In the recent decade, multilevel inverters (MLI) are in favour of academia as well as industry for high power and medium voltage applications. MLI can synthesize switched waveforms with low harmonic distortions than two level converter. The MLI are commonly deployed in grid connected renewable energy sources (RES) to providing better reliability and reducing the harmonics. This paper uses 31 level MLI with reduced switch topology for minimizing Total Harmonic Distortion (THD). This inverter performance is completely tuned by SLDB controller. The SLDB controller is the combination of Deep Belief Network (DBN) and Sea Lion Optimization algorithm (SLnO). The procedure of DBN is optimized by SLnO which is utilized to handle the switching angle of MLI easily and also it generate low harmonic voltage and get better THD. The SLnO algorithm is very competitive because SLnO was conducted on 23 mathematical optimization problems for analyse the exploitation, exploration phases and suggested methods convergence behaviours. This is the most suitable reason for selecting the SLnO algorithm to optimize the DBN. MATLAB/Simulink platform is used to design the work and some parameters like inverter voltage, inverter current, grid voltage, grid current and the analysis of THD, power loss and efficiency are taken and which is compared with other existing algorithms. The proposed THD is 0.1%, efficiency is 99.2% and the power loss is 0.005 (pu), the proposed method gives better performance than existing works.

近十年来，多电平逆变器 (MLI) 在高功率和中压应用方面受到学术界和工业界的青睐。与两电平转换器相比，MLI 可以合成具有低谐波失真的开关波形。MLI 通常部署在并网可再生能源 (RES) 中，以提供更好的可靠性并减少谐波。本文使用具有简化开关拓扑的 31 级 MLI 来最小化总谐波失真 (THD)。此逆变器性能完全由 SLDB 控制器调整。SLDB 控制器是深度信念网络 (DBN) 和海狮优化算法 (SLnO) 的组合。DBN的程序由SLnO优化，用于轻松处理MLI的开关角，并产生低谐波电压并获得更好的THD。SLnO 算法非常有竞争力，因为 SLnO 是针对 23 个数学优化问题进行的，用于分析开发、探索阶段和建议的方法收敛行为。这是选择 SLnO 算法优化 DBN 的最合适的原因。采用MATLAB/Simulink平台进行工作设计，并对逆变器电压、逆变器电流、电网电压、电网电流等参数以及THD、功率损耗和效率进行了分析，并与其他现有算法进行了比较。所提出的 THD 为 0.1%，效率为 99.2%，功率损耗为 0.005 (pu)，所提出的方法比现有工作具有更好的性能。这是选择 SLnO 算法优化 DBN 的最合适的原因。采用MATLAB/Simulink平台进行工作设计，并对逆变器电压、逆变器电流、电网电压、电网电流等参数以及THD、功率损耗和效率进行了分析，并与其他现有算法进行了比较。所提出的 THD 为 0.1%，效率为 99.2%，功率损耗为 0.005 (pu)，所提出的方法比现有工作具有更好的性能。这是选择 SLnO 算法优化 DBN 的最合适的原因。采用MATLAB/Simulink平台进行工作设计，并对逆变器电压、逆变器电流、电网电压、电网电流等参数以及THD、功率损耗和效率进行了分析，并与其他现有算法进行了比较。所提出的 THD 为 0.1%，效率为 99.2%，功率损耗为 0.005 (pu)，所提出的方法比现有工作具有更好的性能。