Worawat Sa-Ngiamvibool
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Worawat Sa-Ngiamvibool has a total of 50 co-authors. Here are the top co-authors by number of shared publications:
- Sitthisak Audomsi
Faculty of Engineering, Mahasarakham University, Maha Sarakham 44150, Thailand
15 shared publications - Jagraphon Obma
Rajamangala University of Technology ISAN, Khonkaen Campus
8 shared publications - Supannika Wattana
Electrical and Computer Engineering Research Unit, Faculty of Engineering, Mahasarakham University, Maha Sarakham 44150, Thailand
8 shared publications - Buncha Wattana
Electrical and Computer Engineering Research Unit, Faculty of Engineering, Mahasarakham University, Maha Sarakham 44150, Thailand
6 shared publications - Kittipong Ardhan
Kalasin University
5 shared publications - Chonlatee Photong
Mahasarakham University
4 shared publications
Latest Publications
- Open Access
Comparative Scenario Analysis of Decarbonization Pathways Toward Carbon Neutrality for Guangdong’s Energy System Using the LEAP Model
Guangdong is China’s largest provincial electricity consumer (912.1 TWh in 2024). Using a Low Emissions Analysis Platform (LEAP) model calibrated to the 2024 provincial yearbook, this paper compares a Business-as-Usual (BAU) reference with three decarbonization pathways over 2024–2060: Policy-Instrument-Coupled (PIC), Deep-Electrification-and-Hydrogen (DEH) and Digital-economy-with-CCUS (DCN). From 862 Mt CO2 in 2024, BAU peaks at 1143 Mt in 2040, whereas the pathways fall to 139, 99 and 35 Mt of direct emissions in 2060. After BECCS removals, only DCN approaches neutrality (21 Mt net, comparable to a conservative provincial land-sink estimate). Forty-three control scenarios show that end-use efficiency and fuel switching deliver 85–91% of the 2060 reduction relative to BAU and the capacity mix 8–10%. In 10,000 correlated Monte Carlo draws over the tested ranges, the ranking DCN < DEH < PIC never changes. In an hourly check (three ERA5 weather years, stylized load), the pathways serve every hour with at most 17 GW of four-hour storage (20–32 GW with plan-level offshore wind and nuclear), whereas the BAU capacity path falls short from 2040. By 2060, 38–57% of the pathways’ must-run output is surplus, and electricity-supply cost per MWh rises in the reverse order of emission depth.
Open AccessComparative Scenario Analysis of Decarbonization Pathways Toward Carbon Neutrality for Guangdong’s Energy System Using the LEAP Model
Guangdong is China’s largest provincial electricity consumer (912.1 TWh in 2024). Using a Low Emissions Analysis Platform (LEAP) model calibrated to the 2024 provincial yearbook, this paper compares a Business-as-Usual (BAU) reference with three decarbonization pathways over 2024–2060: Policy-Instrument-Coupled (PIC), Deep-Electrification-and-Hydrogen (DEH) and Digital-economy-with-CCUS (DCN). From 862 Mt CO2 in 2024, BAU peaks at 1143 Mt in 2040, whereas the pathways fall to 139, 99 and 35 Mt of direct emissions in 2060. After BECCS removals, only DCN approaches neutrality (21 Mt net, comparable to a conservative provincial land-sink estimate). Forty-three control scenarios show that end-use efficiency and fuel switching deliver 85–91% of the 2060 reduction relative to BAU and the capacity mix 8–10%. In 10,000 correlated Monte Carlo draws over the tested ranges, the ranking DCN < DEH < PIC never changes. In an hourly check (three ERA5 weather years, stylized load), the pathways serve every hour with at most 17 GW of four-hour storage (20–32 GW with plan-level offshore wind and nuclear), whereas the BAU capacity path falls short from 2040. By 2060, 38–57% of the pathways’ must-run output is surplus, and electricity-supply cost per MWh rises in the reverse order of emission depth.
