A narrow linewidth 1200nm semiconductor laser data processing system based on cloud computing

Abstract

The present invention provides a cloud computing-based narrow linewidth 1200nm semiconductor laser data processing system, which relates to the field of data processing technology, including: a laser emission module, which uses a semiconductor laser to emit a laser beam with a narrow linewidth of 1200nm; a data acquisition module, which is connected to the laser emission module, and obtains the state parameters of the laser beam in real time, including power, wavelength, linewidth and working environment data, and pre-processes and digitally converts the signal output by the sensor; a cloud processing platform module, which is used to receive and process the data of the data acquisition module, realize real-time analysis, anomaly detection, trend prediction and remote monitoring of the data, and the analysis results of the cloud processing platform. The present invention realizes remote intelligent monitoring and automatic optimization and adjustment of the laser, and improves the data processing efficiency and the stability of the laser performance.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a narrow linewidth 1200nm semiconductor laser data processing system based on cloud computing. Background Art

[0002] Traditional narrow-linewidth semiconductor lasers mainly rely on local equipment to monitor and control the lasers, which severely limits the monitoring range. Once the laser is deployed far away from the monitoring center, it is difficult to achieve real-time and efficient monitoring and management, which increases the difficulty and cost of operation and maintenance. Since the data needs to be collected locally and then transmitted to the monitoring center by wired or wireless means for analysis and processing, this process is often time-consuming and results in slow response. In application scenarios that require fast response, such as real-time adjustment of optical communication systems or emergency ranging of lidar, this delay may cause serious performance problems.

[0003] The data processing capabilities of local devices are often limited, making it difficult to process large-scale, high-dimensional laser status data. This limits the in-depth analysis and optimization of laser performance, and also affects the accuracy and timeliness of anomaly detection. Traditional technologies require the deployment of a large number of local monitoring devices and the deployment of professional operation and maintenance personnel for regular inspection and maintenance. This not only increases hardware costs, but also increases labor costs, making the overall maintenance cost high. With the expansion of application scale and the increase in the number of lasers, traditional technologies are difficult to expand quickly to meet new needs. Adding new monitoring equipment requires additional investment and time, and may face compatibility and integration issues. There is a lack of intelligent management functions, such as remote parameter adjustment, fault warning, trend prediction, etc. This makes it difficult for operation and maintenance personnel to manage lasers in a refined manner, and also limits the full performance of lasers. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a cloud computing-based narrow-linewidth 1200nm semiconductor laser data processing system to improve the stability and accuracy of the laser, reduce maintenance costs, and improve application efficiency.

[0005] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0006] A narrow linewidth 1200nm semiconductor laser data processing system based on cloud computing, comprising:

[0007] A laser emission module, used to emit a laser beam with a narrow line width of 1200nm using a semiconductor laser;

[0008] The data acquisition module is used to connect with the laser emission module to obtain the state parameters of the laser beam in real time, including power, wavelength, line width and working environment data, and pre-process and digitally convert the signal output by the sensor;

[0009] The cloud processing platform module is used to receive and process the data from the data acquisition module, realize real-time data analysis, anomaly detection, trend prediction and remote monitoring, and the analysis results of the cloud processing platform;

[0010] User interface module, used to remotely access the cloud processing platform, view real-time data, historical records, analysis reports, and remotely adjust laser operating parameters or trigger early warning plans;

[0011] The feedback control system module is used to automatically adjust the working state of the laser emission module according to the analysis results of the cloud processing platform, including adjusting the laser power and stabilizing the wavelength.

[0012] Furthermore, it is connected with the laser emission module to obtain the state parameters of the laser beam in real time, including power, wavelength, line width and working environment data, and pre-process and digitally convert the signal output by the sensor, including:

[0013] Monitor the laser beam in real time and convert the monitored physical quantities, including light intensity and wavelength shift, into electrical signals;

[0014] The signal conditioning unit pre-processes the electrical signal, including signal amplification, signal filtering and conditioning, to obtain a processed electrical signal;

[0015] The processed electrical signal enters the analog-to-digital conversion unit to be converted into a digital signal.

