, thereby maintaining relatively low system complexity while achieving high prediction accuracy.
3Quantity of substance
If electricity demand is predicted using conventional whole-plant data, then the prediction covers total usage, but it cannot effectively reduce non-utility generation and contract electricity demand due to large differences between predicted and actual values
Solution Approach 1:
The system incorporates feedback mechanisms by using actual electricity usage data from each device and product type to continuously refine and update the electricity amount calculation models. The model creation process utilizes historical actual values to calibrate prediction coefficients, and the system can update models based on deviations between predicted and actual consumption patterns. This feedback loop significantly improves prediction reliability while maintaining comprehensive coverage of total electricity demand.
Data Source
AI summary
An electricity demand prediction system includes electricity usage data collection means, production schedule storage means, model creation means, and short-term electricity demand prediction means. The electricity usage data collection means collects electricity usage data of each device and causes electricity usage data storage means to store the collected electricity usage data. The model creation means creates an electricity amount calculation model of each prescribed product type for each device on the basis of the electricity usage data stored in the electricity usage data storage means and a past production schedule stored in the production schedule storage means. The short-term electricity demand prediction means computes future electricity demand for each device on the basis of the electricity amount calculation model created by the model creation means and a future production schedule stored in the production schedule storage means.