- Open Access
Pathways Towards Net-Zero Transport in Guangxi, China: A LEAP-Based Scenario Analysis of Electrification, Modal Shift, and Power-Sector Decarbonisation
Transport emissions are still rising in fast-motorising Chinese provinces. In Guangxi, the vehicle fleet grew 32% between 2020 and 2024, while the grid is already half non-fossil. A bottom-up model built on the Low Emissions Analysis Platform (LEAP) and calibrated to official 2024 statistics compares four pathways to 2060, with the grid emission factor calculated from each scenario’s generating mix. In 2024, the sector consumed 567 PJ and emitted 41.9 Mt CO2. Under business as usual, emissions peak at 48.0 Mt around 2036 and remain at 37.4 Mt in 2060; the three policy pathways peak by 2027 and reach 21.5, 16.8, and 11.7 Mt in 2060 (42–69% below business as usual), cutting cumulative emissions by 24–45%. An order-independent Shapley decomposition attributes 50% of the 2060 abatement to vehicle electrification, 36% to power-sector decarbonisation, and 11% to modal shift. The two leading levers are complements and jointly deliver 86% of it (Monte Carlo range 84–88%), but their relative size is not settled: if charging draws disproportionately on thermal generation, the grid’s share exceeds that of electrification. Even the deepest, net-zero-aligned pathway leaves 11.7 Mt that must be balanced outside the sector.
Open AccessPathways Towards Net-Zero Transport in Guangxi, China: A LEAP-Based Scenario Analysis of Electrification, Modal Shift, and Power-Sector Decarbonisation
Transport emissions are still rising in fast-motorising Chinese provinces. In Guangxi, the vehicle fleet grew 32% between 2020 and 2024, while the grid is already half non-fossil. A bottom-up model built on the Low Emissions Analysis Platform (LEAP) and calibrated to official 2024 statistics compares four pathways to 2060, with the grid emission factor calculated from each scenario’s generating mix. In 2024, the sector consumed 567 PJ and emitted 41.9 Mt CO2. Under business as usual, emissions peak at 48.0 Mt around 2036 and remain at 37.4 Mt in 2060; the three policy pathways peak by 2027 and reach 21.5, 16.8, and 11.7 Mt in 2060 (42–69% below business as usual), cutting cumulative emissions by 24–45%. An order-independent Shapley decomposition attributes 50% of the 2060 abatement to vehicle electrification, 36% to power-sector decarbonisation, and 11% to modal shift. The two leading levers are complements and jointly deliver 86% of it (Monte Carlo range 84–88%), but their relative size is not settled: if charging draws disproportionately on thermal generation, the grid’s share exceeds that of electrification. Even the deepest, net-zero-aligned pathway leaves 11.7 Mt that must be balanced outside the sector.
- Open Access
- 16 July 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (14) , 7143
Enhanced Speed Control of PMSM Using Sliding Mode Controller Optimized by Chess Optimization Algorithm
This paper introduces an enhanced speed control strategy for permanent magnet synchronous motors (PMSMs) through the application of sliding mode control (SMC) optimized by the Chess Optimization Algorithm (COA). Additionally, the proposed method is compared with other validated algorithms, such as Harris Hawk Optimization (HHO) and the Whale Optimization Algorithm (WOA). The SMC parameters are optimized using Integral of Time-weighted Absolute Error (ITAE) and Integral of Time-weighted Square Error (ITSE) as objectives function, and the controllers are validated in MATLAB/Simulink analyzing transient response, speed variation, and load disturbance conditions. The results show that COA-tuned SMC provides better performance, with the lowest overshoot of 0.1929% for ITAE and 3.4800% for ITSE, the fastest settling times of 0.00307 s for ITSE, and the lowest MAE and RMSE in most evaluated scenarios, offering the most balanced overall performance among the optimization algorithms tested. Comparison to PI and PID controllers demonstrates that SMC provides enhanced accuracy, robustness, and disturbance rejection. COA-SMC provides an efficient and stable solution for PMSM applications that demand quick and accurate speed regulation.