[0016] Furthermore, the data from the data acquisition module is received and processed to achieve real-time data analysis, anomaly detection, trend prediction and remote monitoring. The analysis results of the cloud processing platform include:

[0017] The cloud processing platform receives the laser beam status parameter data transmitted by the data acquisition module in real time, including power, wavelength, line width and working environment data;

[0018] Preprocessing the laser beam state parameter data to obtain processed data;

[0019] Based on the processed data, the cloud processing platform calculates key indicators and statistics in real time, including standard deviation, maximum value, and minimum value, to reflect the current state of the laser beam;

[0020] Use key indicators and statistics to build box plots, analyze real-time data, automatically identify abnormal data points, and trigger alarm solutions to obtain real-time data analysis results;

[0021] Based on real-time data analysis results and historical data, using time series, a forecasting model is established to predict future trends;

[0022] Through real-time data analysis, anomaly detection and trend prediction, the cloud processing platform obtains analysis results, including key indicator statistics, anomaly point identification and future trend prediction.

[0023] Furthermore, key indicators and statistics are used to construct box plots, analyze real-time data, automatically identify abnormal data points, and trigger alarm solutions to obtain real-time data analysis results, including:

[0024] According to the box plot, determine whether the real-time data point falls within the normal range. If the data point falls within the abnormal value area of ​​the box plot or exceeds the range of the box plot, it is identified as an abnormal data point;

[0025] Set the initial parameters of simulated annealing, including initial temperature, cooling rate, and termination temperature, and use real-time data and abnormal data points as input data for simulated annealing;

[0026] By analyzing the box plot and identifying abnormal data points, the objective function value is calculated, and a new solution is generated based on the current solution through perturbation;

[0027] Evaluate whether the data points corresponding to the new solution are closer to the normal range, and compare them with the objective function value of the current solution. According to the acceptance probability of simulated annealing, determine whether to accept the new solution. If the new solution is accepted, update the current solution to the new solution.

[0028] The temperature is lowered according to the set cooling rate, and a new solution is generated by repeated perturbations. The current solution is evaluated and updated until the preset number of iterations is reached. The final solution is output and used as the laser operating parameter adjustment scheme.

[0029] The laser operating parameters are adjusted according to the final solution to obtain real-time data analysis results.

[0030] Furthermore, the calculation formula of the objective function value is: ;

[0031] in, represents the objective function value; A quantity representing the laser beam state parameters; Indicates The weight of each state parameter; Indicates Deviation of state parameters; Indicates The penalty coefficient of the state parameter; Indicates An index of the deviation of a state parameter; Indicates Trend prediction items of state parameters; Indicates The state parameters and The interaction coefficient between the state parameters; Indicates Deviation of state parameters; , Represents an index variable.

[0032] Furthermore, the calculation formula for the acceptance probability is: ;

[0033] in, Indicates the current temperature; Indicates the dynamic adjustment coefficient; Indicates the current iteration number; Indicates the maximum number of iterations; Represents the historical average objective function value difference; Indicates the historical maximum objective function value difference; Represents the difference in objective function value between the new solution and the current solution; represents the adjustment parameter; Represents the threshold parameter.

[0034] Furthermore, based on the real-time data analysis results and historical data, a forecasting model is established using time series to predict future trends, including:

[0035] Integrate real-time data analysis results and historical data, including laser beam power, wavelength, line width and working environment data, as well as abnormal point identification information obtained through box plot analysis, to obtain an integrated data set;

[0036] Arrange the integrated data set in chronological order to construct time series data;

[0037] Determine an exponential smoothing model and use time series data to train the forecasting model. During the training process, the exponential smoothing model learns the time dependency and trend in the data to predict future data.

[0038] Use the exponential smoothing model as the forecasting model and train the exponential smoothing model using time series data;

[0039] During the training process, the exponential smoothing model predicts future data based on the changing patterns of historical data.

[0040] Furthermore, according to the analysis results of the cloud processing platform, the working state of the laser emission module is automatically adjusted, including adjusting the laser power and stabilizing the wavelength, including:

[0041] A set of initial solutions is randomly generated in the solution space of laser power and wavelength, each solution represents a set of set values ​​of laser power and wavelength;

[0042] Based on the analysis results of the cloud processing platform, a fitness function is defined to evaluate the performance of each individual and obtain the fitness evaluation result;

[0043] According to the fitness evaluation results, the individual is determined as the parent, and a crossover operation is performed on the determined parent to generate a new individual by combining the genes of the parent.

