Deep learning-based silicon controlled rectifier dimming circuit zero crossing point detection method and system

Abstract

The invention provides a silicon controlled rectifier dimming circuit zero crossing point detection method and system based on deep learning, and relates to the technical field of circuit zero crossing point detection. The method comprises the steps that firstly, working condition information and voltage and current signals of a circuit are obtained, time and phase information of an initial zero crossing point are recognized through a pre-trained deep learning model (a circuit zero crossing point detection model), and training data of the model comprise a large number of working conditions and voltage and current samples with labels; after the silicon controlled rectifier circuit starts to work, the system detects the stroboscopic condition, and compares the actual stroboscopic value with the standard value to obtain the deviation degree. And finally, based on the initial zero crossing point information and the stroboscopic deviation value, calculating final zero crossing point information for adjusting and optimizing the working state of the circuit. The double verification mechanism obviously improves the accuracy and reliability of zero crossing point detection, so that the conduction opportunity of the silicon controlled rectifier is more accurate, a more stable dimming effect is realized, and the influence of a stroboflash phenomenon on the illumination quality is reduced.

Description

Technical Field

[0001] This application relates to the technical field of circuit zero-crossing detection, and particularly to a method and system for zero-crossing detection of a thyristor dimming circuit based on deep learning. Background Art

[0002] In a lighting system, due to the sinusoidal wave characteristic of alternating current, the voltage fluctuating above and below the zero-phase point will cause the lamps to produce stroboscopic phenomena. Such stroboscopies not only affect the human visual experience, but have a more significant impact on livestock. Livestock are far more sensitive to light source stroboscopies than humans. Stroboscopies can cause them to have abnormal behaviors such as panic and stress, seriously affecting their growth status and welfare level. Especially in places such as farms and livestock rooms, the lighting stroboscopy problem is directly related to the breeding efficiency. To reduce the influence of stroboscopies, thyristor dimming technology is widely used in lighting control. Its core principle is to adjust the output voltage by precisely controlling the conduction moment of the thyristor, and the control of the conduction moment completely depends on the accurate detection of the zero-crossing point of the alternating current power supply.

[0003] Currently, a hardware circuit is usually used to detect the zero-crossing point of alternating current. Specifically, a comparator is used to compare the voltage or current signal with the zero level. When the voltage or current value of the signal changes from positive to negative or from negative to positive and passes through the zero level, it is determined that the zero-crossing point has been detected. Then, based on the detected zero-crossing point, the conduction and cut-off times of the alternating current are controlled to adjust the brightness of the light, thereby alleviating the stroboscopic phenomenon to a certain extent.

[0004] However, the usage scenarios of alternating current lighting systems are becoming more and more complex. Various different electrical devices are connected to the power grid, the ambient temperature will also change continuously, and there are peak and off-peak periods for the load time. In such a situation, the zero-crossing point detection method based on hardware circuits such as comparators exposes disadvantages. For example, in a high-temperature environment, the component parameters in the hardware circuit will change, greatly reducing the accuracy of detecting the zero-crossing point. Moreover, during the peak electricity consumption period, a large number of electrical devices are running simultaneously, and the interference signals in the circuit increase. The comparator is easily affected by these interference signals and misjudges the zero-crossing point. In addition, this detection method can only detect the time point of the zero-crossing point and cannot obtain the phase information of the zero-crossing point. Therefore, it is difficult for traditional methods to achieve precise detection of the zero-phase point under dynamically changing conditions. Summary of the Invention

[0005] This application provides a method and system for zero-crossing detection of a thyristor dimming circuit based on deep learning, which is used to accurately detect the zero-crossing point in an alternating current circuit and solve the technical problems of low zero-crossing detection accuracy and poor anti-interference ability in the prior art.

[0006] In a first aspect, the present application provides a method for detecting the zero-crossing point of a thyristor dimming circuit based on deep learning, which is applied to a zero-crossing point detection system. The method includes: obtaining working condition information; obtaining the voltage signal and current signal of the thyristor dimming circuit; combining the working condition information, the voltage signal and the current signal, and determining the initial zero-crossing point information through a circuit zero-crossing point detection model. The initial zero-crossing point information at least includes the time point information and phase information of the zero-crossing point. The circuit zero-crossing point detection model is pre-constructed through deep learning according to multiple sets of working condition information, voltage signals and current signals with zero-crossing point annotation information; after starting the thyristor dimming circuit to work according to a set strategy, determining stroboscopic information; comparing the stroboscopic information with a preset stroboscopic standard to determine the stroboscopic deviation value; combining the initial zero-crossing point information, and determining the final zero-crossing point information according to the stroboscopic deviation value.

[0007] By adopting the above technical solution, the deep learning model is trained with a large amount of historical data with annotations, enabling it to accurately identify the zero-crossing point characteristics under various working conditions. Combining the working condition information, voltage and current signals obtained in real time, the model can quickly output the initial zero-crossing point information. On this basis, the system dynamically corrects the initial zero-crossing point information through real-time monitoring of stroboscopic information and deviation value analysis to form the final zero-crossing point information. This dual verification mechanism significantly improves the accuracy and reliability of zero-crossing point detection, effectively reduces the detection error, makes the conduction timing of the thyristor more accurate, thus achieving a more stable dimming effect and reducing the impact of stroboscopic phenomena on lighting quality.

[0008] Combined with some embodiments of the first aspect, in some embodiments, in the step of obtaining the working condition information, it specifically includes: obtaining the current time information and current positioning information; obtaining the information of the electrical equipment; obtaining the current ambient temperature information; combining the current time information and current positioning information to determine the load period information of the current thyristor dimming circuit in operation. The load period information at least includes the peak load period and the off-peak load period; combining the information of the electrical equipment, the current ambient temperature information and the load period information, and determining the working condition information of the current thyristor dimming circuit through a preset AC working condition model. The AC working condition model is pre-constructed through deep learning according to multiple sets of electrical equipment information, temperature information and load period information with working condition information annotations.

[0009] By adopting the above technical solutions, the system has established a complete operation environment perception ability by obtaining multi-dimensional real-time information, including key parameters such as time, location, electrical equipment, and ambient temperature. In particular, through the precise division of load periods, the system can perform differential processing according to the electrical characteristics of different periods. The introduction of the AC working condition model enables the system to accurately predict and adapt to various complex electrical environments, thereby providing a more accurate working condition reference basis for subsequent zero-crossing detection and improving the adaptability and operation stability of the system under different load conditions.

[0010] In combination with some embodiments of the first aspect, in some embodiments, after the step of obtaining the voltage signal and current signal of the thyristor dimming circuit, it further includes: combining the working condition information, determining an anti-interference strategy through a preset working condition detection strategy mapping table; performing data cleaning operations on the voltage signal and current signal according to the anti-interference strategy.