- 16 July 2026
- Published by MDPI AG in Applied Sciences
- Vol. 16 (14) , 7143
Open AccessEnhanced Speed Control of PMSM Using Sliding Mode Controller Optimized by Chess Optimization Algorithm
This paper introduces an enhanced speed control strategy for permanent magnet synchronous motors (PMSMs) through the application of sliding mode control (SMC) optimized by the Chess Optimization Algorithm (COA). Additionally, the proposed method is compared with other validated algorithms, such as Harris Hawk Optimization (HHO) and the Whale Optimization Algorithm (WOA). The SMC parameters are optimized using Integral of Time-weighted Absolute Error (ITAE) and Integral of Time-weighted Square Error (ITSE) as objectives function, and the controllers are validated in MATLAB/Simulink analyzing transient response, speed variation, and load disturbance conditions. The results show that COA-tuned SMC provides better performance, with the lowest overshoot of 0.1929% for ITAE and 3.4800% for ITSE, the fastest settling times of 0.00307 s for ITSE, and the lowest MAE and RMSE in most evaluated scenarios, offering the most balanced overall performance among the optimization algorithms tested. Comparison to PI and PID controllers demonstrates that SMC provides enhanced accuracy, robustness, and disturbance rejection. COA-SMC provides an efficient and stable solution for PMSM applications that demand quick and accurate speed regulation.
- Open Access
Integrated Decarbonization Pathways for Regional Power Systems: A LEAP–NEMO Analysis in Guangxi, China
Achieving deep decarbonization while maintaining power system operational feasibility represents a critical challenge in regions with a high penetration of renewable energy. In this study, we developed an integrated LEAP–NEMO modeling framework to evaluate long-term decarbonization pathways and power system flexibility in Guangxi Province, China, from 2022 to 2060. Four scenarios were designed, the Business-as-Usual Scenario (BAS), the Policy Strengthening Scenario (PSS), the Technology-Centered Scenario (TCS), and the Integrated Optimization Scenario (IOS), to assess the impacts of policy, technology, and system-level coordination. Our results show that under the IOS, renewable electricity reaches approximately 78% by 2050 and further increases to 88% by 2060, while coal consumption declines by over 60% compared with the baseline. Energy storage deployment under the IOS reaches 25 GW by 2060 (rising to approximately 28 GW under a higher-resolution 48-representative-day validation), approximately five times the level in the BAS (5 GW), significantly enhancing system flexibility and reducing renewable curtailment from 10.2% to 1.8%. Carbon emissions peak in 2027 at 228 Mt CO2, approximately 20% lower than the baseline, and decline to 52 Mt CO2 by 2060, achieving a cumulative reduction of 3.85 Gt CO2. Economically, the IOS pathway reduces total system costs to USD 815 billion, saving approximately USD 95 billion compared with the BAS, with the levelized cost of electricity declining to USD 63/MWh by 2060. These results demonstrate that integrated optimization of generation and storage resources is essential in achieving deep decarbonization while maintaining system deterministic capacity adequacy and economic efficiency.