[0044] Perform mutation operations on new individuals and repeat the selection, crossover and mutation operations until the preset number of iterations is reached to determine the final parameters;

[0045] Based on the final parameters, the cloud processing platform generates control instructions, including the adjusted laser power and wavelength values, and sends the control instructions to the feedback control system module;

[0046] The feedback control system module receives control instructions from the cloud processing platform and automatically adjusts the laser power and wavelength to the final set values ​​by controlling the power supply and temperature controller components in the laser emission module according to the control instructions.

[0047] Furthermore, the calculation formula of the fitness function is: ;

[0048] in, Representation solution The fitness value of represents the solution vector; Indicates The state parameters are input The actual value of Indicates The target value of each state parameter; Indicates The tolerance range of each state parameter; Indicates the number of trend items; Indicates the index of the trend item; Indicates The weight of the trend item; Indicates Trends in input The actual value of Indicates The target value of a trend.

[0049] Furthermore, according to the fitness evaluation result, an individual is determined as a parent, and a crossover operation is performed on the determined parent to generate a new individual by combining the genes of the parent, including:

[0050] According to the fitness evaluation results, a fitness value list is obtained;

[0051] Based on the fitness value list, use roulette to determine the individual as the parent;

[0052] Pair the parent individuals determined by roulette, and for each pair of paired parents, perform a crossover operation to generate two new individuals

[0053] The above solution of the present invention includes at least the following beneficial effects:

[0054] Through the close connection between the data acquisition module and the laser emission module, the system can monitor the power, wavelength, line width and working environment data of the laser beam in real time to ensure the stable operation of the laser. At the same time, the user interaction interface module enables operation and maintenance personnel to remotely access the cloud processing platform and view real-time data, historical records and analysis reports at any time, realizing remote management of the laser and greatly improving management efficiency and response speed. With its powerful computing power, the cloud processing platform module can efficiently process the large amount of data uploaded by the data acquisition module, and perform real-time analysis, anomaly detection and trend prediction. This not only improves the accuracy and timeliness of data analysis, but also provides strong support for performance optimization and fault warning of the laser.

[0055] The feedback control system module can automatically adjust the working state of the laser emission module according to the analysis results of the cloud processing platform, such as adjusting the laser power and stabilizing the wavelength, to ensure that the laser always maintains the best working state. At the same time, the system can also trigger the early warning plan through the user interaction interface module, and promptly notify the operation and maintenance personnel to deal with potential faults, effectively avoiding system interruption or performance degradation caused by laser failure. Traditional local monitoring equipment requires a lot of manpower and material resources for maintenance, while this system realizes remote monitoring and management through the cloud processing platform, reducing the deployment and maintenance costs of local equipment. At the same time, the intelligent adjustment and early warning functions also reduce the maintenance costs and production stoppage losses caused by failures.

[0056] The cloud computing platform has good scalability and can easily cope with the increase in the number of lasers and the expansion of the application scale. Operation and maintenance personnel can dynamically adjust cloud resources according to actual needs to ensure that the system always meets application requirements. At the same time, the system also supports the access of various types of lasers and sensors, which improves the flexibility and compatibility of the system. The implementation of this system not only improves the performance and stability of the laser, but also provides strong support for the innovation and application expansion of laser technology. Through data analysis and processing on the cloud computing platform, operation and maintenance personnel can deeply explore the potential performance of the laser. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a schematic diagram of a narrow linewidth 1200nm semiconductor laser data processing system based on cloud computing provided by an embodiment of the present invention.

