LED Diode Fault Self-Diagnosis Method and System

Abstract

The present invention relates to the technical field of LED diode fault analysis, and discloses an LED diode fault self-diagnosis method and system, including collecting the environmental data of the LED diode, calculating the environmental factor through environmental analysis, correcting the environmental factor with reference to the standard environmental data, preprocessing the collected data parameters, receiving the corrected environmental factor, classifying the preprocessed data parameters, the LED diode detection model receiving the classified data parameters, outputting the real-time data parameters and predicted data of the LED diode, the LED diode model optimization unit performing loss optimization on the LED diode detection model through the test set, optimizing the LED diode detection model in real time, classifying the LED diode faults, and performing diode testing and adjustment through the classified faults and environmental factors, realizing the intelligence of LED diode fault self-diagnosis.

Description

Technical Field

[0001] The present invention relates to the technical field of LED diode fault analysis, and discloses an LED diode fault self-diagnosis method and system. Background Art

[0002] The development of LED diode fault self-diagnosis methods and systems is an ever-progressing field, aiming to improve the reliability of LED lamps, extend their service life, and reduce maintenance costs. With the popularization of LED technology and the expansion of its application scope, the demand for intelligent management and maintenance of LED lamps is also increasing. The following outlines the current development status and existing deficiencies of LED fault self-diagnosis methods and systems. Although the influence of the environment on LED diodes is sometimes considered, there is a lack of a comprehensive model, and it is impossible to classify LED diode faults according to environmental factors. Without predicting the data of the LED diode at the next moment, it is impossible to accurately judge the cause of the LED diode fault. Currently, expert experience is commonly used in the market to judge the faults of LED diodes, which lacks objectivity and practicality, and is prone to misjudgment and missed judgment due to large subjective factors.

[0003] For example, the Chinese patent application with the publication number CN118872381A discloses an LED lighting system, an automotive lighting system, and a method for detecting faults in the LED lighting system. The LED lighting system includes an LED lighting circuit and a fault detection circuit. The LED lighting circuit includes a first string of a first plurality of LEDs serially electrically coupled to a first inductor and a second string of a second plurality of LEDs serially electrically coupled to a second inductor. Each of the first string and the second string has an equal total forward voltage. The fault detection circuit is configured to detect faults in the LED lighting circuit based on the detection of the voltage difference between the first inductor and the second inductor.

[0004] Although the above patent is simple and easy to implement, it has obvious deficiencies in terms of comprehensiveness and reliability. Users need to manually check and handle faults, which increases the maintenance cost. It can only detect the existence of faults, cannot classify the fault types, lacks a closed-loop control mechanism, and cannot automatically adjust the working state of the LEDs. It does not consider the influence of environmental factors such as temperature and humidity. It mainly relies on the voltage difference of the inductor to detect faults. This method is relatively single and can only detect specific types of faults such as open circuits or short circuits, and cannot comprehensively cover other types of faults, and may produce misjudgment or missed detection in different environments. Summary of the Invention

[0005] The purpose of this part is to outline some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this part, as well as in the abstract and title of the present application, to avoid obscuring the purpose of this part, the abstract, and the title of the invention. However, such simplifications or omissions shall not be used to limit the scope of the present invention.

[0006] To solve the above technical problems, the main object of the present invention is to provide an LED diode fault self-diagnosis method and system. Among them, the LED diode fault self-diagnosis method includes:

[0007] S1. Collect the environmental data of the LED diode, calculate the environmental factor through environmental analysis, and correct the environmental factor with reference to the standard environmental data;

[0008] S2. Collect the data parameters of the LED diode, preprocess the collected data parameters, receive the corrected environmental factor, and classify the preprocessed data parameters;

[0009] S3. The LED diode detection model receives the classified data parameters, establishes a test set and a training set, and outputs the real-time data parameters and prediction data of the LED diode through the LED diode detection model;

[0010] S4. The LED diode model tuning unit tunes the loss of the LED diode detection model through the test set, and optimizes the LED diode detection model in real time;

[0011] S5. Output the real-time data parameters and prediction data of the LED diode through the LED diode detection model, and classify the LED diode faults;

[0012] S6. Perform diode testing through the classified faults and environmental factors. If the tuning test passes, adjust the LED diode through the adjustment unit. If the tuning test fails, give a fault alarm.

[0013] As a preferred solution of the LED diode fault self-diagnosis method of the present invention, among them:

[0014] The environmental data includes temperature data, humidity data, light data, power supply voltage data, working current data, and LED aging data;

[0015] Calculate the environmental factor through the environmental analysis model, and the calculation expression of the environmental analysis model is as follows:

[0016]

[0017] Among them, is the influence coefficient of the environment on the LED diode output by the comprehensive model.