[0011] By adopting the above technical solutions, the introduction of the working condition detection strategy mapping table realizes the intelligent matching of signal processing strategies. According to different working condition characteristics, the system can automatically select the most suitable anti-interference strategy and perform targeted data cleaning on the voltage and current signals. This adaptive signal processing mechanism can effectively filter out various interference factors, improve the signal-to-noise ratio of the signal, and ensure the quality of the input data for subsequent zero-crossing detection.

[0012] In combination with some embodiments of the first aspect, in some embodiments, after the thyristor dimming circuit is started to work according to the set strategy, the step of determining the stroboscopic information specifically includes: obtaining the light intensity change data through a stroboscopic detection device; performing denoising processing on the light intensity change data by using a filtering algorithm; determining the stroboscopic information according to the light intensity change data, and the stroboscopic information at least includes stroboscopic frequency, stroboscopic depth, and periodic information of the stroboscope.

[0013] By adopting the above technical solutions, the stroboscopic detection device collects the light intensity change data in real time, removes the influence of environmental noise through a filtering algorithm, and accurately obtains the key characteristic parameters of the stroboscope. The system comprehensively analyzes multiple dimensions such as stroboscopic frequency, depth, and periodicity, and constructs a complete stroboscopic characteristic map. This multi-dimensional stroboscopic characteristic analysis method enables the system to more accurately evaluate the dimming effect and provide a reliable basis for subsequent precise zero-crossing adjustment.

[0014] In some embodiments in combination with some embodiments of the first aspect, in the step of determining the final zero-crossing information according to the stroboscopic deviation value in combination with the initial zero-crossing information, it specifically includes: if the stroboscopic deviation value is less than a preset deviation threshold, determining the initial zero-crossing information as the final zero-crossing information; if the stroboscopic deviation value is greater than or equal to the preset deviation threshold, determining an adjustment amount for the initial zero-crossing information according to the magnitude and direction of the stroboscopic deviation value and according to a preset adjustment strategy.

[0015] By adopting the above technical solution, the system establishes a threshold-based intelligent judgment mechanism. When the stroboscopic deviation value is within an acceptable range, the initial detection result is maintained to ensure the stability of the system; when the deviation value exceeds the threshold, directional adjustment is performed according to the specific characteristics of the deviation. This hierarchical adjustment strategy not only avoids system fluctuations caused by overcorrection but also ensures timely correction at necessary moments, achieving the optimal balance between the accuracy of zero-crossing detection and the stability of the system.

[0016] In some embodiments in combination with some embodiments of the first aspect, after the step of determining the final zero-crossing information according to the stroboscopic deviation value in combination with the initial zero-crossing information, it further includes: generating a trigger signal according to the final zero-crossing information, where the trigger signal is used to control the conduction and cutoff of the thyristor; collecting actual dimming effect data of the thyristor dimming circuit after adjusting the zero-crossing point; comparing the actual dimming effect data with preset dimming target data to calculate a dimming error value; if the dimming error value is greater than a preset error threshold, based on the dimming error value, in combination with current working condition information, current voltage signal, and current current signal, performing secondary optimization adjustment on the final zero-crossing information through a pre-trained dimming optimization model, where the dimming optimization model is obtained through deep learning training based on dimming error data, working condition information, voltage signals, and current signals under multiple different working conditions, and the dimming optimization model is used to output a zero-crossing point adjustment scheme that can minimize the dimming error; generating a new trigger signal according to the zero-crossing information after secondary optimization adjustment to control the operation of the thyristor dimming circuit until the dimming error value is less than or equal to the preset error threshold.

[0017] By adopting the above technical solution, the system constructs a complete closed-loop control mechanism. The generation of the trigger signal and the real-time monitoring of the actual effect form a feedback loop, and the detection result is dynamically optimized through the dimming optimization model. Especially when the error exceeds the limit, the system can automatically start the secondary optimization process and continuously adjust until the expected effect is achieved. This continuous optimization mechanism significantly improves the adaptive ability and control accuracy of the system, realizing the continuous optimization and stable improvement of the dimming effect.

[0018] In some embodiments in combination with some embodiments of the first aspect, after the step of generating a new trigger signal according to the zero-crossing information adjusted by the secondary optimization to control the operation of the thyristor dimming circuit until the dimming error value is less than or equal to the preset error threshold, the method further includes: continuously monitoring the operating state of the thyristor dimming circuit, where the operating state at least includes the temperature, current magnitude, and voltage fluctuation of the thyristor dimming circuit; when it is monitored that the temperature of the thyristor exceeds the preset temperature threshold, or the current and voltage show abnormal fluctuations, recording the abnormal data, where the abnormal data at least includes the time of the abnormality occurrence, the type of the abnormality, and the duration of the abnormality; according to the abnormal data, determining the suspected fault cause through a fault diagnosis model, where the fault diagnosis model is obtained through deep learning training based on multiple historical abnormal data and corresponding fault cause annotations; if the suspected fault cause is inaccurate zero-crossing detection, automatically switching to a backup zero-crossing detection strategy according to the corresponding fault type.

[0019] By adopting the above technical solutions, through real-time monitoring of key parameters such as temperature, current, and voltage, combined with the intelligent analysis of the fault diagnosis model, the system can timely identify and locate potential problems. Especially when abnormal zero-crossing detection is found, it can automatically switch to a backup strategy to ensure the continuous and reliable operation of the system. This multiple safeguard mechanism significantly improves the reliability and fault tolerance of the system and ensures the stable operation of the dimming system.

[0020] In a second aspect, the present application provides a zero-crossing detection system, where the zero-crossing detection system includes: one or more processors and a memory; the memory is coupled to the one or more processors, and the memory is used to store computer program code, where the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the zero-crossing detection system to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0021] In a third aspect, the present application provides a computer-readable storage medium, including instructions, which when running on the zero-crossing detection system, cause the zero-crossing detection system to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0022] In a fourth aspect, the present application provides a computer program product, which when running on the zero-crossing detection system, causes the zero-crossing detection system to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0023] One or more technical solutions provided in the embodiments of the present application at least have the following technical effects or advantages: 1. Due to the adoption of a double-verification mechanism that uses a zero-crossing detection model based on deep learning for initial detection and combines stroboscopic information for dynamic correction, the technical problems of low accuracy and poor reliability in zero-crossing detection in the prior art are effectively solved. Furthermore, the precise control of the thyristor conduction timing is achieved, the stroboscopic phenomenon of the lighting system is significantly reduced, and the stability of the dimming effect is improved.

[0024] 2. Due to the adoption of an intelligent judgment mechanism based on a preset threshold and a hierarchical adjustment strategy, the technical problems of easy fluctuation of zero-crossing detection results and unreasonable correction in the prior art are effectively solved. Furthermore, high accuracy in zero-crossing detection and high stability in system operation are achieved, enabling the detection results to neither be over-corrected nor fail to correct deviations in a timely manner.

