为应对大规模新能源并网给电力系统调度运行带来的挑战，实现“十二五”期间污染物总量减排的目标，同时满足深入开展节能发电调度的客观需求，迫切要求用户侧参与发电调度的统一优化，要求将传统的发电计划延伸至用户侧，扩展为发用电计划，同时内嵌式考虑网损的影响，这无疑改变了当前电力系统调度运行方式。针对这一需求，本文深入研究了内嵌网损的日前发用电计划模式、模型与方法。本文首先提出了内嵌网损的日前发用电计划模式。以机制设计理论为指导，设计了日前发用电计划中引导用户理性报价的用户移峰成本申报与结算机制；设计了“三层次，两阶段”的日前发用电计划分解协调模式；提出了以“内嵌网损的建模方法、规范化的用户响应建模方法和节点响应函数、大规模电网日前发用电计划的分解协调模型与方法”为核心的理论研究框架。本文提出了在日前发用电计划中内嵌网损的建模方法。传统的发电计划编制方法通常采用历史运行方式下的网损系数，由此决策的发电计划缺乏精准性和公平性。本文将网损表达为线路电导与线路两端相角差平方的乘积，基于直流潮流分别建立了内嵌网损的日前发电计划和发用电计划模型，提出了基于二次约束规划的高效求解方法，应用表明该模型与方法计算精度高、鲁棒性好。本文提出了规范化的用户响应建模方法和节点响应函数。为提升发用电计划中用户响应模型的简洁性和规范性，本文借鉴机组组合问题中发电机的建模方式，提出了规范化的用户响应建模方法，通过引入0-1整数变量，精确表达用户的复杂用电约束，能够包容大部分类型用户的响应特性。在此基础上，构造了能够反映节点内部用户移峰组合效益的节点响应函数。本文提出了高效、高精度的日前发用电计划分解协调模型与方法。该方法基于节点等效移峰成本曲线，将系统峰荷时段的移峰需求合理分解至不同的节点。各节点进一步以节点移峰成本增加最小、同时兼顾对发电侧成本影响最小的方式优化决策其内部用户的用电方案。该分解协调模型在保证求解精度的前提下，极大地提升了计算效率。本文的研究工作契合了推进智能电网与电力市场的建设、深入开展节能发电调度和“十二五”期间污染物总量减排的需求。希望本文的工作能够为日前发用电计划一体优化的新型调度方式提供理论与方法支撑。

In response to the challenges brought on by the integration of large-scale of renewable energy, the realization of the pollution reduction objective during the 12th Five Year Plan period, and satisfying the requirement of energy-conservation based generation dispatch (ECGD), the demand side of the power system needs to participate in the optimization of generation dispatch. The conventional generation scheduling will extend to embody the demand side, and forms the generation and load scheduling. This undoubtly changes the operation of power systems remarkably. According to such requirements, the mode, model and method of Day-Ahead Network Loss embedded Generation and Load Scheduling (DA-NLGLS) are studied in this thesis.The mode of DA-NLGLS is proposed in this thesis. Using the mechanism design theory as a guide, the declaration and settlement mechanism of customer shifting costs are designed. This mechanism can realize customers’ rational bidding. A “three-layer, two-stage” decomposition and coordination mode is designed. The theoretical study framework consists of the DA-NLGLS modeling, the standardized customer response modeling and node response functions, the decomposition and coordination modeling and algorithm of large-scale DA-NLGLS models are proposed.The method of embedding network loss in day-ahead generation and load scheduling is proposed. Conventional generation scheduling utilizes the network loss factors calculated with historical operating conditions. Such scheduling result is lack of accuracy and fairness. In this thesis, the network loss for each line is expressed by multiplying the conductance and the square of phase angle differences across the line. The day-ahead network loss embedded generation scheduling model and day-ahead network loss embedded load scheduling model are constructed respectively based on DC power flow. An efficient algorithm based on quadratically constrained programming is proposed. The application test demonstrates that the proposed model and method have good performance in terms of computational accuracy and robustness.A standardized customer response modeling method and the node response function are proposed. To ensure the concise and standardized modeling of customer response in generation and load scheduling, we draw the lessons from the generating unit modeling method in the unit commitment problem. The proposed customer response modeling method can embody most of the customer consumption constraints and response characteristics by introducing the 0-1 integer variables. Based on the customer response modeling method, the node response function is proposed which can reflect benefits of the combination of customer load shifting within the node.An efficient and accurate decomposition and coordination model and algorithm of DA-NLGLS are proposed, which are based on nodal equivalent shifting cost curves. The total peak load shifting requirement of the system can be allocated to nodes reasonably. Each node then optimizes the shifting scheduling of customers within it, considering the minimum nodal shifting cost increase and the impact of customer load shifting on the cost of the generation side. The proposed method improves the computational efficiency significantly while acquiring the accurate result of the DA-NLGLS problem.The research of this thesis fits the needs of constructing the Smart Grid and electricity markets, implementing ECGD, and realizing the pollution reduction objective during the 12th Five Year Plan period in China. Hopefully, this thesis would support a new type of scheduling for the co-optimization of day-ahead generation and load.