Adaptive Neuro-Fuzzy Based Control of Solar Photovoltaic Powered Modular Multilevel Inverter for Marine Water Pumping Applications

Abstract

This paper presents an Adaptive Neuro-Fuzzy Inference System (ANFIS) based control strategy for a solar photovoltaic (PV) powered Modular Multilevel Inverter (MMI) designed for marine water pumping applications. Conventional fuzzy logic controllers (FLCs) perform reasonably well in nonlinear renewable-energy systems; however, their fixed rule base and lack of adaptive capability limit their performance under rapidly changing solar and load conditions. The proposed ANFIS controller integrates neural learning and fuzzy inference to offer improved voltage regulation, reduced total harmonic distortion (THD), and superior transient response. A complete solar PV–MMI–induction motor system is modeled in MATLAB/Simulink, and comparative analyses are carried out between FLC and ANFIS. The results demonstrate a significant reduction in settling time and THD,confirming that the ANFIS controller enhances system stability, waveform quality, and pumping performance. This study establishes ANFIS as a robust control solution for renewable-energy-driven marine pumping systems

Introduction

The paper focuses on improving the performance of solar PV–fed marine water pumping systems by replacing a conventional Fuzzy Logic Controller (FLC) with an Adaptive Neuro-Fuzzy Inference System (ANFIS). Solar-powered pumping is well suited for marine and remote applications, but fluctuating solar irradiation and dynamic load conditions introduce instability, nonlinearity, and power quality issues in inverter-fed motor drives. Modular Multilevel Inverters (MMIs) are used to ensure high-quality AC output with low harmonic distortion, which is critical for smooth induction motor operation.

A detailed MATLAB/Simulink model is developed consisting of a PV array, DC–DC boost converter, DC-link capacitor, eleven-level MMI, single-phase induction motor, and centrifugal marine pump. The key contribution is the integration of an ANFIS controller, which combines neural network learning with fuzzy logic reasoning to adapt control parameters in real time. Unlike traditional FLCs with fixed rules, ANFIS dynamically responds to variations in irradiance, load, and system disturbances, ensuring improved voltage regulation, faster dynamic response, and reduced harmonics.

The ANFIS controller uses a structured 7×7 rule base with error and change-in-error as inputs to generate optimized PWM switching signals. Simulation results demonstrate clear performance improvements over FLC control. Motor settling time is reduced from about 0.8 seconds (FLC) to 0.3 seconds (ANFIS), and Total Harmonic Distortion (THD) decreases from 8.57% to 7.27%. Additionally, voltage and speed waveforms show enhanced stability with reduced oscillations.

Overall, the study confirms that ANFIS-based control significantly enhances dynamic response, power quality, and operational stability of PV-fed modular multilevel inverter systems, making it a robust and efficient solution for renewable-energy-based marine water pumping applications.

Conclusion

The research successfully demonstrates that implementing an Adaptive Neuro-Fuzzy Inference System (ANFIS) controller in a solar PV-fed Modular Multilevel Inverter (MMI) significantly enhances the dynamic and steady-state performance of an induction motor-driven marine water pumping system. By overcoming the limitations of the conventional fuzzy logic controller, the proposed ANFIS controller provides faster learning capability, improved adaptability, and more precise decision-making under fluctuating irradiation and load variations. Quantitatively, the motor’s settling time improves from 0.8 s to 0.3 s, reflecting a 62.5% enhancement in transient response, which directly results in smoother and more stable motor operation. In addition, the Total Harmonic Distortion (THD) is reduced from 8.57% to 7.27%, yielding a 15.1% improvement in output harmonic quality. This reduction contributes to a cleaner inverter waveform, reduced harmonic stress on the motor windings, and minimised torque ripple, thereby improving overall mechanical reliability. Although THD reduction does not drastically change efficiency, it supports a 1–2% improvement in overall system efficiency, combined with better voltage regulation and reduced power losses. These quantified enhancements validate that the proposed ANFIS-based control strategy is a reliable and high-performance solution for real-time solar-powered pumping applications. Its ability to maintain stable operation under dynamic conditions makes it highly suitable for marine environments, while also offering strong potential for agricultural and industrial pumping systems.

References

[1] H. Lan, Y. Bai, S. Wen, D. C. Yu, Y.-Y. Hong, J. Dai, and P. Cheng, ‘‘Modeling and stability analysis of hybrid PV/diesel/ESS in ship power system,’’ Inventions, vol. 1, no. 5, pp. 1–16, 2016, doi: 10.3390/inventions1010005. [2] S. G. Jayasinghe, L. Meegahapola, N. Fernando, Z. Jin, and J. M. Guerrero, ‘‘Review of ship microgrids: System architectures, storage technologies and power quality aspects,’’ Inventions, vol. 2, no. 4, pp. 1–19, 2017, doi: 10.3390/inventions2010004. [3] R. Kumar and B. Singh, ‘‘Single stage solar PV fed brushless DC motor driven water pump,’’ IEEE J. Emerg. Sel. Topics Power Electron., vol. 5, no. 3, pp. 1337–1385, Sep. 2017, doi: 10.1109/JESTPE.2017.2699918. [4] S. Shukla and B. Singh, ‘‘Single-stage PV array fed speed sensorless vector control of induction motor drive for water pumping,’’ IEEE Trans. Ind. Appl., vol. 54, no. 4, pp. 3575–3585, Jul./Aug. 2018, doi: 10.1109/TIA.2018.2810263. [5] C.-L. Su, W.-L. Chung, and K.-T. Yu, ‘‘An energy-savings evaluation method for variablefrequency-drive applications on ship central cooling systems,’’ IEEE Trans. Ind. Appl., vol. 50, no. 2, pp. 1286–1297, Mar./Apr. 2014, doi: 10.1109/TIA.2013.2271991.

Copyright

Copyright © 2025 Bodukuriwar Shiva Prasad, Sharma K. V., Naik M. T.. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.