[0025] 3. Due to the adoption of a dimming optimization model based on deep learning and a closed-loop control mechanism, and through real-time monitoring and secondary optimization for dynamic adjustment, the technical problems of difficult continuous optimization of the dimming effect and poor system adaptability in the prior art are effectively solved. Furthermore, precise control and continuous optimization of the dimming effect are achieved, significantly enhancing the system's adaptive ability and control accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is a flowchart of a zero-crossing detection method for a thyristor dimming circuit based on deep learning in an embodiment of the present application; Figure 2 is another flowchart of a zero-crossing detection method for a thyristor dimming circuit based on deep learning in an embodiment of the present application; Figure 3 is a schematic structural diagram of an entity device of a zero-crossing detection system in an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0027] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification and appended claims of the present application, the singular forms "a", "an", "the", "above", "said", "this" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used in the present application refers to and includes any and all possible combinations of one or more of the listed items.

[0028] Hereinafter, the terms "first" and "second" are only used for descriptive purposes and should not be construed as implying or indicating relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the embodiments of the present application, unless otherwise stated, the meaning of "a plurality" is two or more.

[0029] For ease of understanding, the method provided in this embodiment will be described in terms of its process below. Please refer to Figure 1 , which is a schematic flowchart of a zero-crossing detection method for a thyristor dimming circuit based on deep learning in an embodiment of the present application.

[0030] S101. Obtain operating condition information; The zero-crossing detection system will first obtain the current time information and location information through the internal clock module and positioning module. The time information is crucial for the operation analysis of the power system. In different time periods, the power load characteristics vary greatly. In the poultry farming scenario, different times and geographical locations will result in significant differences in power load characteristics. For example, in a northern farm in winter, the night temperature is extremely low, and a large number of heating devices need to be turned on, which will put the circuit in a high-load state; while in a southern farm in summer, a large number of ventilation and refrigeration devices are used for cooling during the day, which will also increase the circuit load. By obtaining this information, the system can initially judge the possible load conditions and environmental characteristics faced by the circuit in the poultry farm. The location information is equally crucial. The system uses the Global Positioning System (GPS) or other indoor positioning technologies to determine the geographical location of the thyristor dimming circuit. The power supply stability, electricity consumption habits, and seasonal characteristics in different regions are different, and the impact on the operating conditions of the dimming circuit is also different. For example, in winter in the north, heating devices are used frequently, which will increase the power load; in summer in the south, the high temperature leads to a sharp increase in electricity demand due to the use of air conditioners. By obtaining accurate time and location information, the system can initially judge the possible load conditions and environmental characteristics faced by the circuit.

[0031] The system obtains the information of the electrical equipment connected to the thyristor dimming circuit by interacting with an intelligent electricity meter or a power monitoring device. These information include the device type, power size, usage frequency, etc. The zero-crossing detection system also uses a temperature sensor to continuously monitor the temperature of the environment where the thyristor dimming circuit is located. The environmental temperature has a significant impact on the performance of circuit components, thereby affecting the accuracy of zero-crossing detection. Taking a thermistor as an example, its resistance value will change with the temperature. If the temperature is too high, the change in the resistance value will cause the circuit parameters to drift, affecting the accuracy of voltage and current signals. In a high-temperature environment, the on and off characteristics of the thyristor will also change, which may lead to deviations in zero-crossing detection.

[0032] The system comprehensively analyzes the obtained time and location information to determine the load period information for the thyristor dimming circuit operation. Through the learning and analysis of historical data, the system can judge whether the circuit is in a peak load period or a low load period at the current time and geographical location. For example, the different poultry farming activities at different times will lead to different electricity demands. For example, in large-scale poultry farms, early in the morning is the concentrated period for feeding, and at this time, the automatic feeding equipment, lighting equipment, and ventilation equipment in the farm will be turned on simultaneously. The automatic feeding equipment needs electricity to drive the feed conveyance, the lighting equipment provides sufficient light for the operators, and the ventilation equipment ensures air circulation. The simultaneous operation of these equipment greatly increases the circuit load, and at this time, the circuit is in a peak load period. In addition, at night, in order to ensure the rest environment of the poultry, most of the lighting equipment will be turned off, but some temperature control equipment and monitoring equipment are still running. Compared with the electricity consumption in the early morning, the circuit load at this time is lower, and it is in a low load period. The determination of the load period helps the system to adjust the detection strategy in advance.

[0033] The zero-crossing detection system takes the previously obtained electrical equipment information, environmental temperature information, and load period information as inputs and feeds them into a pre-trained AC working condition model. This model is constructed through deep learning algorithms based on a large amount of data annotated with working condition information. The AC working condition model can comprehensively consider factors such as the operating state of the equipment, environmental temperature, and load period to accurately judge the working condition of the circuit. If, in a high-temperature environment, high-power production equipment is running simultaneously during the peak load period, the model will judge the possible voltage fluctuations, current overloads, etc. faced by the circuit based on this information and output the corresponding working condition information. These working condition information include the load type (resistive, inductive, or capacitive load), load size, potential interference sources, etc. of the circuit. The system provides a more accurate reference for the subsequent zero-crossing detection based on these working condition information to ensure accurate zero-crossing detection and stable dimming control under different working conditions.

[0034] S102. Obtain the voltage signal and current signal of the thyristor dimming circuit; The zero-crossing detection system obtains voltage signals and current signals by installing voltage sensors and current sensors at key nodes of the thyristor dimming circuit. These sensors have the characteristics of high precision and fast response, and can collect the voltage and current changes in the circuit in real time. For the acquisition of voltage signals, a resistive voltage divider sensor is usually selected. In the main circuit of the thyristor dimming circuit, a voltage division circuit is formed by reasonably configuring high-precision resistors to convert the high voltage into a voltage range suitable for sensor measurement according to a certain ratio. In terms of the acquisition of current signals, in the thyristor dimming circuit, the primary winding of the current transformer is connected in series in the circuit, and the small current output by the secondary winding is converted into a voltage signal through a sampling resistor and then sent to the zero-crossing detection system. The current transformer uses the principle of electromagnetic induction to convert the large current in the main circuit into a small current for measurement.

[0035] S103. Combine the working condition information, the voltage signal, and the current signal, and determine the initial zero-crossing information through the circuit zero-crossing detection model. The initial zero-crossing information at least includes the time point information and phase information of the zero-crossing point. The circuit zero-crossing detection model is constructed in advance through deep learning based on multiple sets of working condition information, voltage signals, and current signals with zero-crossing annotation information. After obtaining the working condition information, voltage signal, and current signal, the zero-crossing detection system uses the circuit zero-crossing detection model to determine the initial zero-crossing information. The model is constructed based on deep learning and specifically uses a convolutional neural network (CNN) to achieve feature extraction and pattern recognition capabilities. In the model training stage, a large number of sets of working condition information, voltage signals, and current signals with zero-crossing annotation information are collected. These data cover various different working conditions, such as different load types (resistive, inductive, capacitive loads), different load sizes, different ambient temperatures, and different load periods, etc. Through preprocessing these data, including data cleaning, normalization, etc., they are converted into a format suitable for model training. In the model structure design, the CNN includes multiple convolutional layers, pooling layers, and fully connected layers. The convolutional layer slides the convolutional kernel on the input data for convolution to extract local features in the voltage signal and current signal, such as signal waveform changes, phase features, etc. The pooling layer downsamples the output of the convolutional layer to reduce the data dimension, while retaining key features, improving the computational efficiency and generalization ability of the model. The fully connected layer integrates the features after convolution and pooling processing and outputs the final zero-crossing time point information and phase information.