[0018] Preprocess and clean the collected data parameters through data cleaning, data standardization, data transformation, and feature selection to prepare the data;

[0019] Data classification is used to classify the preprocessed data according to different environmental conditions;

[0020] Load the preprocessed and classified LED diode parameters from the classified CSV file, and randomly divide the data into a training set and a test set;

[0021] Train the LED diode detection model with the training set data, evaluate it on the test set, and use the trained model to predict new real-time data;

[0022] The convolutional neural network performs a convolution operation on the LED diode data parameters through the convolutional layer, inputs the maximum pooling layer for dimensionality reduction, the data after dimensionality reduction is input into the flattening layer, the multi-dimensional data is flattened into a one-dimensional vector, and the one-dimensional vector is input into the fully connected layer to perform high-level abstraction processing on the features of the LED diode.

[0023] As a preferred solution of the LED diode fault self-diagnosis method of the present invention, wherein:

[0024] The calculation expression of the convolutional layer is as follows:

[0025]

[0026] Wherein, is an element of the output feature map, is the classified data of the jth group of input LED diodes, b is the bias term, is the weight corresponding to the classified data of the jth group of LED diodes, j is the number of groups of classified data of LED diodes, k is the total number of groups of classified data of LED diodes, f() is the activation function, and i is the number of groups of feature maps;

[0027] The calculation expression of the maximum pooling layer is as follows:

[0028]

[0029] Wherein, is an element of the output feature map, is a local area of the input data, max() is the maximum value function, and s is the step size from the ith feature map to the next feature map;

[0030] The calculation expression of the fully connected layer is as follows:

[0031]

[0032] Wherein, y is the output LED diode vector, x is the input LED diode vector, W is the weight matrix, b is the bias vector, and f() is the activation function.

[0033] As a preferred solution of the LED diode fault self-diagnosis method of the present invention, wherein:

[0034] Tune the loss of the LED diode detection model through the test set and optimize the model in real time, including defining the loss function, selecting the optimization algorithm, training and evaluating the model, and updating the model parameters in real time;

[0035] Calculate the loss of the LED diode detection model by defining the loss function, and the calculation expression of the loss function is as follows:

[0036]

[0037] Wherein, is the loss value of the LED diode detection model output by the loss function, N is the number of samples, p is the number of the p-th LED diode sample, is the true value corresponding to the number of the p-th LED diode sample, is the predicted value corresponding to the number of the p-th LED diode sample;

[0038] The optimization algorithm is used to minimize the loss function. Calculate the adaptive learning rate by multiplying the output gradient and hyperparameters of the LED diode detection model, and then calculate the difference between the adaptive learning rate and the momentum of the real-time output predicted value of the LED diode detection model to update the LED diode detection model. The calculation expression is as follows:

[0039]

[0040] Wherein, is the first-order estimation matrix of the z-th group, is the second-order estimation matrix of the z-th group, is the first-order matrix hyperparameter, is the second-order matrix hyperparameter, is the output gradient value of the LED diode detection model, is the optimized first-order estimation matrix, is the optimized second-order estimation matrix, is a small constant used to prevent overfitting, is the parameter of the LED diode detection model at the h-th step, is the optimized model parameter, is the first-order estimation matrix of the (z-1)-th group, is the second-order estimation matrix of the (z-1)-th group.

[0041] As a preferred solution of the LED diode fault self-diagnosis method of the present invention, wherein:

[0042] Fault detection determines whether an LED diode has a fault by comparing the predicted value with the actual value, including setting a threshold. When the difference between the predicted value and the actual value exceeds this threshold, it is considered that the LED diode may have a fault. Calculate the difference between the predicted value and the actual value, and determine whether a fault has occurred based on the difference value;

[0043] Test the LED diode through the classified fault types and environmental factors to verify the accuracy of the fault types and environmental factors.

[0044] An LED diode fault self-diagnosis system, including:

[0045] An environmental factor module, including an environment recognition unit, an environmental factor unit, and an environmental test unit;

[0046] A data module, including a data acquisition unit, a data processing unit, and a data classification unit;

[0047] A detection module, including an LED diode detection model and an LED diode model tuning unit;

[0048] A fault recognition module, including a data receiving unit and a fault classification unit;

[0049] An adaptive module, including a test unit and an adjustment unit.