Open AccessIntegrated Decarbonization Pathways for Regional Power Systems: A LEAP–NEMO Analysis in Guangxi, China
Achieving deep decarbonization while maintaining power system operational feasibility represents a critical challenge in regions with a high penetration of renewable energy. In this study, we developed an integrated LEAP–NEMO modeling framework to evaluate long-term decarbonization pathways and power system flexibility in Guangxi Province, China, from 2022 to 2060. Four scenarios were designed, the Business-as-Usual Scenario (BAS), the Policy Strengthening Scenario (PSS), the Technology-Centered Scenario (TCS), and the Integrated Optimization Scenario (IOS), to assess the impacts of policy, technology, and system-level coordination. Our results show that under the IOS, renewable electricity reaches approximately 78% by 2050 and further increases to 88% by 2060, while coal consumption declines by over 60% compared with the baseline. Energy storage deployment under the IOS reaches 25 GW by 2060 (rising to approximately 28 GW under a higher-resolution 48-representative-day validation), approximately five times the level in the BAS (5 GW), significantly enhancing system flexibility and reducing renewable curtailment from 10.2% to 1.8%. Carbon emissions peak in 2027 at 228 Mt CO2, approximately 20% lower than the baseline, and decline to 52 Mt CO2 by 2060, achieving a cumulative reduction of 3.85 Gt CO2. Economically, the IOS pathway reduces total system costs to USD 815 billion, saving approximately USD 95 billion compared with the BAS, with the levelized cost of electricity declining to USD 63/MWh by 2060. These results demonstrate that integrated optimization of generation and storage resources is essential in achieving deep decarbonization while maintaining system deterministic capacity adequacy and economic efficiency.
- Open Access
- 6 June 2026
- Published by Elsevier in Results in Control and Optimization
- Vol. 23
Enhanced Ant Lion Optimizer for cascade (1+PI)-PIDF control in hydro-thermal power systems with capacitive storage
- 6 June 2026
- Published by Elsevier in Results in Control and Optimization
- Vol. 23
Open AccessEnhanced Ant Lion Optimizer for cascade (1+PI)-PIDF control in hydro-thermal power systems with capacitive storage
- Open Access
Short-Term Photovoltaic Power Forecasting Based on CNN–LSTM–Attention Mechanism
Accurate short-term photovoltaic (PV) power forecasting is crucial for maintaining grid stability. However, existing hybrid deep learning models suffer from inherent limitations owing to static feature-weighting mechanisms, often displaying significant phase lag and peak clipping under severe meteorological fluctuations. To address this, in this study, we adopt a hybrid CNN–LSTM–Attention forecasting framework incorporating an SE-based attention strategy. Field validation at a 150 kW PV power plant in Ningxia, China, demonstrated that the adopted model achieved a Root Mean Square Error (RMSE) convergence of 2.157 kW. Notably, this represented a 41.92% reduction in error compared to the standard LSTM benchmark and a further 16.46% improvement over the suboptimal CNN-LSTM baseline, explicitly confirming the specific contribution of the SE-based attention mechanism. Moreover, multi-weather evaluations and ablation studies confirm the framework’s robustness. Dynamic Time Warping (DTW) and Diebold–Mariano (DM) tests establish its statistical superiority and the reduction in phase lag against baselines. Residual analysis reveals a leptokurtic distribution with white noise properties, confirming the reduction in systematic bias. Consequently, this high-fidelity tracking allows precise minute-level ramping detection and decreases spinning reserve demands in practical power dispatch.
Open AccessShort-Term Photovoltaic Power Forecasting Based on CNN–LSTM–Attention Mechanism
Accurate short-term photovoltaic (PV) power forecasting is crucial for maintaining grid stability. However, existing hybrid deep learning models suffer from inherent limitations owing to static feature-weighting mechanisms, often displaying significant phase lag and peak clipping under severe meteorological fluctuations. To address this, in this study, we adopt a hybrid CNN–LSTM–Attention forecasting framework incorporating an SE-based attention strategy. Field validation at a 150 kW PV power plant in Ningxia, China, demonstrated that the adopted model achieved a Root Mean Square Error (RMSE) convergence of 2.157 kW. Notably, this represented a 41.92% reduction in error compared to the standard LSTM benchmark and a further 16.46% improvement over the suboptimal CNN-LSTM baseline, explicitly confirming the specific contribution of the SE-based attention mechanism. Moreover, multi-weather evaluations and ablation studies confirm the framework’s robustness. Dynamic Time Warping (DTW) and Diebold–Mariano (DM) tests establish its statistical superiority and the reduction in phase lag against baselines. Residual analysis reveals a leptokurtic distribution with white noise properties, confirming the reduction in systematic bias. Consequently, this high-fidelity tracking allows precise minute-level ramping detection and decreases spinning reserve demands in practical power dispatch.