[0036] During the actual operation process, the working condition information, voltage signal, and current signal obtained in real time are input into the trained circuit zero-crossing detection model. The model first extracts features from the input data, and the convolutional layer will automatically learn the feature patterns related to the zero-crossing of the voltage signal and current signal under different working conditions. For example, under the inductive load working condition, the model can identify the phase lag feature of the current signal near the zero-crossing; under the high-temperature environment working condition, the model can capture the relationship between the signal distortion feature caused by temperature influence and the zero-crossing. Then, through the multi-layer calculation and processing of the model, the time point information and phase information of the initial zero-crossing are finally output.

[0037] In order to continuously improve the performance and adaptability of the model, the model also needs to be updated and optimized regularly. Collect new working condition data and signal data, and retrain the model so that it can better adapt to the changing circuit operation environment and working condition conditions, and ensure the detection accuracy and reliability of the initial zero-crossing information.

[0038] S104. After starting the thyristor dimming circuit to work according to the set strategy, determine the stroboscopic information; After the thyristor dimming circuit starts to work according to the set strategy, the zero-crossing detection system determines the stroboscopic information through the following series of technical means: First, use a special stroboscopic detection device to obtain the light intensity change data. Common stroboscopic detection devices use photoelectric sensors, such as silicon photodiodes. This sensor is very sensitive to light intensity changes and can quickly respond to changes in illumination intensity. It is installed near the lamp to ensure that the light intensity change signal emitted by the lamp can be accurately collected. For example, in a lighting system, the photoelectric sensor is installed at a certain distance below the lamp to avoid interference from other light sources to obtain accurate light intensity change data of the lamp.

[0039] After collecting the light intensity change data, use a filtering algorithm to denoise it. The median filtering algorithm is a commonly used denoising method. It replaces the value of each point in the data sequence with the median value of the data in its neighborhood, which can effectively remove discrete noises such as salt-and-pepper noise. In actual applications, the window size of the median filter is reasonably set according to the characteristics of the light intensity change data and the noise situation. For example, for light intensity change data with a high sampling frequency, a smaller window size, such as 3 or 5, can be selected; for cases with more noise and larger data fluctuations, the window size is appropriately increased, such as 7 or 9, to better remove noise and retain the true signal of the light intensity change.

[0040] The stroboscopic information is determined based on the denoised light intensity change data. By performing spectral analysis on the light intensity change data and using the Fast Fourier Transform (FFT) algorithm, the light intensity change data in the time domain is converted to the frequency domain to obtain the stroboscopic frequency. For example, after the FFT transformation, the frequency component with a larger amplitude in the frequency spectrum diagram is the stroboscopic frequency, and the stroboscopic depth is determined by calculating the ratio of the difference between the maximum and minimum values of the light intensity change to the sum of the maximum and minimum values. For the periodic information of stroboscopy, it is obtained by analyzing the periodic characteristics of the light intensity change data. For example, by observing the interval time when the same waveform repeats in the light intensity change curve, the period of stroboscopy is determined. By accurately obtaining and analyzing this multi-dimensional stroboscopic information, it provides a strong basis for subsequent adjustment of the zero-crossing information.

[0041] S105. Compare the stroboscopic information with a preset stroboscopic standard to determine the stroboscopic deviation value; In the poultry farming scenario, poultry are extremely sensitive to the stroboscopic light of the light source. When the zero-crossing detection system determines the stroboscopic deviation value, it needs to make an accurate comparison according to the stroboscopic standards specifically formulated for poultry lighting. These standards are determined based on a large number of poultry farming experiments and research on the visual characteristics of poultry. For example, research shows that when the stroboscopic frequency is higher than 20Hz, poultry may have a stress response, which affects growth. Therefore, in poultry lighting, the preset stroboscopic frequency standard is usually set below 20Hz.

[0042] The zero-crossing detection system retrieves the stroboscopic standard data for poultry lighting from the storage unit. This data not only includes the ideal values of the stroboscopic frequency and stroboscopic depth but also covers the requirements for the stable range of stroboscopic periodicity. For the stroboscopic depth, considering the sensitivity of poultry to the change in light intensity, the preset stroboscopic depth standard may be set within 3%. If the actually detected stroboscopic depth is 6%, the system calculates the stroboscopic depth deviation value as 6% - 3% = 3%. For the periodicity of stroboscopy, the system will strictly judge whether it meets the preset stable period requirements. In poultry lighting, a stable stroboscopic period is crucial for the biological clock and growth rhythm of poultry. If the preset stroboscopic period should be stable at 0.05 seconds (corresponding to a frequency of 20Hz), and the actually detected stroboscopic period fluctuates greatly and deviates significantly from the standard period, the system will record this periodic deviation and calculate the deviation degree according to the preset rules.

[0043] To visually present the comparison results, the system will generate a dedicated poultry lighting stroboscopic deviation report. The report is presented in the form of a combination of charts and data, clearly listing the actual stroboscopic information, the preset stroboscopic standard, and the calculated deviation value. For example, a bar chart is used to compare the actual and standard stroboscopic frequencies and stroboscopic depths, and a line chart is used to show the change in the stroboscopic cycle, enabling the operator to quickly understand the specific situation of the stroboscopic deviation. The stroboscopic deviation value obtained from each comparison will be stored in the database and associated with information such as the lighting time, ambient temperature, and poultry growth stage at that time. By analyzing historical data, the system can predict the development trend of the stroboscopic deviation. If it is found that the stroboscopic deviation value continues to rise, the system will issue a warning signal to prompt the staff to check the lighting system in a timely manner to prevent adverse effects on poultry due to stroboscopic problems.

[0044] S106. Combine the initial zero-crossing information and determine the final zero-crossing information according to the stroboscopic deviation value.

[0045] When the stroboscopic deviation value is less than the preset deviation threshold, it indicates that the current initial zero-crossing information can provide a relatively stable lighting environment for poultry and will not have an obvious stress impact on poultry. At this time, the system will directly determine the initial zero-crossing information as the final zero-crossing information to maintain the stable operation of the system and avoid interference with poultry caused by unnecessary adjustments. For example, if the preset stroboscopic frequency deviation threshold is ±3 Hz, and the actually calculated stroboscopic frequency deviation value is within this range, the system will not modify the initial zero-crossing information.

[0046] If the stroboscopic deviation value is greater than or equal to the preset deviation threshold, the system will start an accurate adjustment program. For the stroboscopic frequency deviation, if the actual stroboscopic frequency is higher than the preset standard, it means that the on and off timing of the thyristor needs to be adjusted to reduce the flashing frequency of the lamp. The system will calculate the adjustment amount of the initial zero-crossing time point according to the preset adjustment rules and in combination with the frequency deviation value. For example, after multiple experimental verifications, in the poultry lighting system, when the stroboscopic frequency is 1 Hz higher than the standard, the zero-crossing time point needs to be advanced by 0.02 milliseconds. Assuming that the actual stroboscopic frequency is 8 Hz higher than the standard, then the system will calculate that the initial zero-crossing time point needs to be advanced by 0.16 milliseconds.

