Optimal Collection Scheme of Private Data When Using Blockchain to Estimate Baseline Load in Demand Response

Abstract

When estimating baseline load in demand response (DR), the accuracy of the results can be greatly improved if private data of resource owners are available. However, this may bring the risk of privacy leakage. Therefore, mainstream approaches advocate against using private data for baseline load estimation. To this end, this paper proposes a blockchain-based baseline load estimation and supervision framework, making it possible to use private data for baseline load estimation. Then, three principles for private data collection are established by analyzing the problems caused by collecting excessive or insufficient data, that is, first, the data collected should help calculate the baseline load effectively. Second, the data collected should ensure that the cost of data falsification by resource owners is significantly higher than the rewards they would gain in DR. Third, the rewards should far exceed the costs resource owners pay for data collection, transmission, and storage. Furthermore, this paper converts the problem of finding an optimal data collection scheme into an optimization model. A particle swarm optimization (PSO) algorithm with nonlinear inertia weight is then used to solve this problem, determining the optimal types and frequencies of data collection. Finally, this paper analyzes a case study where a shopping mall participates in DR, and points out the optimal collection types and frequencies of private data for such a scenario. The proposed model also shows effectiveness and robustness through sensitivity analysis and robustness test.