[0058] Figure 2 The present invention provides a cloud computing-based narrow linewidth 1200nm semiconductor laser data processing system for receiving and processing data from a data acquisition module, realizing real-time data analysis, anomaly detection, trend prediction and remote monitoring, and a flow chart of the analysis results of a cloud processing platform. DETAILED DESCRIPTION

[0059] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0060] like Figure 1 As shown, an embodiment of the present invention provides a narrow linewidth 1200nm semiconductor laser data processing system based on cloud computing, comprising:

[0061] The laser emission module 11 is used to emit a laser beam with a narrow line width of 1200 nm using a semiconductor laser;

[0062] The data acquisition module 12 is used to connect with the laser emission module to obtain the state parameters of the laser beam in real time, including power, wavelength, line width and working environment data, and pre-process and digitally convert the signal output by the sensor;

[0063] The cloud processing platform module 13 is used to receive and process the data from the data acquisition module, realize real-time data analysis, anomaly detection, trend prediction and remote monitoring, and the analysis results of the cloud processing platform;

[0064] The user interaction interface module 14 is used to remotely access the cloud processing platform, view real-time data, historical records, analysis reports, and remotely adjust the laser operating parameters or trigger the early warning plan;

[0065] The feedback control system module 15 is used to automatically adjust the working state of the laser emission module according to the analysis results of the cloud processing platform, including adjusting the laser power and stabilizing the wavelength.

[0066] In an embodiment of the present invention, a semiconductor laser is used to emit a laser beam with a narrow line width of 1200nm, which ensures the high purity, high stability and high directivity of the laser and improves the overall performance and reliability of the system. Through close connection with the laser emission module, the data acquisition module can obtain the power, wavelength, line width and working environment data of the laser beam in real time, providing an accurate and reliable basis for subsequent data analysis and processing. At the same time, the preprocessing and digital conversion of the signal improve the transmission efficiency and anti-interference ability of the data. The cloud processing platform can receive and process a large amount of data from the data acquisition module, and realize real-time analysis, anomaly detection, trend prediction and remote monitoring of the data. This not only improves the efficiency and accuracy of data analysis, but also provides a scientific basis for performance optimization and fault warning of the laser. The cloud processing platform can dynamically allocate computing resources according to actual needs to ensure the timeliness and efficiency of data processing. At the same time, the storage and backup functions of the cloud also ensure the security and reliability of the data.

[0067] The user interaction interface module enables operation and maintenance personnel to remotely access the cloud processing platform, view real-time data, historical records and analysis reports at any time, and realize remote monitoring and management of the laser. This convenient operation method improves management efficiency and response speed. Through the user interaction interface, operation and maintenance personnel can remotely adjust the operating parameters of the laser, such as laser power, wavelength, etc., to meet different application requirements. At the same time, when the system detects an abnormality or potential fault, it can also trigger the early warning plan in time to notify the operation and maintenance personnel to handle it. The feedback control system module can automatically adjust the working state of the laser emission module according to the analysis results of the cloud processing platform, such as adjusting the laser power, stabilizing the wavelength, etc. This intelligent adjustment method ensures that the laser always remains in working condition and improves the stability and accuracy of the system.

[0068] Through real-time monitoring and automatic adjustment, the feedback control system module can promptly detect and prevent potential faults. When a fault occurs, the system can also respond quickly and take appropriate measures to repair or warn, reducing the impact and loss of the fault on the system.

[0069] In a preferred embodiment of the present invention, the laser emission module is connected to obtain the state parameters of the laser beam in real time, including power, wavelength, line width and working environment data, and the signal output by the sensor is preprocessed and digitally converted, which may include:

[0070] Monitor the laser beam in real time and convert the monitored physical quantities, including light intensity and wavelength shift, into electrical signals;

[0071] The signal conditioning unit pre-processes the electrical signal, including signal amplification, signal filtering and conditioning, to obtain a processed electrical signal;

[0072] The processed electrical signal enters the analog-to-digital conversion unit to be converted into a digital signal.