[0047] In terms of phase information adjustment, if the stroboscopic depth deviates significantly, it indicates that the phase relationship between the voltage and current signals needs to be optimized. The system will adjust according to the direction and magnitude of the stroboscopic depth deviation value, combined with the phase characteristics of the poultry lighting circuit. If the stroboscopic depth is too large and it is analyzed that it is caused by the phase lag of the current signal, the system will appropriately adjust the phase of the zero-crossing point to make the current signal cross zero at a more appropriate moment, thereby reducing the stroboscopic depth and the fluctuation of the light intensity. During the adjustment process, the system will use advanced control algorithms to accurately modify the zero-crossing point information. These algorithms have been tested and optimized a large number of times to ensure the accuracy and stability of the adjustment. After the adjustment is completed, the system will monitor the dimming effect in real time, detect the stroboscopic information again, and verify whether the adjusted zero-crossing point information effectively reduces the stroboscopic deviation value. If the adjustment effect is not ideal, the system will optimize the adjustment strategy again according to the feedback information until the stroboscopic deviation value reaches an acceptable range.

[0048] In addition, the system will also consider the different requirements in different growth stages of poultry. In the chick stage, poultry are more sensitive to stroboscopes, and the system will adopt more stringent adjustment standards; while in the adult stage, the standards can be appropriately relaxed. The system will dynamically adjust the adjustment strategy of the zero-crossing point information according to the growth stage information of poultry to ensure that a suitable lighting environment can be provided for poultry in different growth stages, promote the healthy growth of poultry, and improve the breeding efficiency.

[0049] In the embodiments of this application, by comprehensively applying advanced sensor technology, intelligent deep learning models, and precise adjustment strategies, the accurate detection and dynamic optimization of the zero-crossing point of the thyristor dimming circuit are realized, which not only effectively reduces the stroboscopic phenomenon in the poultry lighting system, but also greatly improves the quality of the growth environment of poultry and the breeding efficiency.

[0050] After combining the above content, the following is a further and more specific process description of the method provided in this embodiment. Please refer to Figure 2 , which is another process schematic diagram of the zero-crossing point detection method for the thyristor dimming circuit based on deep learning in the embodiments of this application.

[0051] S201. Generate a trigger signal according to the final zero-crossing point information, where the trigger signal is used to control the conduction and cutoff of the thyristor; After the zero-crossing point detection system determines the final zero-crossing point information, it will generate a trigger signal for controlling the conduction and cutoff of the thyristor based on this information.

[0052] The system first analyzes the final zero-crossing information to obtain key data such as the time point and phase of the zero crossing. These data are transmitted to a dedicated trigger signal generation module, which works in coordination with hardware circuits and software algorithms. At the hardware level, a high-precision clock generator is used, which can generate a stable and accurate clock signal to provide a time reference for the generation of trigger signals. In terms of software, complex logic control algorithms are used to calculate the delay time and pulse width of the trigger signal based on the zero-crossing information.

[0053] In practical applications, such as in the poultry farming lighting scenario, the system adjusts the trigger signal according to different farming requirements and lighting conditions. If the poultry is in the rest stage and requires a darker lighting environment, the system will, based on the final zero-crossing information, extend the cut-off time of the thyristor to make the lamp output a lower brightness. Specifically, assuming the current final zero-crossing time point is t0, the system calculates that the trigger signal needs to be delayed by Δt time according to the dimming requirement, and the pulse width is set to T. By controlling the clock generator and the logic circuit, a pulse signal with a width of T is generated at the moment of t0 + Δt as the trigger signal.

[0054] To ensure the accuracy and stability of the trigger signal, the system will perform multiple checks. On the one hand, the time accuracy of the generated trigger signal is verified by comparing the actual emission time of the trigger signal with the expected time. If there is an error, the parameters of the clock generator and the logic control algorithm are adjusted in a timely manner. On the other hand, the pulse width of the trigger signal is monitored to ensure that it meets the set requirements. At the same time, the system also monitors the working state of the thyristor in real time, such as the current and voltage conditions, and feedback-adjusts the trigger signal to adapt to the dynamic changes of the circuit.

[0055] S202. Collect the actual dimming effect data of the thyristor dimming circuit after adjusting the zero crossing; After the zero-crossing detection system triggers the thyristor dimming circuit to work, it is necessary to collect its actual dimming effect data to evaluate the actual dimming effect and provide a basis for subsequent optimization and adjustment.

[0056] During the collection process, the system mainly obtains data through optical sensors and other relevant monitoring devices. In the poultry farming lighting scenario, the installation location of the optical sensor is crucial. To measure the light intensity comprehensively and accurately, the system will evenly distribute multiple optical sensors in the farming area, such as installing them in different poultry house areas and at different height positions to ensure that the entire farming space can be covered and light data at different positions can be obtained.

[0057] Light sensors are selected with high sensitivity and wide dynamic range, capable of accurately measuring the changes in light intensity under different brightness levels. These sensors convert light signals into electrical signals and then convert them into digital signals through analog-to-digital conversion circuits for easy processing and analysis by the system. In addition to light intensity, the system also collects data on light uniformity. By collecting light intensity data at different positions, statistical quantities such as the standard deviation of light intensity are calculated to evaluate light uniformity. Regarding the setting of the time interval for data collection, the system will make dynamic adjustments according to the actual situation.

[0058] S203. Compare the actual dimming effect data with the preset dimming target data to calculate the dimming error value; the system reads the actual dimming effect data and the preset dimming target data from the database. The preset dimming target data is formulated according to the different stages and requirements of poultry farming. For example, in the chick stage, to promote its growth and development, a higher light intensity and a specific color rendering index are required; while in the adult chicken stage, the requirements for light intensity and color rendering index are different. These target data are stored in the parameter configuration table of the system for easy access at any time.

[0059] For light intensity, the system compares the average value of the actually collected light intensity with the preset target light intensity. Suppose the preset target light intensity is Lux0 and the average value of the actually collected light intensity is Lux1, then the dimming error value E1 = |Lux1 - Lux0|. If the actual light intensity is higher than the target value, the error value is positive; otherwise, it is negative. In terms of light uniformity, the system determines the error value by calculating the difference between the actual light uniformity index and the preset target uniformity index. The preset light uniformity target may be a specific standard deviation range, such as σ0, and the standard deviation of the actually measured light uniformity is σ1. When σ1 exceeds the σ0 range, the dimming error value E2 is calculated according to the exceeding degree. For the color rendering index, the actually measured color rendering index Ra1 is also compared with the preset target color rendering index Ra0, and the dimming error value E3 = |Ra1 - Ra0|. By comprehensively considering the dimming error values of these different indicators, the system can comprehensively evaluate the deviation between the dimming effect and the target.

