综合能源网规划和运行的基础理论和方法研究

This project is concerned with the research on the key fundamental issues of integrated energy grid (Smart Grid) planning and operation, which consist of three parts: (1) Modeling of integrated energy grid, which includes modeling the dynamic behaviors of various energy devices as well as their connections; (2) development of high-dimensional multi-objective stochastic optimization algorithms and (3) development of decision making systems selected from the Pareto sets of the multi-objective optimization computation for determination of the final optimal solution of planning and operation of integrated energy grids. The developed algorithms and methods will be applied to plan a power network based integrated energy grid and investigate the economy and reliability of the integrated energy grid which could be achieved using distributed CHP and CCHP, heat storages and cool storages, to smooth peaks and valleys of power generation and loads, respectively..For modeling of an integrated energy grid, Individual Based Models (IBM) will be used to model a large number of components which have different dynamic modes in different nature such as linear, nonlinear, stochastic, hierarchical and hybrid characteristics etc. A generalized energy flow concept will be defined to develop the model of integrated energy grid. In IBMs, the local and global variables will be defined to investigate the dynamics of different areas. The topology of the integrated energy grid will be focused in the modeling research. In order to develop high-dimensional multi-objective stochastic optimization algorithms, the selection of independent objective functions will be studied and the computation capability optimization algorithms will also be investigated in particular for real-time optimization. In the study of stochastic optimization, the expectation and variance of stochastic objective functions will be optimized to ensure the more deterministic solutions . The selection of independent evidence for decision making will be investigated together with the study of multiple people and multiple attributes involved in the decision making process..Planning and operation problems of integrated energy grids will be investigated as case studies in this project. The planning problem will be focused on integrated utilization of various energies such as coal, wind power, solar energy, nuclear power, and hydropower for thermal, cool and electric loads. In this investigation, the thermal and cool storage units and their contributions to the economy and reliability of the whole energy grid will be considered. The operation problem of integrated energy grid will be studied to explore the potential of smooth operation of power grids using thermal and cool suppliers and storages and to show that the advantages of integrated utilization of various energies sources in an energy grid would have more economic and reliable benefits.

项目攻克以下关键问题：1）综合能源网建模，研究各种能源设备动态特性及其联接关系；2）高维多目标随机优化算法；3）开发决策系统从可行解集中确定问题最终解。算法用于求解基于电网的综合能源网规划和运行问题，探究含分布式热电联产、冷热电联供和冷热存储的综合能源网的经济可靠性，及电网发电和负荷各自的削峰填谷。.　提出基于个体的建模方法对具有不同动态特性的大量元件进行建模，并采用局部和全局变量以研究不同区域的动态特性。提出能量流的概念并着重对拓扑关系进行建模。优化算法研究包括选取相互独立的多目标函数、提高计算性能及同时优化各随机目标函数的期望方差。决策研究包括选取独立证据及考虑不同人员和多属性决策问题。.　规划问题研究采用含煤核风光水等多种能源提供冷热电负荷，及冷热存储单元和它们对整个能源网经济可靠性的影响。运行问题研究含冷热储联产的电网平滑运行的潜力开发。展示能源综合利用将带来更好的经济可靠性。

本项目攻克综合能源系统规划与运行的共性基础科学问题，基于本团队在个体为本建模、计算智能优化、决策系统三方面的领先成果，研究并推出大规模异质系统建模、优化、决策为一体的理论方法和计算方式，获得主要成果如下：.（1）采用生物学中的个体为本思想，首次建立了具有通用性、开放性的复杂大系统个体为本统一模型，其中个体五元体模型能够详细描述不同性质、层级、属性、动态特性的系统单元，解决了异构网络子系统模型迥异、应用迁移性差的问题；.（2）基于动物行为学中群体搜索思想，首次研究提出了基于降维搜索的大规模多目标随机优化高效计算方法，通过众多目标削减与目标-约束转换，显著降低问题求解复杂度，改进随机搜索策略解决非凸多目标随机优化的可计算性、收敛性与求解效率难题；.（3）基于证据推理方法，首次推出大系统多属性科学决策方法，实现多属性综合运算的透明化科学决策，解决了量化与非量化指标混合的综合能源网规划与运行方案决策问题；.（4）将大规模异质系统建模、优化、决策为一体的计算方法应用于十四个不同综合能源网规划与运行场景，并开发了建模灵活、仿真准确、性能鲁棒的大型综合能源系统生产模拟仿真平台。该平台集成了能源系统建模、复杂网络分析、生产模拟仿真等功能，为我国综合能源网规划与运行提供了迄今较完善的辅助分析工具。.与众多国内外同类项目不同，本项目聚焦共性基础科学问题，基础研究夯实，交叉学科性强，理论有深度，应用有广度，实施有平台，人才培养、项目拓展、社会交流成果丰富。