This study presents a wireless photovoltaic fault monitoring system integrating an STM32 microcontroller with an Improved Horned Lizard Optimization Algorithm(IHLOA)and a Multi-Layer Perceptron(MLP)neural network.The ...This study presents a wireless photovoltaic fault monitoring system integrating an STM32 microcontroller with an Improved Horned Lizard Optimization Algorithm(IHLOA)and a Multi-Layer Perceptron(MLP)neural network.The IHLOA algorithm introduces three key innovations:(1)chaotic initialization to enhance population diversity and global search capability,(2)adaptive random walk strategies to escape local optima,and(3)a cross-strategy mechanism to accelerate convergence and enhance fault detection accuracy and robustness.The system comprises both hardware and software components.The hardware includes sensors such as the BH1750 light intensity sensor,DS18B20 temperature sensor,and INA226 current and voltage sensor,all interfaced with the STM32F103C8T6 microcontroller and the ESP8266 module for wireless data transmission.The software,developed using QT Creator,incorporates an IHLOA-MLP model for fault diagnosis.The user-friendly interface facilitates intuitive monitoring and scalability for multiple systems.Experimental validation on a PV array demonstrates that the IHLOA-MLP model achieves a fault detection accuracy of 94.55%,which is 2.4%higher than the standard MLP,while reducing variance by 63.64%compared to the standard MLP.This highlights its accuracy and robustness.When compared to other optimization algorithms such as BKA-MLP(94.10%accuracy)and HLOA-MLP(94.00%accuracy),the IHLOA-MLP further reduces variance to 0.08,showcasing its superior performance.The system selects voltage as a feature vector to maintain circuit stability,avoiding efficiency impacts from series current sensors.This combined hardware and software approach further reduces false alarms to 0.1%through a consecutive-judgment mechanism,significantly enhancing practical reliability.This work provides a cost-effective and scalable solution for improving the stability and safety of PV systems in real-world applications.展开更多
Considering the noise problem of the acquisition signals frommechanical transmission systems,a novel denoising method is proposed that combines Variational Mode Decomposition(VMD)with wavelet thresholding.The key inno...Considering the noise problem of the acquisition signals frommechanical transmission systems,a novel denoising method is proposed that combines Variational Mode Decomposition(VMD)with wavelet thresholding.The key innovation of this method lies in the optimization of VMD parameters K and α using the improved Horned Lizard Optimization Algorithm(IHLOA).An inertia weight parameter is introduced into the random walk strategy of HLOA,and the related formula is improved.The acquisition signal can be adaptively decomposed into some Intrinsic Mode Functions(IMFs),and the high-noise IMFs are identified based on a correlation coefficient-variance method.Further noise reduction is achieved using wavelet thresholding.The proposed method is validated using simulated signals and experimental signals,and simulation results indicate that the proposed method surpasses original VMD,Empirical Mode Decomposition(EMD),and wavelet thresholding in terms of Signal-to-Noise Ratio(SNR)and Root Mean Square Error(RMSE),and experimental results indicate that the proposedmethod can effectively remove noise in terms of three evaluationmetrics.Furthermore,comparedwith FeatureModeDecomposition(FMD)andMultichannel Singular Spectrum Analysis(MSSA),this method has a better envelope spectrum.This method not only provides a solution for noise reduction in signal processing but also holds significant potential for applications in structural health monitoring and fault diagnosis.展开更多
基金supported by the National Natural Science Foundation of China(12064027,12464010)2022 Jiangxi Province High-level and Highskilled Leading Talent Training Project Selected(No.63)+1 种基金Jiujiang"Xuncheng Talents"(No.JJXC2023032)Jiujiang Natural Science Foundation Project(Key Technologies Research on Autonomous Cruise Solar-Powered UAV-2025-1).
文摘This study presents a wireless photovoltaic fault monitoring system integrating an STM32 microcontroller with an Improved Horned Lizard Optimization Algorithm(IHLOA)and a Multi-Layer Perceptron(MLP)neural network.The IHLOA algorithm introduces three key innovations:(1)chaotic initialization to enhance population diversity and global search capability,(2)adaptive random walk strategies to escape local optima,and(3)a cross-strategy mechanism to accelerate convergence and enhance fault detection accuracy and robustness.The system comprises both hardware and software components.The hardware includes sensors such as the BH1750 light intensity sensor,DS18B20 temperature sensor,and INA226 current and voltage sensor,all interfaced with the STM32F103C8T6 microcontroller and the ESP8266 module for wireless data transmission.The software,developed using QT Creator,incorporates an IHLOA-MLP model for fault diagnosis.The user-friendly interface facilitates intuitive monitoring and scalability for multiple systems.Experimental validation on a PV array demonstrates that the IHLOA-MLP model achieves a fault detection accuracy of 94.55%,which is 2.4%higher than the standard MLP,while reducing variance by 63.64%compared to the standard MLP.This highlights its accuracy and robustness.When compared to other optimization algorithms such as BKA-MLP(94.10%accuracy)and HLOA-MLP(94.00%accuracy),the IHLOA-MLP further reduces variance to 0.08,showcasing its superior performance.The system selects voltage as a feature vector to maintain circuit stability,avoiding efficiency impacts from series current sensors.This combined hardware and software approach further reduces false alarms to 0.1%through a consecutive-judgment mechanism,significantly enhancing practical reliability.This work provides a cost-effective and scalable solution for improving the stability and safety of PV systems in real-world applications.
基金supported by Central Guidance on Local Science and Technology Development Fund of Hebei Province(Grant No.226Z1906G)funded by Science Research Project of Hebei Education Department(CXY2024038)+1 种基金funded by Basic Research Project of Shijiazhuang University in Hebei Province(241791157A)National Natural Science Foundation of China(52206224).
文摘Considering the noise problem of the acquisition signals frommechanical transmission systems,a novel denoising method is proposed that combines Variational Mode Decomposition(VMD)with wavelet thresholding.The key innovation of this method lies in the optimization of VMD parameters K and α using the improved Horned Lizard Optimization Algorithm(IHLOA).An inertia weight parameter is introduced into the random walk strategy of HLOA,and the related formula is improved.The acquisition signal can be adaptively decomposed into some Intrinsic Mode Functions(IMFs),and the high-noise IMFs are identified based on a correlation coefficient-variance method.Further noise reduction is achieved using wavelet thresholding.The proposed method is validated using simulated signals and experimental signals,and simulation results indicate that the proposed method surpasses original VMD,Empirical Mode Decomposition(EMD),and wavelet thresholding in terms of Signal-to-Noise Ratio(SNR)and Root Mean Square Error(RMSE),and experimental results indicate that the proposedmethod can effectively remove noise in terms of three evaluationmetrics.Furthermore,comparedwith FeatureModeDecomposition(FMD)andMultichannel Singular Spectrum Analysis(MSSA),this method has a better envelope spectrum.This method not only provides a solution for noise reduction in signal processing but also holds significant potential for applications in structural health monitoring and fault diagnosis.