[0050] A preferred solution of the LED diode fault self-diagnosis system of the present invention, wherein:

[0051] The environment recognition unit is used to collect environmental data and identify the types of environmental influencing factors;

[0052] The environmental factor unit is used to calculate environmental factors including temperature factor, humidity factor, light intensity factor, power supply voltage factor, working current factor, and LED aging factor;

[0053] The environmental test unit is used to analyze the effects of temperature, humidity, light intensity, power supply voltage, working current, and aging on the brightness of the LED;

[0054] The data acquisition unit is used to collect the environmental temperature, environmental humidity, environmental light intensity, power supply voltage, working current, time, and brightness of the LED;

[0055] The data processing unit is used to preprocess and clean the collected data parameters and prepare the data through data cleaning, data standardization, data transformation, and feature selection;

[0056] The data classification unit is used to classify the preprocessed data according to different environmental conditions for subsequent analysis and modeling;

[0057] The LED diode detection model is used to output real-time monitoring parameters and prediction parameters of the LED diode;

[0058] The LED diode model tuning unit is used to update the LED diode detection model.

[0059] A preferred embodiment of the LED diode fault self-diagnosis system of the present invention, wherein:

[0060] The data receiving unit is used to receive real-time data and prediction data output by the LED diode detection model;

[0061] The fault classification unit is used to identify the received data and detect whether a fault occurs. If a fault occurs, the fault is classified;

[0062] The test unit is used to receive the fault classification data and perform LED diode tests according to the fault classification data and environmental factors;

[0063] The adjustment unit is used to adaptively adjust the performance parameters of the LED diode according to the test results.

[0064] Advantages of the present invention:

[0065] By setting environmental factor correction, the present invention can more accurately reflect the actual working conditions of the LED diode, avoid misjudgment caused by environmental changes, ensure that the LED diode can operate stably in different environments, improve the versatility and adaptability of the system, and can accurately predict the data parameters of the LED diode at the next moment by establishing an LED diode detection model, and detect in advance whether the LED diode fails.

[0066] By classifying the real-time data parameters and prediction data, different types and degrees of faults can be more accurately identified. By tuning the test and adjustment units, the system realizes closed-loop control, ensures that faults are detected and processed in a timely manner, reduces unnecessary maintenance and replacement, and reduces the maintenance cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts. Among them:

[0068] Figure 1 is a flowchart of the LED diode fault self-diagnosis method of the present invention;

[0069] Figure 2 is a composition diagram of the LED diode fault self-diagnosis system of the present invention. Detailed Implementation Modes

[0070] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following provides a detailed description of the specific implementation modes of the present invention in conjunction with the accompanying drawings of the specification.

[0071] In the following description, many specific details are set forth to facilitate a thorough understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art can make similar generalizations without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0072] Secondly, the so-called "one embodiment" or "embodiment" herein refers to specific features, structures, or characteristics that may be included in at least one implementation mode of the present invention. The phrase "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments.

[0073] Embodiment 1

[0074] As shown in Figure 1 , the LED diode fault self-diagnosis method includes:

[0075] S1. Collect the environmental data of the LED diode, calculate the environmental factor through environmental analysis, and correct the environmental factor with reference to the standard environmental data;

[0076] Among them, the environmental data includes temperature data, humidity data, light data, power supply voltage data, working current data, and LED aging data;

[0077] Calculate the environmental factor through the environmental analysis model;

[0078] The environmental factor includes temperature factor, humidity factor, light intensity factor, power supply voltage factor, working current factor, and LED aging factor;

[0079] The influence of temperature on the LED is very significant. It not only affects the brightness but also changes the color temperature and lifespan. The calculation expression of the temperature factor is as follows:

[0080]

[0081] Among them, is the brightness of the LED at temperature T, is the brightness of the LED at the reference temperature , is the maximum tolerable temperature of the LED, and n is an empirical coefficient representing the sensitivity of the influence of temperature on brightness;

[0082] As the temperature rises, the bandgap of semiconductor materials narrows, making it easier to excite electrons. However, it also leads to an increase in the recombination rate of carriers, thereby reducing the luminous efficiency.

[0083] Humidity mainly indirectly affects the brightness of LEDs by influencing the performance of LED packaging materials. High humidity may cause moisture absorption and corrosion of the packaging materials, thus reducing the luminous efficiency of LEDs. Humidity may affect the packaging materials of LEDs, resulting in corrosion or performance degradation. The humidity factor calculation expression is as follows:

[0084]

[0085] Where, is the humidity factor, is the humidity sensitivity coefficient, is the relative humidity, is the reference humidity;

[0086] The light intensity may cause the internal temperature of the LED to rise, thereby affecting its performance. The light intensity factor calculation expression is as follows:

[0087]

[0088] Where, is the incident light intensity, is the light sensitivity coefficient, and e is the exponential constant;

[0089] Fluctuations in the power supply voltage will affect the working current of the LED, thereby affecting the brightness. The power supply voltage factor calculation expression is as follows:

