研究目的
Investigating the use of deep learning based PV generation and load forecasts to improve the economic optimization of nano-grid applications with PV and battery storage systems.
研究成果
The proposed method achieves a 99.5% optimality rate by using forecasted load and PV generation, demonstrating significant economic benefits for nano-grid applications with PV and battery storage systems. The study highlights the potential for further improvements by incorporating socio-economic factors and weather forecasts into the models.
研究不足
The study excludes the aging of the battery from the problem definition due to its nonlinear nature, which could affect long-term optimization results. Additionally, the optimization problem is linearized, which may not capture all nonlinear dynamics of the system.
1：Experimental Design and Method Selection:
The study utilizes deep learning based PV generation and load forecasts to optimize the economic operation of nano-grids. The methodology includes the use of long-short-term-memory (LSTM) for forecasting and mixed integer linear programming (MILP) for optimization.
2：Sample Selection and Data Sources:
The dataset is retrieved from European Network of Transmission System Operators for Electricity (ENTSO-E), specifically from Amprion GmbH, covering load and PV generation data from Germany.
3：List of Experimental Equipment and Materials:
The study considers a hypothetical nano-grid system with PV generation units, battery storage systems (BSS), and grid connection.
4：Experimental Procedures and Operational Workflow:
The process involves data cleaning, forecasting using LSTM models, optimization using MILP, and handling mismatches between forecasted and actual values.
5：Data Analysis Methods:
The analysis includes comparing forecasted and actual values, evaluating the economic benefits of the proposed method, and assessing the performance of different optimization strategies.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
