Mode-Specific Load Prediction for High-Variance Power Demand
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting electrical power consumption, particularly for HVAC systems and other high variance loads, are inefficient due to inaccurate peak and valley predictions, leading to inefficient electrical distribution and increased charges.
Innovation Solution
A hierarchical machine learning approach that classifies input data into different modes of operation, using a classifier to determine the current mode and a regression model to predict power consumption, improving accuracy by training separate models for each mode without requiring expert training or precise measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single predictive model is used for all operational modes, then the device complexity is reduced, but the prediction accuracy deteriorates due to high variance in power consumption across different modes
Solution Approach 1:
The patent segments the operational data into distinct modes of operation based on power consumption characteristics. Multiple predictive models are then trained separately for each mode, allowing each model to specialize in predicting power consumption for its specific operational context. This segmentation resolves the contradiction by improving prediction accuracy through mode-specific modeling while keeping individual models relatively simple.
Solution Approach 2:
The patent applies local quality by creating predictive models with properties tailored to specific operational modes. Each mode receives a model optimized for its particular characteristics rather than a generic model. This allows the system to achieve high prediction accuracy for each local operational context while maintaining manageable complexity through specialized rather than universal modeling.
2Measurement precision
If separate predictive models are trained for each mode of operation, then the prediction accuracy for high variance loads is improved, but the device complexity increases
Solution Approach 1:
The system segments the overall prediction task into multiple smaller sub-tasks, each handled by a dedicated model for a specific operational mode. This segmentation improves accuracy by allowing each model to focus on its specialized domain while managing complexity through division of labor across multiple simpler models rather than one complex model.
Solution Approach 2:
The patent changes the parameter of model specialization by training separate models with parameters optimized for each operational mode. This parameter change approach allows the system to achieve high prediction accuracy for high variance loads by adapting model parameters to specific modes, while the modular structure keeps overall system complexity manageable.
3Productivity
If conventional prediction methods are used for high variance loads, then the system operation is simplified, but the electrical distribution efficiency deteriorates due to inaccurate predictions
Solution Approach 1:
The patent segments the prediction system to handle high variance loads more effectively by creating mode-specific predictive models. This segmentation enables accurate prediction of power consumption patterns, which directly improves electrical distribution efficiency by allowing better load management and resource allocation, while the modular approach keeps implementation complexity manageable.
Solution Approach 2:
The system applies local quality by implementing prediction capabilities tailored to specific operational modes of high variance loads. This localized approach improves electrical distribution efficiency for each mode while maintaining overall system simplicity through the use of dedicated rather than universal prediction mechanisms.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method and system are provided for improving predictions of electrical power usage. In the method and system, load and/or environmental data is classified into data sets that correspond to different modes of operation of an electrical load. Different predictive models are also provided for each set of classified data. The predictive models may provide more efficient and/or more accurate predictions of power usage since each model is limited to a particular mode of operation.