[0073] In an embodiment of the present invention, a specially designed sensor (such as a photodiode, an optical power meter, a wavelength meter, etc.) is closely connected to the laser emission module to ensure that the sensor can accurately capture the real-time state of the laser beam. The sensor starts working and monitors the key physical quantities such as the light intensity (i.e., power) and wavelength shift of the laser beam in real time. These physical quantities reflect the performance and working state of the laser beam. The photosensitive element or electronic element inside the sensor generates a corresponding electrical signal according to the monitored physical quantity. For example, a photodiode generates a current signal according to the change in light intensity, while a wavelength meter may measure the wavelength shift through interference or diffraction phenomena and convert it into an electrical signal. Since the electrical signal output by the sensor is usually weak, it needs to be amplified by a signal conditioning unit. In order to remove noise and interference in the electrical signal, the signal conditioning unit will include a filtering circuit. These circuits can be low-pass filters, high-pass filters, or band-pass filters, which are used to retain useful signals and suppress unnecessary frequency components. In addition to amplification and filtering, signal conditioning also includes level conversion, bias adjustment, linearization and other operations to ensure that the electrical signal meets the requirements of the subsequent analog-to-digital conversion unit and optimize the quality of the signal. The processed electrical signal is sent to the analog-to-digital conversion unit. ADC converts continuous analog electrical signals into discrete digital signals, which involves processes such as sampling, quantization and encoding. The converted digital signals are output to other parts of the data acquisition module, such as a microprocessor, FPGA or digital signal processor (DSP) for further data processing, analysis, storage or transmission.

[0074] For example, a highly sensitive photodiode is selected as a sensor to monitor the intensity of a laser beam. When the laser beam hits the photodiode, the photosensitive element inside the diode generates a current signal that is proportional to the light intensity. This current signal is the original electrical signal output by the sensor. Since the current signal output by the photodiode is usually weak (for example, a few microamperes to tens of microamperes), an operational amplifier is required to amplify it. By adjusting the gain of the amplifier, the current signal can be amplified to a range suitable for subsequent processing (for example, a few milliamperes to tens of milliamperes). In order to remove high-frequency noise and interference in the current signal, a low-pass filter can be connected after the amplifier. The filter allows low-frequency signals to pass through and attenuates high-frequency signals, thereby obtaining a smoother and more stable current signal. The processed current signal is sent to a 16-bit ADC for conversion. The ADC samples the current signal at a certain sampling rate and converts the analog value of each sampling point into a corresponding digital value. These digital values ​​represent the light intensity (power) information of the laser beam at different time points. The converted digital signal is output to the microprocessor of the data acquisition module. The microprocessor can further process and analyze these digital signals, such as calculating average power, monitoring power fluctuations, etc., and display the results through a user interface or use them for remote monitoring and control.

[0075] The system can monitor key physical quantities such as the intensity and wavelength deviation of the laser beam in real time, ensuring the immediacy and accuracy of the data. Through high-precision sensors and analog-to-digital conversion units, the system can accurately convert physical quantities into digital signals, improving the measurement accuracy and reliability of the data. The signal conditioning unit amplifies, filters and conditions the original electrical signal, effectively removing noise and interference and enhancing the quality of the signal. After entering the analog-to-digital conversion unit, the pre-processed electrical signal can be stably converted into a digital signal, reducing errors and uncertainties caused by signal fluctuations. The digitized data can be transmitted to a remote monitoring center or cloud platform through the network, allowing operation and maintenance personnel to access and view the status parameters of the laser beam anytime and anywhere, realizing the convenience of remote monitoring and management.

[0076] The data analysis algorithm based on digital signals can monitor the performance changes of the laser beam in real time, predict potential faults, and take measures in advance to prevent or repair them. The system can automatically adjust the working parameters (such as power, wavelength, etc.) of the laser emission module according to the analysis results to achieve the optimal working state and output effect. This intelligent control method improves the automation level and response speed of the system. By simultaneously monitoring multiple state parameters of the laser beam (such as power, wavelength, line width, etc.), the system can have a more comprehensive understanding of the performance and working state of the laser beam, providing strong support for the optimization and upgrading of the system. The long-term stable operation of the system depends on the reliability and stability of each component. By real-time monitoring and preprocessing signals, the system can detect and deal with potential problems in a timely manner, thereby extending the service life of the system and improving overall reliability.

