ABSTRACT

This manuscript presents a hybrid method for smart energy management (EM) in grid connected microgrid (MG) system. The grid connected micro grid system consists of photovoltaic (PV), wind turbine (WT), micro turbine (MT), and battery. The proposed method is the combined execution of Radial Basis Function Neural Network (RBFNN) and Squirrel Search Algorithm (SSA), hence it is called RBFNN-SSA method. Here, the necessary load demand of grid-connected MG system is constantly monitored by AI strategy. SSA has developed the perfect combination of MG considering the forecast load demand. The major intention of the RBFNN-SSA method is fuel cost involvement, grid power hourly power variation, operation with maintenance cost of grid connected micro grid system. The constraints are the accessibility of renewable energy sources (RES), power requirement and state of charge (SoC) of storage elements. Batteries are used as an energy source, to stabilize and allow the renewable power system units for maintaining constant output power. The proposed method is activated in MATLAB/Simulink working site. Then, the efficiency is assessed with existing methods such as improved artificial bee colony (IABC), bacterial foraging optimizer and artificial neural network (BFOANN), ant lion optimizer (ALO), grasshopper optimization algorithm with particle swarm optimization including artificial neural network (GOAPSNN). The root mean square error (RMSE), mean absolute percentage error (MAPE), and mean bias error (MBE) of proposed and existing methods under 50 and 100 count of trails are also analyzed. The proposed technique achieves the RMSE is 9.3, MAPE is 4.2, and MBE is 2 for 50 number of trails. For 100 count of trails, the proposed method achieves the RMSE is 13.5, MAPE is 3.9 and MBE is 5.7. The mean, median and standard deviation of RBFNN-SSA method achieves 0.9681, 0.9062 and 0.1099. The elapsed time of RBFNN-SSA method attains 30.15 s.

摘要

本手稿提出了一种在并网微电网 (MG) 系统中进行智能能源管理 (EM) 的混合方法。并网微电网系统由光伏（PV）、风力涡轮机（WT）、微型涡轮机（MT）和电池组成。所提出的方法是径向基函数神经网络（RBFNN）和松鼠搜索算法（SSA）的组合执行，因此称为RBFNN-SSA方法。在这里，并网MG系统的必要负载需求由AI策略持续监控。SSA 开发了MG 的完美组合，考虑了预测的负载需求。RBFNN-SSA 方法的主要目的是燃料成本参与、电网功率每小时功率变化、并网微电网系统的运行和维护成本。限制因素是可再生能源 (RES) 的可及性，存储元件的功率要求和充电状态 (SoC)。电池用作能源，以稳定并允许可再生电力系统单元保持恒定的输出功率。所提出的方法在MATLAB/Simulink 工作站点中被激活。然后，使用改进的人工蜂群（IABC）、细菌觅食优化器和人工神经网络（BFOANN）、蚁狮优化器（ALO）、包括人工神经网络的粒子群优化的蚱蜢优化算法（GOAPSNN）等现有方法评估效率。 ）。还分析了在 50 和 100 条路径下提出的和现有方法的均方根误差 (RMSE)、平均绝对百分比误差 (MAPE) 和平均偏差误差 (MBE)。所提出的技术实现了 RMSE 为 9.3，MAPE 为 4.2，对于 50 条路径，MBE 为 2。对于 100 条路径，所提出的方法实现的 RMSE 为 13.5，MAPE 为 3.9，MBE 为 5.7。RBFNN-SSA 方法的均值、中位数和标准差分别达到 0.9681、0.9062 和 0.1099。RBFNN-SSA 方法的运行时间达到 30.15 秒。