[0090]

[0091] Where, is the power supply voltage factor, is the voltage sensitivity coefficient, is the power supply voltage, is the reference voltage;

[0092] Changes in the working current directly affect the brightness and lifespan of the LED. The working current factor calculation expression is as follows:

[0093]

[0094] Where, is the working current factor, is the current sensitivity coefficient, I is the working current, is the reference current, and m is the empirical coefficient, indicating the non-linear degree of the influence of current on brightness;

[0095] Over time, the performance of LEDs gradually degrades. The calculation expression for the LED aging factor is as follows:

[0096]

[0097] Where, is the LED aging factor, λ is the aging rate coefficient, and t is the time;

[0098] The calculation expression for the environmental analysis model is as follows:

[0099]

[0100] Where, is the influence coefficient of the comprehensive model output environment on the LED diode;

[0101] This environmental analysis model describes the effects of temperature, humidity, light intensity, power supply voltage, operating current, and aging on the brightness of LEDs through a series of mathematical functions. Each function is established based on physical principles and experimental data. By combining these functions, a comprehensive and accurate set of environmental influence parameters on LED performance can be constructed.

[0102] To make the model more accurate, parameter calibration is required to adjust each parameter and correct the environmental factors;

[0103] The environmental factors consider the effects of various environmental factors on the performance of LEDs and can more accurately calculate the performance of LEDs under different conditions. By continuously optimizing and calibrating the model parameters, the calculation accuracy of the model can be further improved, providing strong support for the design and application of LEDs;

[0104] Model verification includes experimental verification and simulation verification;

[0105] Experimental verification measures the actual brightness of the LED under different environmental conditions, compares it with the model prediction value, and evaluates the accuracy of the model.

[0106] Simulation verification uses simulation software to simulate the performance of the LED under different environmental conditions and verifies the applicability and reliability of the model.

[0107] S2. Collect the data parameters of the LED diode, preprocess the collected data parameters, receive the corrected environmental factors, and classify the preprocessed data parameters;

[0108] Collect the data parameters of the LED diode through sensors, including:

[0109] The temperature sensor is used to measure the ambient temperature;

[0110] The humidity sensor is used to measure the ambient humidity;

[0111] The light sensor is used to measure the ambient light intensity;

[0112] The voltmeter is used to measure the power supply voltage;

[0113] The ammeter is used to measure the working current;

[0114] The timer is used to record time;

[0115] The brightness sensor is used to measure the brightness of the LED;

[0116] The collected data parameters are preprocessed, cleaned, and prepared for subsequent analysis and modeling through data cleaning, data normalization, data transformation, and feature selection.

[0117] Data cleaning includes removing outliers, filling in missing values, and denoising;

[0118] Removing outliers is used to identify and delete clearly incorrect data points, such as temperature or humidity values outside a reasonable range;

[0119] Filling in missing values is used to fill in missing data points using interpolation methods or other statistical methods;

[0120] Denoising is used to remove noise using filters (such as moving average filters);

[0121] Data normalization includes normalization and standardization;

[0122] Normalization is used to scale all numerical features to the same range, such as [0, 1] or [-1, 1].

[0123] Standardization is used to transform the data into a standard normal distribution with zero mean and unit variance.

[0124] Data transformation includes logarithmic transformation and polynomial transformation;

[0125] Logarithmic transformation is used for data with non-linear relationships, where logarithmic transformation can be used to linearize the relationship.

[0126] Polynomial transformation is used for complex non-linear relationships, where polynomial transformation can be used.

[0127] Feature selection includes correlation analysis and principal component analysis;

[0128] Correlation analysis is used to calculate the correlation between each feature and the brightness of the LED, and select features with high correlation;

[0129] Principal component analysis (PCA) is used for dimensionality reduction to extract the most important features;

[0130] Furthermore, the corrected environmental factors are received to correct the collected environmental factors according to the rated parameters of the LED diodes, so as to improve the accuracy and reliability of the data.

[0131] The corrected temperature is used to correct the temperature data according to the known temperature sensor error model.

[0132] The corrected humidity is used to correct the humidity data according to the known humidity sensor error model.

[0133] The corrected light intensity is used to correct the light intensity data according to the known light sensor error model.

[0134] The corrected voltage and current are used to correct the voltage and current data according to the known voltmeter and ammeter error models.

[0135] The sensor error model is determined by the rated specifications of the sensor;

[0136] Furthermore, data classification is used to classify the preprocessed data according to different environmental conditions for subsequent analysis and modeling.

[0137] By defining classification criteria, the classification concept of the LED diodes is set;

[0138] For example:

[0139] Temperature range: For example, low temperature (<20°C), medium temperature (20 - 30°C), high temperature (>30°C).