[0077] In a preferred embodiment of the present invention, the data of the data acquisition module is received and processed to realize real-time analysis, anomaly detection, trend prediction and remote monitoring of the data. The analysis results of the cloud processing platform may include:

[0078] The cloud processing platform receives the laser beam status parameter data transmitted by the data acquisition module in real time, including power, wavelength, line width and working environment data;

[0079] Preprocessing the laser beam state parameter data to obtain processed data;

[0080] Based on the processed data, the cloud processing platform calculates key indicators and statistics in real time, including standard deviation, maximum value, and minimum value, to reflect the current state of the laser beam;

[0081] Use key indicators and statistics to build box plots, analyze real-time data, automatically identify abnormal data points, and trigger alarm solutions to obtain real-time data analysis results;

[0082] Based on real-time data analysis results and historical data, using time series, a forecasting model is established to predict future trends;

[0083] Through real-time data analysis, anomaly detection and trend prediction, the cloud processing platform obtains analysis results, including key indicator statistics, anomaly point identification, and future trend prediction.

[0084] In an embodiment of the present invention, the cloud processing platform receives in real time the laser beam state parameter data transmitted by the data acquisition module through the network, including power, wavelength, line width and working environment data (such as temperature, humidity, etc.). The received raw data is cleaned, denoised, normalized and other pre-processing operations are performed to eliminate outliers, missing values ​​and noise in the data, so as to improve the quality and availability of the data. The cloud processing platform calculates the key indicators of the laser beam in real time based on the pre-processed data, such as the average value of power, the stability of wavelength, the range of variation of line width, etc. At the same time, statistics are calculated, including standard deviation, maximum value, minimum value, etc., to fully reflect the current state of the laser beam and its fluctuation. Using key indicators and statistics, a box plot is constructed to visualize the real-time data. Through box plot analysis, data points that exceed the normal range, i.e., abnormal data points, are automatically identified. Once an abnormal data point is identified, the alarm scheme is immediately triggered and an alarm message is sent to relevant personnel so that timely measures can be taken to deal with it. The real-time data analysis results are integrated with historical data to form time series data. By using time series analysis techniques, such as ARIMA model and LSTM neural network, a prediction model is established to predict the future state trend of the laser beam. The accuracy and reliability of the prediction results are ensured by continuously iterating and optimizing the model parameters and verifying the model.

[0085] Outputs statistical results of key indicators of laser beam, such as average power, wavelength stability, etc. Generates anomaly point identification report, including time, value, type, etc. of abnormal data points. Provides future trend forecast report, showing future change trend of laser beam state parameters and possible abnormal risks.

[0086] By receiving and processing data from the data acquisition module in real time, the cloud processing platform can quickly reflect the state changes of the laser beam and improve monitoring efficiency. The real-time calculation of key indicators and statistics provides accurate data support for evaluating the performance of the laser beam. Box plot analysis and outlier recognition technology can timely detect outliers in the laser beam state parameters, trigger alarm schemes, and reduce potential safety hazards. The timely transmission of alarm information ensures that relevant personnel can quickly take measures to deal with abnormalities and ensure the stable operation of the system. By establishing a prediction model, the cloud processing platform can predict the future trend of the laser beam state parameters and provide a basis for decision-making. The accuracy of the prediction results helps to improve the predictability of the system, optimize resource allocation and operation and maintenance strategies. The real-time analysis, anomaly detection and trend prediction functions of the cloud processing platform realize the intelligent monitoring and management of the laser beam state. The visual display and report generation of data improve the readability and ease of use of information and reduce the cost and difficulty of manual monitoring. The cloud processing platform supports remote access and monitoring, allowing operation and maintenance personnel to understand the status of the laser beam anytime and anywhere, improving the flexibility and convenience of monitoring. The data sharing and collaboration functions promote information exchange and cooperation between different departments and improve overall work efficiency.

[0087] When specifically applied, this includes:

[0088] The cloud processing platform receives the laser beam state parameter data and working environment data transmitted by the data acquisition module in real time through the network. The received raw data is preprocessed by cleaning, denoising, normalization and other operations. Outliers and missing values ​​in the data are removed. Filtering algorithms, such as sliding average filtering, are used to remove noise in the data. The data is scaled to a specific range, such as [0, 1] or [−1, 1], to improve the convergence speed and stability of the algorithm. Based on the preprocessed data, the key indicators of the laser beam are calculated in real time. Average power ;in, Indicates the number of data points; Indicates The power value of each data point after preprocessing; wavelength stability ;in, Indicates The wavelength value of the data point after preprocessing; Indicates the average value of wavelength Line width variation range ; Indicates the maximum value of line width; Indicates the minimum value of the line width. At the same time, statistics are calculated, including standard deviation, maximum value, minimum value, etc. Standard deviation: ,in, Indicates The value of the data point after preprocessing; Representation parameters The average value of . Maximum value: ; Represents the set of all preprocessed data points. Minimum: .