- Open Access
- 13 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 25590-25603
Chess-Optimized Fuzzy-2DOF-PID Controller for Load Frequency Stabilization in Interconnected Multi-Area Renewable Grids
The Load Frequency Control (LFC) challenge in multi-source power systems increasingly complicates with the integration of renewable energy, owing to frequent nonlinearities, uncertainties, and disturbances. Conventional controllers such as proportional-integral-derivative (PID) and adaptive PI–1PD frequently exhibit constraints in terms of stability and convergence speed. This study proposes an optimal fuzzy logic two-degree-of-freedom PID controller (Optimal–FL–2DOF–PID) tuned via a metaheuristic chess optimizer (CO), which is a novel optimization algorithm introduced in this study. The performance was evaluated in MATLAB/Simulink under two scenarios: (i) a 5% Step Load Perturbation (SLP) and (ii) a Random Step Load Pattern (RSLP) increment-decrement that effectively mirrors real-world applications. The simulation results demonstrate that the Optimal–FL–2DOF–PID controller surpasses all the comparable controllers. In Scenario 1, the proposed controller attains the lowest objective function—Integral of Timeweighted Absolute Error (ITAE) of 4.6692 corresponding to reductions of 25.31%, 43.05%, and 41.92% relative to 2DOF–PID, PID, and PI–1PD, respectively (and 1.92% relative to the None–optimal–FL–2DOF– PID), overshoot (OS), undershoot (US), settling time (ST) 13–15 s, and tie-line power deviations are likewise reduced. In Scenario 2, using the corresponding optimum parameters from Scenario 1, the Optimal–FL– 2DOF–PID controller achieves ITAE = 395.72, which is lower than the None–optimal–FL–2DOF–PID controller design by 46.52%, and lower than 2DOF–PID, PID, and PI–1PD by 29.73%, 22.19%, and 9.68%, respectively. The frequency–channel ST is 150–180 s, whereas tie-line exchanges settled in the mid–150 s with visibly smoother profiles. These results indicate that Optimal–FL–2DOF–PID provides robust and efficient frequency regulation for multi-area systems under significant uncertainty, supporting deployment in future renewable energy-integrated power systems.
- 13 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 25590-25603
Open AccessChess-Optimized Fuzzy-2DOF-PID Controller for Load Frequency Stabilization in Interconnected Multi-Area Renewable Grids
The Load Frequency Control (LFC) challenge in multi-source power systems increasingly complicates with the integration of renewable energy, owing to frequent nonlinearities, uncertainties, and disturbances. Conventional controllers such as proportional-integral-derivative (PID) and adaptive PI–1PD frequently exhibit constraints in terms of stability and convergence speed. This study proposes an optimal fuzzy logic two-degree-of-freedom PID controller (Optimal–FL–2DOF–PID) tuned via a metaheuristic chess optimizer (CO), which is a novel optimization algorithm introduced in this study. The performance was evaluated in MATLAB/Simulink under two scenarios: (i) a 5% Step Load Perturbation (SLP) and (ii) a Random Step Load Pattern (RSLP) increment-decrement that effectively mirrors real-world applications. The simulation results demonstrate that the Optimal–FL–2DOF–PID controller surpasses all the comparable controllers. In Scenario 1, the proposed controller attains the lowest objective function—Integral of Timeweighted Absolute Error (ITAE) of 4.6692 corresponding to reductions of 25.31%, 43.05%, and 41.92% relative to 2DOF–PID, PID, and PI–1PD, respectively (and 1.92% relative to the None–optimal–FL–2DOF– PID), overshoot (OS), undershoot (US), settling time (ST) 13–15 s, and tie-line power deviations are likewise reduced. In Scenario 2, using the corresponding optimum parameters from Scenario 1, the Optimal–FL– 2DOF–PID controller achieves ITAE = 395.72, which is lower than the None–optimal–FL–2DOF–PID controller design by 46.52%, and lower than 2DOF–PID, PID, and PI–1PD by 29.73%, 22.19%, and 9.68%, respectively. The frequency–channel ST is 150–180 s, whereas tie-line exchanges settled in the mid–150 s with visibly smoother profiles. These results indicate that Optimal–FL–2DOF–PID provides robust and efficient frequency regulation for multi-area systems under significant uncertainty, supporting deployment in future renewable energy-integrated power systems.
