Load Data Valuation in Multi-Energy Systems: An End-to-End Approach

Yangze Zhou,Qingsong Wen,Jie Song,Xueyuan Cui,Yi Wang

2023-11-16

Abstract:Accurate load forecasting serves as the foundation for the flexible operation of multi-energy systems (MES). Multi-energy loads are tightly coupled and exhibit significant uncertainties. Many works focus on enhancing forecasting accuracy by leveraging cross-sector information. However, data owners may not be motivated to share their data unless it leads to substantial benefits. Ensuring a reasonable data valuation can encourage them to share their data willingly. This paper presents an end-to-end framework to quantify multi-energy load data value by integrating forecasting and decision processes. To address optimization problems with integer variables, a two-stage end-to-end model solution is proposed. Moreover, a profit allocation strategy based on contribution to cost savings is investigated to encourage data sharing in MES. The experimental results demonstrate a significant decrease in operation costs, suggesting that the proposed valuation approach more effectively extracts the inherent data value than traditional methods. According to the proposed incentive mechanism, all sectors can benefit from data sharing by improving forecasting accuracy or receiving economic compensation.

Systems and Control

What problem does this paper attempt to address?

### Problems the paper attempts to solve This paper aims to solve the problem of quantifying the value of load data in multi - energy systems (MES). Specifically: 1. **Improve the accuracy of load forecasting**: - Multi - energy loads are closely coupled and there is significant uncertainty, so cross - domain information needs to be utilized to improve the forecasting accuracy. - However, data owners are usually reluctant to share their data unless it can bring substantial benefits. 2. **Reasonably quantify the value of data**: - In order to encourage data sharing, it is necessary to reasonably quantify the value of data and fairly distribute the resulting additional profits. - Traditional methods usually handle the forecasting and decision - making processes separately, which may lead to the underestimation of the value of data. 3. **Propose an end - to - end framework**: - This paper proposes an end - to - end framework that integrates the forecasting and decision - making processes to more accurately quantify the value of multi - energy load data. - Through this integration, the intrinsic value of data can be more effectively extracted and the operating costs can be reduced. 4. **Design an incentive mechanism**: - A profit - distribution strategy based on the contribution degree is proposed to encourage data sharing in different fields. - The experimental results show that this method can significantly reduce the operating costs, and all fields can benefit from data sharing, either by improving the forecasting accuracy or obtaining economic compensation. ### Summary By proposing an end - to - end data - valuation framework, this paper solves the problem of quantifying the value of load data in multi - energy systems. This framework not only improves the accuracy of load forecasting, but also encourages data - owners' data - sharing behavior through a reasonable profit - distribution strategy, thus achieving multi - field collaborative cooperation and cost savings.

Load Data Valuation in Multi-Energy Systems: An End-to-End Approach

Integrated Demand Response for a Load Serving Entity in Multi-Energy Market Considering Network Constraints

Data Valuation for Decision-Making with Uncertainty in Energy Transactions: A Case of the Two-Settlement Market System

Multi-energy Users Valuation Analysis Based on Multi-Attribute Decision-Making Theories

Valuation of Load Forecasts in an Ensemble Model

Multi-objective Modeling of Price and Pollution in Large-Scale Energy Hubs with Load Management

A novel data-driven multi-energy load forecasting model

From Uncertainty Elimination to Profit Enhancement: Role of Data in Demand Response

Trading and Valuation of Day-Ahead Load Forecasts in an Ensemble Model

Monetizing Customer Load Data for an Energy Retailer: A Cooperative Game Approach

Cost-oriented load forecasting

Data Sharing in Energy Systems

Multi-level Peer-to-peer Collaborative Optimization of Smart Energy System Based on Big Data Analysis

A multi-energy load forecasting method based on complementary ensemble empirical model decomposition and composite evaluation factor reconstruction

Bilateral Electricity Trade Between Smart Grids and Green Datacenters: Pricing Models and Performance Evaluation

A novel short-term multi-energy load forecasting method for integrated energy system based on two-layer joint modal decomposition and dynamic optimal ensemble learning

Data-driven approaches to achieve energy transition: Statistical analysis of demand-side management and smart decision making in urban energy hub networks

Cloud Energy Storage in Multi Energy Systems: Optimal Scheduling and Profit-Sharing Approaches

Adaptive data decomposition based quantile-long-short-term memory probabilistic forecasting framework for power demand side management of energy system

Multi-Energy Coupling Load Forecasting in Integrated Energy System with Improved Variational Mode Decomposition-Temporal Convolutional Network-Bidirectional Long Short-Term Memory Model

A Data-Driven Machine Learning Approach for Consumer Modeling with Load Disaggregation