[0140] Humidity range: For example, low humidity (<50%), medium humidity (50 - 70%), high humidity (>70%).

[0141] Light intensity range: For example, low light (<100 lux), medium light (100 - 1000 lux), high light (>1000 lux).

[0142] Voltage range: For example, low voltage (<4.5V), medium voltage (4.5 - 5.5V), high voltage (>5.5V).

[0143] Current range: For example, low current (<20 mA), medium current (20 - 30 mA), high current (>30 mA).

[0144] Time range: For example, short term (<1 hour), medium term (1 - 10 hours), long term (>10 hours).

[0145] By using a programming language (such as Python) to write a script, the data is classified according to the defined classification criteria, and the classified data is stored in different files or database tables;

[0146] S3. The LED diode detection model receives the classified data parameters, establishes a test set and a training set, and outputs the real-time data parameters and predicted data of the LED diode through the LED diode detection model.

[0147] Load the preprocessed and classified LED diode parameters from the classified CSV file, randomly divide the data into a training set and a test set, and set the ratio to 80% for the training set and 20% for the test set.

[0148] Using a Convolutional Neural Network (CNN) to establish an LED diode detection model is an effective method. For the environmental data parameters of LED diodes, which are usually one-dimensional time series data, a one-dimensional convolutional neural network (1D CNN) can be used to process them.

[0149] Convert the training set into a data format suitable for input to the convolutional layer. For 1D CNN, the input data usually needs to be reshaped into the format of (number of samples, number of time steps, number of features). Establish a 1D CNN model through the keras library, train the model with the training set data, and evaluate it on the test set. Use the trained model to predict new real-time data.

[0150] Furthermore, the convolutional layer performs a convolution operation on the input data through a sliding window and extracts local features. The calculation expression of the convolutional layer is as follows:

[0151]

[0152] Among them, is an element of the output feature map, is the j-th group of classified data of the input LED diode, b is the bias term, is the weight corresponding to the j-th group of classified data of the LED diode, j is the number of groups of classified data of the LED diode, k is the total number of groups of classified data of the LED diode, f() is the activation function, and i is the number of groups of feature maps.

[0153] Furthermore, the convolutional layer performs a convolution calculation on the input LED data parameters by multiplying the input vector with the weight matrix and adding the bias term, and extracts the local feature values of the LED data parameters.

[0154] The maximum pooling layer selects the maximum value of the local features after convolution, reduces the data dimension, and retains the important features. The calculation expression of the maximum pooling layer is as follows:

[0155]

[0156] Among them, is an element of the output feature map, is a local area of the input data, max() is the maximum value function, and s is the stride from the i-th feature map to the next feature map;

[0157] The multi-dimensional data is flattened into a one-dimensional vector through a flattening layer, and the one-dimensional vector is input into a fully connected layer. The fully connected layer performs high-level abstraction processing on the features of the LED diode through linear combination and activation functions. The calculation expression of the fully connected layer is as follows:

[0158]

[0159] Among them, y is the output LED diode vector, x is the input LED diode vector, W is the weight matrix, b is the bias vector, and f() is the activation function;

[0160] The activation functions include linear activation functions and ReLU;

[0161] Through the LED diode detection model, a 1D CNN model is constructed to detect and predict the performance of the LED diode. This model uses convolutional layers to extract local features, reduces the data dimension through max pooling layers, and finally performs regression prediction through fully connected layers. The LED diode detection model can effectively process time series data, improve the accuracy and robustness of prediction, extract features in a timely manner according to the data under various environmental impacts, ensure that the LED diode detection is minimally affected by various environments, the data is more accurate, and can accurately predict the data of the LED diode at the next moment according to the current environment, ensuring that it can accurately detect whether the LED diode will fail at the next moment;

[0162] S4. The LED diode model tuning unit tunes the loss of the LED diode detection model through the test set and optimizes the LED diode detection model in real time;

[0163] The loss of the LED diode detection model is tuned through the test set, and the model is optimized in real time, including defining the loss function, selecting the optimization algorithm, training and evaluating the model, and updating the model parameters in real time.

[0164] Furthermore, the loss of the LED diode detection model is calculated by defining the loss function. The loss function (LossFunction) is used to measure the difference between the model prediction value and the true value. The calculation expression of the loss function is as follows:

[0165]

[0166] Among them, is the loss value of the LED diode detection model output by the loss function, N is the number of samples, p is the number of the p-th LED diode sample, is the true value corresponding to the number of the p-th LED diode sample, is the predicted value corresponding to the number of the p-th LED diode sample.