- 1 January 2026
- Published by Engineered Science Publisher in Engineered Science
A State-Space Fuzzy Self-Tuning Proportional-Integral-Derivative Controller Optimized for Automatic Generation Control in Interconnected Hydro-Thermal Power Systems
Optimal design of a fuzzy logic-based proportional-integral-derivative controller for automatic generation control in a two-area hydro-thermal power system using a hybrid grey wolf optimizer–particle swarm optimization algorithm.
- 1 January 2026
- Published by Engineered Science Publisher in Engineered Science
A State-Space Fuzzy Self-Tuning Proportional-Integral-Derivative Controller Optimized for Automatic Generation Control in Interconnected Hydro-Thermal Power Systems
Optimal design of a fuzzy logic-based proportional-integral-derivative controller for automatic generation control in a two-area hydro-thermal power system using a hybrid grey wolf optimizer–particle swarm optimization algorithm.
- Open Access
- 8 December 2025
- Published by Engineering, Technology & Applied Science Research in Engineering, Technology & Applied Science Research
- Vol. 15 (6) , 28965-28971
Design of a Discontinuous SVPWM Strategy-Based Current Controller in DQ Synchronous Frame for the Grid-Connected 3L-T-Type Inverter
This study proposes an innovative discontinuous Space Vector Pulse Width Modulation (SVPWM) technique integrated with a DQ-based current control scheme for a three-level T-type (3L-T-Type) inverter running in grid-connected mode. The discontinuous modulation technique employs the Nearest Three Vectors (NTVs) from the space vector diagram to provide appropriate switching signals, which substantially decreases the commutation of the Controllable Power Switches (CPSs), thus reducing switching losses. To enhance the output power quality prior to grid injection, a transformation from standard DQ current control to active and reactive Power Control (PQ) is implemented. The proposed method allows for separate control of the d-axis and q-axis components while keeping them in sync with the grid voltage. A mathematical model was developed to compute the stay time of each switching state across different regions and sectors, facilitating the generation of discontinuous modulation waves for each phase leg. The simulation results in MATLAB/Simulink demonstrate improved performance with a 33% reduction in switching commutations, a Total Harmonic Distortion (THD) of 4.67% for the current, proper alignment with the grid voltage at the main frequency of 50 Hz, and effective balancing of the DC-link voltage. These findings confirm the proposed controller's efficacy and its appropriateness for high-performance grid-connected applications, including renewable energy integration and electric vehicle infrastructure.