[0060] S204. If the dimming error value is greater than the preset error threshold, then based on this dimming error value, combined with the current working condition information, the current voltage signal, and the current current signal, the final zero-crossing information is secondarily optimized and adjusted through a pre-trained dimming optimization model. This dimming optimization model is obtained through deep learning training based on dimming error data, working condition information, voltage signals, and current signals under multiple different working conditions. This dimming optimization model is used to output a zero-crossing adjustment scheme that can minimize the dimming error. When the dimming error value calculated by the zero-crossing detection system is greater than the preset error threshold, it indicates that there is a large deviation between the current dimming effect and the expected target. At this time, the system starts the secondary optimization adjustment process. The system first integrates and processes the dimming error value, the current working condition information, the current voltage signal, and the current current signal. In the actual poultry farming scenario, the working condition information includes data such as the temperature and humidity of the farming area where the thyristor dimming circuit is located. These information are obtained by the previous acquisition module and stored in the system. The voltage and current signals are collected in real time by sensors installed at the key nodes of the thyristor dimming circuit.

[0061] Next, the system inputs the integrated data into a pre-trained dimming optimization model, which is constructed based on deep learning, such as adopting a deep neural network (DNN) architecture. During the training phase, the model uses a large amount of dimming error data, working condition information, voltage signals, and current signals under different working conditions. These data are labeled with the corresponding optimal zero-crossing adjustment schemes, and the parameters of the model are continuously optimized through the backpropagation algorithm, enabling the model to learn the mapping relationship between different input data and the optimal zero-crossing adjustment schemes.

[0062] In terms of the model structure, the DNN contains multiple hidden layers, and each hidden layer consists of a large number of neurons. The input layer receives the integrated data, and the neurons perform weighted summation and nonlinear transformation on the data through weights and biases, and transfer the processed data to the next layer. The neurons in the hidden layer gradually extract the key information related to dimming error and zero-crossing adjustment by continuously learning the characteristics of the data. The output layer outputs the zero-crossing adjustment scheme that can minimize the dimming error according to the output of the hidden layer. This scheme includes information such as the adjustment amount of the zero-crossing time point and the adjustment amount of the phase. Taking poultry farming lighting as an example, if the temperature in the current farming area is relatively high and the dimming error indicates that the light intensity is too strong, the model will analyze that it is caused by the too long conduction time of the thyristor in combination with the current voltage and current signal conditions. Through learning a large amount of historical data, the adjustment scheme output by the model may be to appropriately advance the zero-crossing time point and fine-tune the phase at the same time, so that the thyristor is cut off in advance, thereby reducing the brightness of the lamp.

[0063] S205. Generate a new trigger signal according to the zero-crossing information after the secondary optimization adjustment to control the thyristor dimming circuit to work until the dimming error value is less than or equal to the preset error threshold; After obtaining the zero-crossing information after the secondary optimization adjustment, the zero-crossing detection system starts to generate a new trigger signal to control the thyristor dimming circuit to work, and continues this process until the dimming error value reaches the preset standard.

[0064] The system transmits the zero-crossing information after secondary optimization adjustment to the trigger signal generation module. This module generates a trigger signal based on the time point and phase adjustment amount in the zero-crossing information, combined with the hardware circuit and software algorithm. In terms of hardware, the high-precision clock generator still serves as the time reference to ensure the time accuracy of the trigger signal. The software algorithm calculates the delay time of the trigger signal according to the adjusted zero-crossing time point, and determines the pulse width and phase relationship of the trigger signal according to the phase adjustment amount.

[0065] In an actual poultry farming lighting system, assume that the zero-crossing time point after secondary optimization adjustment is t1. The system calculates that the trigger signal needs to be issued at the moment of t1 + Δt1, and the pulse width is adjusted to T1. By controlling the clock generator and logic circuit, a pulse signal that meets the requirements is generated at the specified moment as the new trigger signal. The new trigger signal is transmitted to the thyristor through the drive circuit. The drive circuit amplifies and isolates the signal to ensure that the trigger signal can reliably control the conduction and cut-off of the thyristor. During the transmission process, the system uses shielded cables to connect, reducing the impact of external electromagnetic interference on the trigger signal.

[0066] After each new trigger signal is sent, the system will collect the actual dimming effect data of the thyristor dimming circuit again, including indicators such as illuminance, lighting uniformity, and color rendering index. After the collection is completed, these data are compared with the preset dimming target data, and the dimming error value is recalculated. If the dimming error value is still greater than the preset error threshold, the system will input the current dimming error value, operating conditions information, voltage signal, and current signal into the dimming optimization model again for adjustment to generate a new trigger signal, and so on in a loop.

[0067] The system also has a feedback adjustment mechanism. During the dimming process, it monitors the working state of the thyristor in real time, such as current and voltage changes. If it is found that the thyristor is working abnormally, such as excessive current or severe voltage fluctuations, the system will adjust the parameters of the trigger signal in a timely manner to ensure the safe and stable operation of the thyristor and the entire dimming circuit. Through continuous adjustment and optimization, until the dimming error value is less than or equal to the preset error threshold, at this time, it is considered that the dimming effect reaches the expected goal, and further adjustment is stopped.

[0068] S206. Continuously monitor the operating state of the thyristor dimming circuit, and this operating state at least includes the temperature, current magnitude, and voltage fluctuation condition of the thyristor dimming circuit; The system monitors the operating status through various sensors installed in the thyristor dimming circuit. For temperature monitoring, a thermistor sensor is used. The resistance value of the thermistor changes significantly with temperature. Installed near the thyristor element, it can sense the temperature change of the thyristor in real time. A current transformer is used to obtain the current magnitude of the thyristor dimming circuit. For monitoring voltage fluctuations, a voltage sensor is adopted. The voltage sensor can collect voltage signals in the circuit in real time, and the system analyzes the collected voltage signals to calculate the amplitude, frequency, and fluctuation conditions of the voltage.

[0069] S207. When it is detected that the temperature of the thyristor exceeds the preset temperature threshold, or the current or voltage shows abnormal fluctuations, record the abnormal data, which at least includes the time of the abnormality occurrence, the type of abnormality, and the duration of the abnormality. During the continuous monitoring of the operating status of the thyristor dimming circuit by the zero-crossing detection system, once it is found that the temperature of the thyristor exceeds the preset temperature threshold, or the current or voltage shows abnormal fluctuations, the abnormal data recording program will be immediately started. Various sensors built into the thyristor dimming circuit continuously collect temperature, current, and voltage data. For temperature monitoring, the thermistor sensor plays a key role. The thermistor is tightly installed on the heat sink of the thyristor element or other key heat-generating parts, and its resistance value changes significantly with the change of temperature. The system calculates the real-time temperature of the thyristor accurately by measuring the resistance value of the thermistor and based on the pre-calibrated resistance-temperature conversion relationship. In terms of current monitoring, the current transformer converts the large current in the main circuit into a small current, which is converted into a voltage signal after passing through a sampling resistor and then collected and quantified by the analog-digital conversion module of the system. The system calculates parameters such as the amplitude and change rate of the current in real time. When the current amplitude exceeds the preset normal range (such as 1.2 times the rated current), or the current change rate is abnormal (such as a sharp rise or fall in current within a short period of time), the system determines that the current shows abnormal fluctuations. The system also analyzes the voltage signal in real time to monitor indicators such as the amplitude, frequency, and harmonic content of the voltage. If the voltage amplitude exceeds the normal operating range (such as ±10% of the rated voltage), or the voltage frequency deviates from the standard value (such as for a 50Hz power grid, the frequency deviation exceeds ±0.5Hz), or the voltage harmonic content is too high, the system judges that the voltage has abnormal fluctuations.