[0167] The optimization algorithm is used to minimize the loss function. The adaptive learning rate is calculated by multiplying the output gradient of the LED diode detection model by the hyperparameter, and then the difference between the adaptive learning rate and the momentum of the real-time output predicted value of the LED diode detection model is calculated to update the LED diode detection model. The calculation expression is as follows:

[0168]

[0169] Among them, is the first-order estimation matrix of the z-th group, is the second-order estimation matrix of the z-th group, is the hyperparameter of the first-order matrix, is the hyperparameter of the second-order matrix, is the output gradient value of the LED diode detection model, is the optimized first-order estimation matrix, is the optimized second-order estimation matrix, is a small constant used to prevent overfitting, is the parameter of the LED diode detection model at the h-th step, is the optimized model parameter, is the first-order estimation matrix of the (z - 1)-th group, is the second-order estimation matrix of the (z - 1)-th group;

[0170] The first-order estimation matrix and the second-order estimation matrix of the z-th group are calculated through the first-order estimation matrix and the second-order estimation matrix of the (z - 1)-th group, and the optimized model parameter is calculated through the first-order estimation matrix and the second-order estimation matrix of the z-th group;

[0171] Real-time optimization of the model means that during the operation of the model, the model parameters are continuously adjusted according to new data;

[0172] S5. Classify the LED diode faults by outputting the real-time data parameters and predicted data of the LED diode through the LED diode detection model;

[0173] Fault detection determines whether there is a fault in the LED diode by comparing the difference between the predicted value and the actual value, including setting a threshold. When the difference between the predicted value and the actual value exceeds the threshold, it is considered that the LED diode may have a fault. Calculate the difference between the predicted value and the actual value, and determine whether a fault has occurred according to the difference value;

[0174] If a failure occurs, further fault classification is carried out. By further analyzing the fault type, the faults are divided into different categories, and features helpful for classification are extracted, such as temperature, humidity, light intensity, voltage, current, etc. A classifier is selected according to the influence coefficient of the LED diode, and the classifier is trained using fault data with labels. The trained classifier is used to classify the detected faults, and real-time data parameters and predicted data are output through the trained LED diode detection model, and the faults of the LED diode are detected and classified. This method can effectively identify and classify the faults of the LED diode, improving the reliability and maintenance efficiency of the system.

[0175] S6. Perform diode testing through the classified faults and environmental factors. If the tuning test passes, the LED diode is adjusted through the adjustment unit; if the tuning test fails, a fault alarm is issued.

[0176] Test the LED diode through the classified fault type and environmental factors to verify the accuracy of the fault type and environmental factors, including loading the fault data from the classified data, setting the test conditions according to the fault type and environmental factors, testing the LED diode under the set test conditions, and recording the test results.

[0177] Furthermore, according to the test results, perform a tuning test on the LED diode to verify the performance after tuning, including selecting appropriate tuning parameters according to the fault type and environmental factors, testing the LED diode under the set tuning parameters, and recording the test results after tuning.

[0178] The adjustment unit is used to adjust the LED diode through the adjustment unit according to the tuning test results.

[0179] If the tuning test fails, trigger a fault alarm and record the fault information.

[0180] Through S6, the performance of the LED diode can be effectively managed and maintained, improving the reliability and stability of the system.

[0181] Embodiment 2

[0182] As Figure 2 shown, the LED diode fault self-diagnosis system includes:

[0183] The environmental factor module includes an environmental identification unit, an environmental factor unit, and an environmental test unit.

[0184] The environmental identification unit is used to collect environmental data and identify the types of environmental influencing factors.

[0185] The environmental factor unit is used to calculate environmental factors, including temperature factor, humidity factor, light intensity factor, power supply voltage factor, working current factor, and LED aging factor;

[0186] The environmental test unit is used to analyze the effects of temperature, humidity, light intensity, power supply voltage, working current, and aging on the brightness of the LED;

[0187] The data module includes a data acquisition unit, a data processing unit, and a data classification unit;

[0188] The data acquisition unit is used to collect the ambient temperature, ambient humidity, ambient light intensity, power supply voltage, working current, time, and the brightness of the LED diode;

[0189] The data processing unit is used to preprocess and clean the collected data parameters through data cleaning, data standardization, data transformation, and feature selection to prepare the data;

[0190] The data classification unit is used to classify the preprocessed data according to different environmental conditions for subsequent analysis and modeling;

[0191] The detection module includes an LED diode detection model and an LED diode model tuning unit;

[0192] The LED diode detection model is used to output real-time monitoring parameters and prediction parameters of the LED diode;

[0193] The LED diode model tuning unit is used to update the LED diode detection model;

[0194] The fault identification module includes a data receiving unit and a fault classification unit;

[0195] The data receiving unit is used to receive the real-time data and prediction data output by the LED diode detection model;

[0196] The fault classification unit is used to identify the received data and detect whether a fault occurs. If a fault occurs, the fault is classified;

[0197] The adaptive module includes a test unit and an adjustment unit;

[0198] The test unit is used to receive the fault classification data and perform LED diode tests based on the fault classification data and environmental factors;

[0199] The adjustment unit is used to adaptively adjust the performance parameters of the LED diode according to the test results.

