Hierarchical Energy Forecasting via Smart Meter Aggregation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Accurate energy forecasting in hierarchical energy systems is challenging due to the complexity of hierarchical data organization and the time-consuming process of parameter estimation in mathematical models, especially with the integration of intermittent renewable energy sources.
Innovation Solution
A novel hierarchical forecasting approach that reuses forecast models from lower hierarchy levels to create efficient global forecast models, utilizing smart meters to decentralize forecasting calculations and a communication framework that enhances flexibility and efficiency, allowing for rapid and accurate energy demand and supply balancing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional parameter estimation using optimization algorithms is applied to forecast models, then forecast accuracy can be achieved, but the computational time and resources increase exponentially with the number of model parameters
Solution Approach 1:
The patent segments the hierarchical forecasting problem into multiple levels (individual, household, building, district, city). At each level, forecast models are created independently using local data, avoiding the need to estimate parameters for the entire hierarchical system at once. This segmentation reduces the exponential computational complexity by breaking it into manageable sub-problems that can be solved in parallel.
Solution Approach 2:
The patent performs preliminary actions by pre-calculating and storing forecast models at lower hierarchical levels (individual and household levels) before needing them for higher-level aggregation. These pre-computed models are then reused and combined to create forecasts at building, district, and city levels, eliminating the need to re-estimate parameters from scratch at each hierarchical level.
2Measurement precision
If forecast models are created for each level of the hierarchy independently, then local forecast accuracy is improved, but ensuring consistency across hierarchical levels becomes more complex
Solution Approach 1:
The patent merges forecast models from lower hierarchical levels to create models at higher levels. Specifically, individual-level forecasts are combined to form household-level forecasts, which are then merged to create building-level forecasts, and so on up the hierarchy. This merging process ensures consistency across levels because higher-level forecasts are derived from lower-level forecasts using aggregation rules, rather than being created independently.
Solution Approach 2:
The patent implements a nested structure where forecast models at each hierarchical level contain and utilize models from lower levels. The individual-level forecast models are nested within household-level models, which are nested within building-level models, and so forth. This nested arrangement ensures that forecasts at any level are consistent with forecasts at lower levels, as the higher-level models are built upon and incorporate the lower-level models.
3Measurement precision
If traditional centralized forecasting approaches are used for hierarchical energy systems, then comprehensive system-wide forecasts can be generated, but the computational burden and time required increase significantly
Solution Approach 1:
The patent segments the centralized forecasting problem into distributed forecast model creation at multiple hierarchical levels. Instead of processing all system-wide data through a single centralized model, the system creates and maintains separate forecast models at individual, household, building, district, and city levels. This segmentation enables parallel processing and reduces the computational burden on any single system component.
Solution Approach 2:
The patent performs preliminary actions by pre-calculating forecast models at lower hierarchical levels and storing them for reuse. These pre-computed models are then aggregated to create higher-level forecasts, eliminating the need to re-process all raw data from scratch at each forecasting request. This preliminary computation significantly speeds up system-wide forecast generation while maintaining accuracy.
Data Source
AI summary
Examples of energy forecasting in hierarchical energy systems are provided herein. A global forecast model instance for a hierarchical energy system can be determined through aggregation of energy forecast model data from individual energy smart meters. Energy forecast model data can include values for energy forecast model parameters used by the individual smart meters. The energy smart meters include measurement, forecasting, and calculation capabilities. The smart meters locally determine a forecast model instance used by the smart meter and provide corresponding information to higher levels in the energy system hierarchy. A global forecast model instance is determined based on the provided information.