[0070] Once an abnormality is detected, the system quickly starts the recording program. For the type of abnormality, the system identifies it according to the data collected by the sensors and the preset judgment logic. If the temperature exceeds the threshold, the type of abnormality is marked as "too high temperature"; if the current amplitude is too large, it is marked as "overcurrent"; if the voltage amplitude is abnormal, it is marked as "overvoltage" or "undervoltage", and if the voltage frequency is abnormal, it is marked as "frequency abnormality", etc.

[0071] S208. Based on the abnormal data, determine the suspected cause of the fault through a fault diagnosis model, which is obtained through deep learning training based on multiple historical abnormal data and corresponding fault cause annotations; After the zero-crossing detection system records the abnormal data, it will immediately call a pre-trained fault diagnosis model to determine the suspected cause of the fault. The fault diagnosis model is constructed based on deep learning technology. By learning a large amount of historical abnormal data and corresponding fault cause annotations, it has powerful pattern recognition and fault diagnosis capabilities. The system preprocesses the recorded abnormal data to make it meet the input requirements of the fault diagnosis model. The preprocessed data is input into the fault diagnosis model. The model adopts architectures such as deep neural network (DNN) or convolutional neural network (CNN). Taking DNN as an example, the model contains multiple hidden layers, and each hidden layer consists of a large number of neurons. The input layer receives the preprocessed abnormal data, and the neurons perform weighted summation and non-linear transformation on the data through weights and biases, and transfer the processed data to the next layer. The neurons in the hidden layer continuously learn the features of the data and gradually extract the key information related to the fault cause.

[0072] In the training stage, the model uses a large number of different types of abnormal data and corresponding accurate fault cause annotations. For example, for the abnormal situation of too high temperature, the possible causes of the fault include cooling system failure, overloading, etc.; the abnormal current may be caused by thyristor element damage, circuit short circuit, etc. Through the backpropagation algorithm, the model continuously adjusts the weights and biases of the neurons to make the output of the model as close as possible to the actual fault cause.

[0073] In actual operation, when the current abnormal data is input, the model goes through multiple layers of calculation and processing and outputs the prediction result of the fault cause. The model will give multiple suspected fault causes and assign a confidence score to each cause. For example, the model predicts that the suspected cause of the current abnormal current is "thyristor element damage" with a confidence of 0.8; the confidence of "circuit short circuit" is 0.2. The system preferentially outputs the fault cause with the highest confidence as the main suspected fault cause according to the confidence score.

[0074] S209. If the suspected cause of the fault is inaccurate zero-crossing detection, automatically switch to the backup zero-crossing detection strategy according to the corresponding fault type.

[0075] When the zero-crossing detection system determines through the fault diagnosis model that the suspected cause of the fault is inaccurate zero-crossing detection, in order to ensure the stable operation of the system and the dimming effect is not affected, it will automatically switch to the backup zero-crossing detection strategy according to the corresponding fault type. [[ID=IS]]

[0076] The system first analyzes and judges the results output by the fault diagnosis model. If it is determined that the suspected fault cause is inaccurate zero-crossing detection, the system will further identify the fault type. For different fault types, the system calls the corresponding backup zero-crossing detection strategy. If the fault type is a sensor fault, the system will switch to a backup detection strategy based on hardware redundancy. In the hardware design, the system has pre-set multiple sets of voltage and current sensors. When the main sensor fails, the system automatically switches to the backup sensor for signal acquisition. If the fault type is the deviation of the deep learning model under specific working conditions, the system will switch to a backup zero-crossing detection strategy based on traditional algorithms. For example, the zero-crossing detection method based on a comparator is used as a backup solution. Although the traditional method has certain limitations, it can still provide basic zero-crossing detection functions in some cases. The system uses a comparator to compare the voltage or current signal with the zero level. When the voltage or current value of the signal changes from positive to negative or from negative to positive and passes through the zero level, it is determined that the zero-crossing point is detected. To improve the accuracy of the traditional method, the system will also combine some filtering and anti-interference measures. For example, before the signal is input to the comparator, a low-pass filter is used to remove high-frequency noise interference.

[0077] After the switch is completed, the system will monitor the operation effect of the backup zero-crossing detection strategy in real time. By collecting the actual dimming effect data of the dimming circuit, such as the light intensity, stroboscopic situation, etc., and comparing it with the preset standard, the effectiveness of the backup strategy is evaluated. If the backup strategy can meet the basic requirements of the system, the system continues to run; if the backup strategy still cannot solve the problem, the system will issue a higher-level fault alarm to prompt the technical personnel to conduct a comprehensive inspection and repair.

[0078] In the embodiments of this application, by constructing a complete closed-loop control mechanism, introducing multiple deep learning models, and setting multiple safeguard measures, the precise control and continuous optimization of the thyristor dimming circuit are realized. It not only effectively solves the technical problems of difficult continuous optimization of the dimming effect, poor system adaptability, low zero-crossing detection accuracy, and poor reliability in the prior art, but also realizes the precise control and continuous optimization of the dimming effect, and significantly improves the adaptive ability and control accuracy of the system.

[0079] The following describes the zero-crossing detection system in the embodiments of the present invention application from the perspective of hardware processing. Please refer to Figure 3 , which is a schematic structural diagram of an entity device of the zero-crossing detection system in the embodiments of this application.

[0080] It should be noted that Figure 3 The structure of the zero-crossing detection system shown is only an example and should not bring any limitations to the functions and usage scopes of the embodiments of the present invention.

[0081] Such as Figure 3As shown, the zero-crossing detection system includes a Central Processing Unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the Read-Only Memory (ROM) 302 or the program loaded from the storage section 308 into the Random Access Memory (RAM) 303, such as executing the method described in the above embodiments. In the RAM 303, various programs and data required for system operation are also stored. The CPU 301, ROM 302, and RAM 303 are connected to each other via a bus 304. An Input / Output (I / O) interface 305 is also connected to the bus 304.

[0082] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, a button switch, etc.; an output section 307 including a Liquid Crystal Display (LCD), an audio output device, an indicator light, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read from it can be installed into the storage section 308 as needed.

[0083] Specifically, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network via the communication section 309, and / or installed from the removable medium 311. When the computer program is executed by the Central Processing Unit (CPU) 301, various functions defined in the present invention are executed.