[0200] Importantly, it should be noted that the construction and arrangement of the present application shown in multiple different exemplary embodiments are merely illustrative. Although only two embodiments are described in detail in this disclosure, those who refer to this disclosure should easily understand that many modifications are possible without substantially departing from the novel teachings and advantages of the subject matter described in this application. For example, changes in the dimensions, scales, structures, shapes and proportions of various elements, as well as parameter values (such as temperature, pressure, etc.), installation arrangements, use of materials, colors, orientations, etc. For example, an element shown as integrally formed may be composed of multiple parts or elements, the position of the element may be inverted or otherwise changed, and the nature, number or position of discrete elements may be altered or changed. Accordingly, all such modifications are intended to be included within the scope of the present invention. The order or sequence of any process or method steps may be changed or reordered according to alternative embodiments. Any "means-plus-function" clause is intended to cover the structures that perform the functions described herein, and not only structural equivalents but also equivalent structures. Other substitutions, modifications, changes and omissions may be made in the design, operating conditions and arrangement of the exemplary embodiments without departing from the scope of the present invention. Accordingly, the present invention is not limited to a particular embodiment, but extends to various modifications that still fall within the scope of the appended claims.

[0201] In addition, in order to provide a concise description of the exemplary embodiments, not all features of the actual embodiments may be described (i.e., those features that are not relevant to the currently contemplated best mode of carrying out the present invention or those features that are not relevant to the implementation of the present invention).

[0202] It should be understood that in the development of any actual implementation, as in any engineering or design project, numerous specific implementation decisions may be made. Such development efforts may be complex and time-consuming, but for those of ordinary skill in the art who benefit from this disclosure, the development efforts will be a routine task of design, manufacture and production without excessive experimentation.

[0203] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention may be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A self-diagnosis method for LED diode faults, characterized in that: include: S1. Collect the environmental data of the LED diode, calculate the environmental factors through environmental analysis, and calibrate the environmental factors by comparing with the standard environmental data; The environmental data includes temperature data, humidity data, illumination data, power supply voltage data, operating current data, and LED aging data; The environmental factors are calculated through the environmental analysis model. The calculation expression of the environmental analysis model is as follows: ； in, The comprehensive model outputs the influence coefficient of the environment on the LED diode. is the LED brightness at temperature T, T is the temperature, is the humidity sensitivity coefficient, is the relative humidity, is the reference humidity, is the reference temperature The LED brightness under is the light sensitivity coefficient, is the incident light intensity, e is the exponential constant, is the power supply voltage factor, is the current sensitivity coefficient, I is the working current, is the reference current, m is the empirical coefficient; S2, collecting LED diode data parameters, preprocessing the collected data parameters, receiving the corrected environmental factors, and classifying the preprocessed data parameters; S3, the LED diode detection model receives the classified data parameters, establishes a test set and a training set, and outputs the LED diode real-time data parameters and predicted data through the LED diode detection model; S4, the LED diode model tuning unit performs loss tuning on the LED diode detection model through the test set, and optimizes the LED diode detection model in real time; S5, outputting LED diode real-time data parameters and prediction data through the LED diode detection model, and classifying LED diode faults; S6. Perform diode test based on the classified faults and environmental factors. If the tuning test passes, adjust the LED diode through the adjustment unit. If the tuning test fails, a fault alarm is issued.

2. The LED diode fault self-diagnosis method according to claim 1, characterized in that: Preprocess and clean the collected data parameters and prepare the data through data cleaning, data standardization, data transformation and feature selection; Data classification is used to classify the preprocessed data according to different environmental conditions.

3. The LED diode fault self-diagnosis method according to claim 2, characterized in that: Load the preprocessed and classified LED diode parameters from the classified CSV file and randomly divide the data into training and test sets; The LED diode detection model is trained using the training set data and evaluated on the test set. The trained model is used to predict new real-time data. The convolutional neural network performs convolution operations on the LED diode data parameters through the convolution layer, inputs the maximum pooling layer for dimensionality reduction processing, and inputs the data after dimensionality reduction processing into the flattening layer to flatten the multi-dimensional data into a one-dimensional vector, and the one-dimensional vector is input into the fully connected layer to perform high-level abstract processing on the characteristics of the LED diode.