[0089] Use key indicators and statistics to construct a box plot to visualize real-time data. The box plot can intuitively display the distribution characteristics of the data, including the median, quartiles, outliers, etc. Through box plot analysis, data points that are beyond the normal range, i.e. abnormal data points, are automatically identified.

[0090] Once an abnormal data point is identified, the alarm scheme is triggered immediately. The real-time data analysis results are integrated with historical data to form time series data. The prediction model is established using time series analysis techniques, such as ARIMA model, LSTM neural network, etc. Let the prediction model be M, the input be time series data T, and the output be the predicted value of the future laser beam state parameter. ARIMA model: ;in, Indicates at a point in time The response variable here refers to the time series data that you want to predict, such as a state parameter of a laser beam; , , represents the autoregression (AR) coefficient, which shows the impact of past values ​​on current values ​​in the time series; represents the order of the autoregressive part; Indicates at a point in time White noise; , , represents the moving average (MA) coefficient; Indicates the order of the sliding average part, that is, the number of past error terms considered in the model. By training the LSTM network, the long-term dependencies of time series data are learned and predicted. The model parameters are optimized through continuous iteration, such as learning rate, number of iterations, etc., and the model is validated, such as cross validation, holdout validation, etc., to ensure the accuracy and reliability of the prediction results. Output the statistical results of key indicators of the laser beam, such as average power , wavelength stability Etc. Generate anomaly point identification report, including the time, type and other information of abnormal data points. Provide future trend prediction report, showing the future change trend of laser beam state parameters and possible abnormal risks. The report can include predicted value, prediction error range, possible abnormal type and probability in the future, etc.

[0091] In another preferred embodiment of the present invention, key indicators and statistics are used to construct a box plot, analyze real-time data, automatically identify abnormal data points, and trigger an alarm scheme to obtain real-time data analysis results, which may include:

[0092] According to the box plot, determine whether the real-time data point falls within the normal range. If the data point falls within the abnormal value area of ​​the box plot or exceeds the range of the box plot, it is identified as an abnormal data point;

[0093] Set the initial parameters of simulated annealing, including initial temperature, cooling rate, and termination temperature, and use real-time data and abnormal data points as input data for simulated annealing;

[0094] By analyzing the box plot and identifying abnormal data points, the objective function value is calculated, and a new solution is generated based on the current solution through perturbation;

[0095] Evaluate whether the data points corresponding to the new solution are closer to the normal range, and compare them with the objective function value of the current solution. According to the acceptance probability of simulated annealing, determine whether to accept the new solution. If the new solution is accepted, update the current solution to the new solution.

[0096] The temperature is lowered according to the set cooling rate, and a new solution is generated by repeated perturbations. The current solution is evaluated and updated until the preset number of iterations is reached. The final solution is output and used as the laser operating parameter adjustment scheme.

[0097] The laser operating parameters are adjusted according to the final solution to obtain real-time data analysis results.

[0098] In an embodiment of the present invention, real-time data of key indicators (such as power, wavelength, line width, etc.) are obtained from the laser monitoring system. Statistics such as the mean, median, standard deviation, and quartiles of the real-time data are calculated to construct a box plot. Based on the calculated statistics, a box plot is drawn, including the upper quartile. , lower quartile , Median, Upper Bound , Nether and outlier regions. Determine whether the real-time data point falls in the outlier region of the box plot or exceeds the range of the box plot. If so, mark it as an outlier data point. Determine the initial temperature, cooling rate, termination temperature and number of iterations of simulated annealing. Use the real-time data and the identified outlier data points as the input data of simulated annealing. According to the real-time data and the outlier data points, calculate the objective function value of the current solution, which can reflect the deviation between the laser operating parameters and the expected state. Perturb the current solution to generate a new solution. The perturbation method can be to randomly adjust one or more values ​​of the laser operating parameters. Calculate whether the data point corresponding to the new solution is closer to the normal range, that is, calculate the objective function value of the new solution and compare it with the objective function value of the current solution. According to the acceptance probability of simulated annealing (related to the current temperature), determine whether to accept the new solution. If the new solution is accepted, update the current solution to the new solution. Reduce the temperature according to the set cooling rate. Repeat the perturbation to generate a new solution until the preset number of iterations is reached or the temperature drops below the termination temperature. After the iteration, output the final solution, which is the optimized laser operating parameter adjustment plan. According to the final solution, adjust the actual operating parameters of the laser, such as power, wavelength, line width, etc. Compare the adjusted laser operating parameters with the real-time data, analyze the adjustment effect, and generate a real-time data analysis report.

