绿色低碳发展共识推动了能源结构低碳化转型，电力系统中可再生能源渗透率不断上升，对灵活性资源的需求大幅提升。可再生能源制氢技术具有快速响应能力，可作为灵活性负荷提升电力系统调节能力。碱性电制氢技术凭借单机规模大、价格低的优势，是大规模可再生能源制氢的一种可行路径。然而碱性电制氢系统跟随可再生能源出力变负载运行时，制氢功率波动对工艺流程中传热传质过程造成扰动，温度、杂质等关键状态量频繁越限，危害系统安全运行。变负载运行对系统工艺控制提出了高要求，然而目前国内外相关优化控制研究极少，面向控制的系统动态模型存在不足。在此背景下，本文针对碱性电制氢系统变负载运行需求，开展系统级建模与优化控制研究。本文主要研究内容如下： 首先，从系统工艺流程模型的角度，本文针对温度和杂质比例两个关键状态量，建立系统级动态模型。所建立的模型具有时滞微分代数方程组的统一形式，能够准确预测碱性电制氢系统变负载运行时电解槽和关键辅机处的温度与杂质比例动态变化过程，为后续优化控制研究奠定模型基础。 其次，从主工艺装置电解槽模型的角度，本文建立等效电路模型，以求解特定工况下电解槽内部的电流分布和流量分布。传统集总模型存在计算得到的电氢转换效率偏高、无法衡量电解液分布均匀性的问题。本文通过分布式建模方法，计入旁路电流的影响，实现电氢转换效率的准确预测。 然后，针对变负载运行时电解槽温度频繁波动、温度越限损害隔膜的问题，提出基于前馈思想的新型温度控制器设计，实践了电流前馈PID温度控制器和基于线性变参数方法的模型预测控制器两种解决方案。通过及时或提前开启冷水阀门，降低负载突增时的温度超调，实现稳定的温度控制。 最后，针对低载时氢氧混合比接近燃爆点、系统灵活性受限的问题，提出拓宽可调功率区间的压力控制策略，并基于前述建立的杂质动态模型，实践了最优运行曲线跟踪压力控制器和模型预测控制器两种方案。通过在低负载时降压减少杂质流量，有效降低电制氢系统负荷下限，拓宽可调功率范围。 本文针对碱性电制氢系统变负载运行难题，系统性地提出了全工艺流程的建模和优化控制方法，有效提升系统安全性、灵活性和经济性，提升电制氢机组并网灵活调节能力，同时为含碱性电制氢场站的新能源电网的优化调度提供模型基础。
The consensus on green and low-carbon development has promoted the low-carbon transformation of the energy structure. In the power system, the demand for flexible resources has increased significantly with increasing penetration rate of renewable energy. Renewable-to-hydrogen technology has quick response ability, which can be used as flexible loads to improve the stability of the power system. Alkaline electrolysis is a feasible path for large-scale renewable-to-hydrogen production with the advantages of large scale and low price. However, when an alkaline electrolysis system tracks the fluctuating renewable output, the power fluctuation causes disturbances on the heat and mass transfer processes in the system, and the key state variables such as temperature and impurity will exceed the limits frequently, which will endanger the safe operation of the system. Load-varying operation puts forward high requirements for system control. However, in the literature, the optimization and control studies are limited, and the existing dynamic models are insufficient. Therefore, this thesis conducts a system-level modeling and optimal control study to meet the load-varying operation requirement of alkaline electrolysis systems. The primary contributions of this thesis are as follows: Firstly, in terms of the system model, this paper establishes a system-level dynamic model for two key state variables including the temperature and impurity ratio. The established model has a unified form of differential algebraic equations with time-delays, which can predict the dynamic of the stack and key auxiliaries during load-varying operation and be used for subsequent controller design. Secondly, in terms of the stack model, this paper establishes an equivalent circuit model and a flow network model to solve the current distribution and flow distribution inside the stack. The traditional lumped models overestimate the system efficiency and cannot evaluate the uniformity of electrolyte distribution. In this paper, a distributed modeling method is adopted which takes the shunt current into account and can predict the system efficiency more accurately. Thirdly, in order to solve the temperature fluctuation problem during load-varying operation, a new temperature controller design method based on the idea of feedforward is proposed. Using the established thermal dynamic model, two novel temperature controllers are realized including a current feedforward proportion-integration-differentiation temperature controller and a model predictive controller. The proposed approaches can open the cooling water valve in time or in advance to reduce the temperature overshoot during loading-varying operation, hence the temperature control performance is improved. Finally, to overcome the flexibility limitation at low loads caused by hydrogen impurity, a pressure control strategy is proposed to widen the adjustable loading range. Based on the established impurity model, two realizations including an optimal operation curve tracking controller and a model predictive controller are achieved. The proposed approaches can reduce the system pressure during low-loading periods, which lower the impurity flow rate and widen the adjustable loading range. This thesis proposes a systematical modeling and optimal control method aiming at the difficulties in the load-varying operation of alkaline electrolysis systems. The proposed methods can improve the safety, flexibility and economy of alkaline electrolysis systems, and provide models for the optimal dispatch of power grids with alkaline electrolysis systems.