- 8 December 2025
- Published by Engineering, Technology & Applied Science Research in Engineering, Technology & Applied Science Research
- Vol. 15 (6) , 28965-28971
Open AccessDesign of a Discontinuous SVPWM Strategy-Based Current Controller in DQ Synchronous Frame for the Grid-Connected 3L-T-Type Inverter
This study proposes an innovative discontinuous Space Vector Pulse Width Modulation (SVPWM) technique integrated with a DQ-based current control scheme for a three-level T-type (3L-T-Type) inverter running in grid-connected mode. The discontinuous modulation technique employs the Nearest Three Vectors (NTVs) from the space vector diagram to provide appropriate switching signals, which substantially decreases the commutation of the Controllable Power Switches (CPSs), thus reducing switching losses. To enhance the output power quality prior to grid injection, a transformation from standard DQ current control to active and reactive Power Control (PQ) is implemented. The proposed method allows for separate control of the d-axis and q-axis components while keeping them in sync with the grid voltage. A mathematical model was developed to compute the stay time of each switching state across different regions and sectors, facilitating the generation of discontinuous modulation waves for each phase leg. The simulation results in MATLAB/Simulink demonstrate improved performance with a 33% reduction in switching commutations, a Total Harmonic Distortion (THD) of 4.67% for the current, proper alignment with the grid voltage at the main frequency of 50 Hz, and effective balancing of the DC-link voltage. These findings confirm the proposed controller's efficacy and its appropriateness for high-performance grid-connected applications, including renewable energy integration and electric vehicle infrastructure.
- Open Access
- 8 December 2025
- Published by Engineering, Technology & Applied Science Research in Engineering, Technology & Applied Science Research
- Vol. 15 (6) , 28584-28590
Optimized Machine Learning for Induction Motor Fault Diagnosis Using Vibration and Frequency-Domain Features
This study uses vibration signal analysis to assess and contrast several machine learning models for identifying defects in induction motors. A comprehensive dataset obtained from TDMS-format sensor recordings included seven inter-turn short circuit fault conditions and one normal state. Three experimental settings were investigated: (i) multiclass classification with basic time-domain features, (ii) binary classification using enhanced time and Fast Fourier Transform (FFT) frequency-domain features, and (iii) optimal binary classification using hyperparameter tuning and advanced boosting models. KNN, Random Forest (RF), and ensemble models (XGBoost, LightGBM, CatBoost) were trained and evaluated using accuracy, MSE, RMSE, and R2. The results reveal that while raw time-domain features performed poorly in multiclass tasks (accuracy ~20%), significant gains were obtained utilizing FFT features and binary classification (accuracy up to 80%). Using hyperparameter tuning and gradient boosting techniques, additional enhancements drove the accuracy to 87%, with CatBoost and LightGBM excelling among others. These results highlight the importance of frequency-domain characteristics and model optimization in increasing fault detection dependability. In conclusion, this study helps to encourage the inclusion of intelligent monitoring systems into predictive maintenance pipelines, therefore prolonging the lifetime of industrial equipment and reducing operational downtime.
- 8 December 2025
- Published by Engineering, Technology & Applied Science Research in Engineering, Technology & Applied Science Research
- Vol. 15 (6) , 28584-28590
Open AccessOptimized Machine Learning for Induction Motor Fault Diagnosis Using Vibration and Frequency-Domain Features
This study uses vibration signal analysis to assess and contrast several machine learning models for identifying defects in induction motors. A comprehensive dataset obtained from TDMS-format sensor recordings included seven inter-turn short circuit fault conditions and one normal state. Three experimental settings were investigated: (i) multiclass classification with basic time-domain features, (ii) binary classification using enhanced time and Fast Fourier Transform (FFT) frequency-domain features, and (iii) optimal binary classification using hyperparameter tuning and advanced boosting models. KNN, Random Forest (RF), and ensemble models (XGBoost, LightGBM, CatBoost) were trained and evaluated using accuracy, MSE, RMSE, and R2. The results reveal that while raw time-domain features performed poorly in multiclass tasks (accuracy ~20%), significant gains were obtained utilizing FFT features and binary classification (accuracy up to 80%). Using hyperparameter tuning and gradient boosting techniques, additional enhancements drove the accuracy to 87%, with CatBoost and LightGBM excelling among others. These results highlight the importance of frequency-domain characteristics and model optimization in increasing fault detection dependability. In conclusion, this study helps to encourage the inclusion of intelligent monitoring systems into predictive maintenance pipelines, therefore prolonging the lifetime of industrial equipment and reducing operational downtime.