| 摘要: |
| 【目的】采用优化的寻优算法进行高效PI控制参数寻优。【方法】首先采用复合梯度法和复合粒子群法进行单一渠池PI控制参数寻优，对寻优过程中存在的初值不稳定等问题进行了探讨和针对性优化，提出了一种适用于渠道控制的分步寻优算法；随后针对多渠池的多变量问题提出了随机多渠池多变量寻优算法与固定权重的多渠池多变量寻优算法。【结果】单渠池工况下，复合梯度法和复合粒子群法能较好地解决寻优初值不稳定及寻优效率低下的问题。在控制效果相当的情况下，2种算法较均匀网格法的计算效率均得到显著提升，最长寻优耗时仅为网格法的5%左右。复合梯度法所用平均寻优耗时仅仅为复合粒子群算法的35.76%，复合梯度法较复合粒子群算法更为高效。多渠池工况下，随机多渠池多变量寻优算法与固定权重的多渠池多变量寻优算法在显著提高寻优效率的同时保证了较好的控制性能。然而，随机多渠池多变量寻优的控制效果较固定权重的多渠池多变量寻优有较大的改善，随机法较固定权重法的流量偏差小90.48%、闸门开度偏差小77.57%，系统稳定时间小17.05%，水位偏差小50.42%。但随机法所付出的时间成本也较高，寻优耗时较固定权重法多63.19%。【结论】随机多渠池多变量寻优算法的寻优耗时更长但控制性能更佳，使用者可根据其对于控制性能或耗时的要求有针对性地选择。 |
| 关键词: 渠道控制；PID算法；参数寻优；网格法；梯度算法；粒子群算法；多渠池参数寻优 |
| DOI：10.13522/j.cnki.ggps.2019404 |
| 分类号: |
| 基金项目: |
|
| Optimizing Operation Parameters of Irrigation Channels Using Single Objective and Multivariable Simulations |
|
ZHANG Yumeng, GUAN Guanghua
|
|
State Key Laboratory of Water Resources and Hydropower Engineering Science, Wuhan University, Wuhan 430072, China
|
| Abstract: |
| 【Background】 Classical control methods such as the proportional integral derivative (PID) are widely used in engineering due to their simplicity and robustness. There are two methods in designing controller parameters. One is the traditional classical design methods, including construction of control model and combination of the optimal control theory to optimize the PID parameters; the other one is the trial-errors obtained from computer simulation that combined digital simulations and the optimization program to determine the optimal PID parameters. However, because classical controller design theory does not consider the limitations of real channels, nonlinearity and communication delay of the system, its performance is often very poor or even becomes unstable. Currently, for channel control system, most PI control parameters optimizations use the grid search method to tune the controller. Due to the characteristics of canal system, such as nonlinearity and coupling, the initial value of the trial calculation and the setting of the grid density are highly dependent on experience and the calculation could be tedious.【Objective】This paper intends to use the optimized optimization algorithm to carry out the efficient PI control parameter tuning vial simulation.【Method】The compound gradient algorithm and the compound particle swarm optimization algorithm were used to tune the PI control parameters of the channel with a single reach, and the instability induced by the initial value in the optimization process is discussed and optimized. We proposed a step-by-step search algorithm to optimize the channel controller.【Result】For channel with a single reach, the proposed composite gradient algorithm and the composite particle swarm optimization algorithm can resolve the unstable initial value and improve computation efficiency. When the control effect was similar, the two proposed algorithms greatly improved the computation efficiency by using as less as 5% of the time used by the grid search method. In general, the average time used by the composite gradient algorithm was 35.76% of that used by the composite particle swarm optimization (PSO). Random multi-variable optimization algorithm and the fixed-weight multi-variable optimization algorithm were proposed for multi-reach channels. Both methods significantly improved the optimization efficiency and guaranteed a better control performance, despite the former being superior to the later in the control performance. The flow deviation, gate opening derivation, system ability time and water level deviation of the random method was 90.48%, 77.57%, 17.05% and 50.42% respectively less those of the fixed-weight method, but it achieved these by using more than 63.19% executed time.【Conclusion】The algorithm proposed in this paper can be used to design new water delivery control systems and upgrade existing channel control systems. It has implications for modernizing and transforming water distribution systems in irrigation districts in China. |
| Key words: channel control; PID algorithm; parameter optimization; grid search method; gradient algorithm; particle swarm optimization; multi-channel pool parameter optimization |
