基于MSTSO算法的冷水机组负荷分配模型研究

为降低空调系统的运行能耗，优化冷水机组的负荷分配，首先提出了一种多策略改进的金枪鱼优化算法(MSTSO)，引入黄金正弦觅食机制和非线性惯性权重来加强算法对最优解的全局定位能力；通过蜜獾随机搜索策略赋予算法更强的性能以跳出局部最优。接着利用双向长短期记忆网络(BiLSTM)搭建能效预测模型并用MSTSO算法对其初始参数进行寻优从而获得最佳训练效果。最后进一步提出BiLSTM-MSTSO负荷分配模型，对多台冷水机组的负荷进行合理分配与优化。实验结果表明，优化后的BiLSTM预测模型拥有更高的预测精度，MSTSO算法相较其他智能优化算法可以减少更多的能耗并最大化提升冷水机组的运行效率。因此BiLSTM-MSTSO智能模型适用于多冷水机组的能耗预测与优化。

Study on Load Distribution of Chillers Model Based on Multi-strategy Tuna Swarm Algorithm

To reduce the operating energy consumption of air-conditioning systems and optimize the load distribution of chillers, a multi-strategy improved tuna swarm optimization algorithm (MSTSO) is first proposed, which introduces a golden sine foraging mechanism and non-linear inertia weights to enhance the algorithm''s ability to locate the optimal solution globally and uses a honey badger random search strategy to give the algorithm stronger performance to jump out of the local optimum. Then, using a bi-directional long short-term memory (BiLSTM) network to build an energy efficiency prediction model and use the MSTSO algorithm to optimize its initial parameters to obtain the best training results. Finally, the BiLSTM-MSTSO energy consumption optimization model is further proposed to reasonably allocate and optimize the load of multi-chillers. The experimental results show that the optimized BiLSTM prediction model has higher prediction accuracy and the MSTSO algorithm can reduce energy consumption and maximize the operating efficiency of chillers compared to other intelligent optimization algorithms. Therefore, the BiLSTM-MSTSO intelligent model can be used to predict and optimize the energy consumption of multi-chillers.