4. The LED diode fault self-diagnosis method according to claim 3, characterized in that: The convolutional layer calculation expression is as follows: ； in, is an element of the output feature map, is the input data of the jth group of LED diodes after classification, b is the bias term, is the weight corresponding to the data of the jth group of LED diode classification, j is the number of data groups after LED diode classification, k is the total number of data groups after LED diode classification, f() is the activation function, and i is the number of feature map groups; The maximum pooling layer calculation expression is as follows: ； in, is an element of the output feature map, is a local area of ​​the input data, max() is the maximum value function, and s is the step size from the i-th feature map to the next feature map; The calculation expression of the fully connected layer is as follows: ； Among them, y is the output LED diode vector, x is the input LED diode vector, W is the weight matrix, b is the bias vector, and f() is the activation function.

5. The LED diode fault self-diagnosis method according to claim 4, characterized in that: Performing loss tuning on the LED diode detection model through the test set and optimizing the model in real time, including defining a loss function, selecting an optimization algorithm, performing model training and evaluation, and updating model parameters in real time; The loss amount of the LED diode detection model is calculated by defining a loss function, and the loss function calculation expression is as follows: ； in, is the loss value of the LED diode detection model output by the loss function, N is the number of samples, p is the number of p-th LED diode samples, is the true value corresponding to the pth LED diode sample number, is the predicted value corresponding to the pth LED diode sample number; The optimization algorithm is used to minimize the loss function, calculate the adaptive learning rate by multiplying the output gradient of the LED diode detection model and the hyperparameter product, and then calculate the difference between the adaptive learning rate and the momentum of the real-time output prediction value of the LED diode detection model to update the LED diode detection model. The calculation expression is as follows: ； in, is the first-order estimation matrix of the zth group, is the second-order estimation matrix of the zth group, is a first-order matrix hyperparameter, is a second-order matrix hyperparameter, Output gradient value for LED diode detection model, To optimize the first-order estimation matrix, To optimize the second-order estimation matrix, is a small constant used to prevent overfitting. is the LED diode detection model parameter of the hth step, are the optimized model parameters, is the first-order estimation matrix of the z-1th group, is the second-order estimation matrix of the z-1th group.

6. The LED diode fault self-diagnosis method according to claim 5, characterized in that: Fault detection determines whether the LED diode is faulty by comparing the difference between the predicted value and the actual value, including setting a threshold value. When the difference between the predicted value and the actual value exceeds the threshold value, the LED diode is considered to be faulty, calculating the difference between the predicted value and the actual value, and determining whether a fault has occurred based on the difference value; The LED diode is tested according to the classified fault types and environmental factors to verify the accuracy of the fault types and environmental factors.

7. An LED diode fault self-diagnosis system, used to implement the LED diode fault self-diagnosis method according to any one of claims 1 to 6, characterized in that: include: Environmental factor module, including environmental identification unit, environmental factor unit, and environmental testing unit; Data module, including data acquisition unit, data processing unit and data classification unit; A detection module, including an LED diode detection model and an LED diode model tuning unit; A fault identification module, including a data receiving unit and a fault classification unit; The adaptive module includes a testing unit and an adjusting unit.

8. The LED diode fault self-diagnosis system according to claim 7, characterized in that: The environment identification unit is used to collect environmental data and identify the types of environmental influencing factors; The environmental factor unit is used to calculate environmental factors including temperature factor, humidity factor, light intensity factor, power supply voltage factor, operating current factor, and LED aging factor; The environmental testing unit is used to analyze the influence of temperature, humidity, light intensity, power supply voltage, operating current and aging effect on LED brightness; The data acquisition unit is used to collect the LED diode ambient temperature, ambient humidity, ambient light intensity, power supply voltage, operating current, time and LED brightness; The data processing unit is used to pre-process and clean the collected data parameters and prepare data through data cleaning, data standardization, data conversion and feature selection; The data classification unit is used to classify the preprocessed data according to different environmental conditions; The LED diode detection model is used to output real-time monitoring parameters and prediction parameters of the LED diode; The LED diode model tuning unit is used to update the LED diode detection model.

9. The LED diode fault self-diagnosis system according to claim 7, characterized in that: The data receiving unit is used to receive the real-time data and prediction data output by the LED diode detection model; The fault classification unit is used to identify the received data and detect whether a fault occurs, and if a fault occurs, classify the fault; The test unit is used to receive fault classification data and perform LED diode testing according to the fault classification data and environmental factors; The adjustment unit is used to adaptively adjust the performance parameters of the LED diode according to the test results.

Images