In this paper, an effective neuro-fuzzy controller (NFC) technique has been proposed to control the induction motor torque and flux. The NFC hybrid technique is the grouping of the neural network (NN) and fuzzy logic controller (FLC), which generated the target voltages with the corresponding input flux and torque. The novelty of the proposed hybrid technique is highly flexible in nonlinear loads, convenient user interface and logical intellectual and permitting for integrated controlling schemes. Here, the FLC generates the training dataset of the NN technique based on the logical rules. The generated dataset contains the information about the flux and torque deviation parameters and the corresponding reference voltage parameters. The NN has been trained based on training dataset and the testing time which produces the optimal reference voltage parameters depends on the variation of the torque and flux parameters. By using the output of the NFC technique, the space vector modulation (SVM) develops the appropriate control pulses to the five-level inverter and the inverter generates the output voltage signal to the induction motor. The proposed method is designed in the MATLAB/Simulink platform and the outputs are verified through the comparison analysis with the existing techniques.

中文翻译：

---

多电平逆变器与支持向量机的感应电机DTC神经模糊控制器

本文提出了一种有效的神经模糊控制器 (NFC) 技术来控制感应电机的转矩和磁通。NFC 混合技术是神经网络 (NN) 和模糊逻辑控制器 (FLC) 的组合，它产生具有相应输入磁通和转矩的目标电压。所提出的混合技术的新颖性在非线性负载、方便的用户界面和逻辑智能以及允许集成控制方案方面高度灵活。在这里，FLC 根据逻辑规则生成 NN 技术的训练数据集。生成的数据集包含有关磁通和转矩偏差参数以及相应的参考电压参数的信息。神经网络已根据训练数据集进行训练，产生最佳参考电压参数的测试时间取决于转矩和磁通参数的变化。利用 NFC 技术的输出，空间矢量调制 (SVM) 为五电平逆变器产生适当的控制脉冲，逆变器产生输出电压信号给感应电机。该方法是在MATLAB/Simulink平台上设计的，并通过与现有技术的比较分析验证了输出。