[0084] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fibers, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In the present invention, a computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0085] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present invention. Among them, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the above-mentioned module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings.

[0086] Specifically, the zero-crossing detection system of this embodiment includes a processor and a memory. A computer program is stored on the memory. When the computer program is executed by the processor, it implements the zero-crossing detection method for thyristor dimming circuits based on deep learning provided in the above-mentioned embodiment.

[0087] On the other hand, the present invention also provides a computer-readable storage medium, which may be included in the zero-crossing detection system described in the above-mentioned embodiment; or it may exist alone without being assembled into the zero-crossing detection system. The above storage medium carries one or more computer programs. When the above one or more computer programs are executed by a processor of the zero-crossing detection system, the zero-crossing detection system is enabled to implement the zero-crossing detection method for thyristor dimming circuits based on deep learning provided in the above-mentioned embodiment.

[0088] As mentioned above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the various embodiments of the present application.

[0089] As used in the foregoing embodiments, depending on the context, the term "when" can be construed to mean "if" or "after" or "in response to determining" or "in response to detecting". Similarly, depending on the context, the phrase "upon determining" or "if (the stated condition or event) is detected" can be construed to mean "if determined" or "in response to determining" or "when (the stated condition or event) is detected" or "in response to detecting (the stated condition or event)".

[0090] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the foregoing embodiments can be implemented by a computer program instructing relevant hardware. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the foregoing method embodiments. The foregoing storage media include: various media such as ROM or random access memory RAM, magnetic disks, or optical discs that can store program codes.

Claims

1. A zero-crossing detection method for a thyristor dimming circuit based on deep learning, applied to a zero-crossing detection system, characterized in that The method includes: Obtaining operating condition information; Obtaining the voltage signal and current signal of the thyristor dimming circuit; Combining the operating condition information, the voltage signal and the current signal, and determining initial zero-crossing information through a circuit zero-crossing detection model. The initial zero-crossing information at least includes the time point information and phase information of the zero-crossing point. The circuit zero-crossing detection model is constructed in advance through deep learning based on multiple sets of operating condition information, voltage signals and current signals with zero-crossing annotation information; After starting the operation of the thyristor dimming circuit according to a set strategy, determining stroboscopic information; Comparing the stroboscopic information with a preset stroboscopic standard to determine the stroboscopic deviation value; Combining the initial zero-crossing information, and determining the final zero-crossing information according to the stroboscopic deviation value.

2. The method according to claim 1, wherein In the step of obtaining operating condition information, it specifically includes: Obtaining the current time information and current positioning information; Obtaining electrical equipment information; Obtaining the current ambient temperature information; Combining the current time information and current positioning information to determine the load period information of the current thyristor dimming circuit operation. The load period information at least includes peak load periods and off-peak load periods; Combining the electrical equipment information, the current ambient temperature information and the load period information, and determining the operating condition information of the current thyristor dimming circuit through a preset AC power operating condition model. The AC power operating condition model is constructed in advance through deep learning based on multiple sets of electrical equipment information, temperature information and load period information sets with operating condition information annotation; 3. The method according to claim 1, wherein After the step of obtaining the voltage signal and current signal of the thyristor dimming circuit, it further includes: Combining the operating condition information, and determining an anti-interference strategy through a preset operating condition detection strategy mapping table; Performing data cleaning operations on the voltage signal and current signal according to the anti-interference strategy.

4. The method according to claim 1, wherein In the step of determining stroboscopic information after starting the operation of the thyristor dimming circuit according to a set strategy, it specifically includes: Obtaining light intensity change data through a stroboscopic detection device; Performing denoising processing on the light intensity change data by using a filtering algorithm; Determining stroboscopic information according to the light intensity change data. The stroboscopic information at least includes stroboscopic frequency, stroboscopic depth and periodic information of stroboscopy.

5. The method according to claim 1, wherein In the step of combining the initial zero-crossing information and determining the final zero-crossing information according to the stroboscopic deviation value, it specifically includes: If the stroboscopic deviation value is less than a preset deviation threshold, determining the initial zero-crossing information as the final zero-crossing information; If the stroboscopic deviation value is greater than or equal to the preset deviation threshold, determining an adjustment amount for the initial zero-crossing information according to the magnitude and direction of the stroboscopic deviation value and a preset adjustment strategy.

6. The method according to claim 1, wherein After the step of combining the initial zero-crossing information and determining the final zero-crossing information according to the stroboscopic deviation value, it further includes: Generating a trigger signal according to the final zero-crossing information, and the trigger signal is used to control the conduction and cut-off of the thyristor; Collecting the actual dimming effect data of the thyristor dimming circuit after adjusting the zero-crossing point; ​ If the dimming error value is greater than a preset error threshold, based on the dimming error value, in combination with the current working condition information, the current voltage signal, and the current current signal, the final zero-crossing information is secondarily optimized and adjusted through a pre-trained dimming optimization model. The dimming optimization model is obtained through deep learning training based on dimming error data, working condition information, voltage signals, and current signals under multiple different working conditions. The dimming optimization model is used to output a zero-crossing adjustment scheme that can minimize the dimming error; Generate a new trigger signal according to the zero-crossing information after the secondary optimization adjustment to control the operation of the thyristor dimming circuit until the dimming error value is less than or equal to the preset error threshold.

7. The method according to claim 1, characterized in that After the step of generating a new trigger signal according to the zero-crossing information after the secondary optimization adjustment to control the operation of the thyristor dimming circuit until the dimming error value is less than or equal to the preset error threshold, it further includes: Continuously monitor the operating state of the thyristor dimming circuit. The operating state at least includes the temperature of the thyristor dimming circuit, the magnitude of the current, and the voltage fluctuation condition; When it is monitored that the temperature of the thyristor exceeds the preset temperature threshold, or the current and voltage show abnormal fluctuations, record the abnormal data. The abnormal data at least includes the time of the abnormality occurrence, the type of the abnormality, and the duration of the abnormality; According to the abnormal data, determine the suspected fault cause through a fault diagnosis model. The fault diagnosis model is obtained through deep learning training based on multiple historical abnormal data and corresponding fault cause annotations; If the suspected fault cause is inaccurate zero-crossing detection, automatically switch to a backup zero-crossing detection strategy according to the corresponding fault type.

8. A zero-crossing detection system, characterized in that, The zero-crossing detection system includes: one or more processors and a memory; the memory is coupled to the one or more processors. The memory is used to store computer program code, and the computer program code includes computer instructions. The one or more processors call the computer instructions to enable the zero-crossing detection system to execute the method according to any one of claims 1-7.

9. A computer-readable storage medium, comprising instructions, characterized in that, When the instruction runs on the zero-crossing detection system, enable the zero-crossing detection system to execute the method according to any one of claims 1-7.

10. A computer program product, characterized in that, When the computer program product runs on the zero-crossing detection system, enable the zero-crossing detection system to execute the method according to any one of claims 1-7.

Images