[0099] Through box plot analysis, abnormal data points can be identified and processed in time, improving the accuracy and reliability of real-time data. Using simulated annealing algorithm, a better adjustment scheme for laser operating parameters can be found to make the laser performance closer to the expected state. The entire process is highly automated, capable of analyzing data, identifying anomalies, adjusting parameters and outputting results in real time, reducing manual intervention. By continuously optimizing laser operating parameters, the stability and reliability of the laser system can be enhanced and the service life of the equipment can be extended. Real-time data analysis and parameter adjustment can quickly respond to changes in production needs and improve production efficiency and product quality.

[0100] When specifically applied, this includes:

[0101] Obtain real-time data of key indicators (such as power, wavelength, line width, etc.) from the laser monitoring system and record them as data sets. Calculate the statistical mean, median, and standard deviation of the real-time data. The lower quartile is the value of the 25% position after the data set is sorted. The upper quartile is the value of the 75% position after the data set is sorted. Interquartile range . Draw box plots based on the calculated statistics, including:

[0102] Median; lower bound ; Upper bound . Data points that are below the lower bound or above the upper bound. The box plot formula is: ;in, Indicates an outlier. For each real-time data point , determine whether it falls in the outlier area of ​​the box plot or exceeds the range of the box plot: If, ,Right now If it is an abnormal value, it is marked as an abnormal data point. Determine the initial parameters of simulated annealing, initial temperature, cooling rate (values ​​between 0.8 and 0.99), termination temperature (values ​​close to 0), and number of iterations. Use the real-time data and the identified abnormal data points as the input data of simulated annealing, record them as the initial solution, and perturb the current solution to generate a new solution. The perturbation method can be to randomly adjust one or more values ​​of the laser operating parameters. Calculate the objective function value of the new solution. Determine whether to accept the new solution based on the acceptance probability of simulated annealing. The acceptance probability is related to the current temperature. It is related to the change of the objective function value. The temperature is reduced according to the set cooling rate. ;in, Represents the cooling rate. Repeat the steps to generate new solutions by repeated perturbations until the preset number of iterations is reached or the temperature drops below the termination temperature. After the iteration is completed, the final solution is output, which is the optimized laser operating parameter adjustment plan. According to the final solution, adjust the actual operating parameters of the laser, such as power, wavelength, line width, etc. Compare the adjusted laser operating parameters with the real-time data, analyze the adjustment effect, and generate a real-time data analysis report. The report can include parameter comparison before and after adjustment, changes in the objective function value, and the handling of abnormal data points.

[0103] In another preferred embodiment of the present invention, the calculation formula of the objective function value is: ;

[0104] in, represents the objective function value; A quantity representing the laser beam state parameters; Indicates The weight of each state parameter; Indicates Deviation of state parameters; Indicates The penalty coefficient of the state parameter; Indicates An index of the deviation of a state parameter; Indicates Trend prediction items of state parameters; Indicates The state parameters and The interaction coefficient between the state parameters; Indicates Deviation of state parameters; , Represents an index variable.

[0105] In the embodiment of the present invention, it is necessary to determine the specific values ​​of all parameters involved in the objective function. This includes: the number of laser beam state parameters The weight of each state parameter The deviation of each state parameter , which is calculated by the difference between the real-time measured value and the expected value. The penalty coefficient for each state parameter , used to impose additional penalties on deviations. The exponential of the deviation of each state parameter , used to adjust the impact of the deviation. Trend prediction item for each state parameter , which is the future trend predicted based on historical data. The interaction coefficient between state parameters , used to describe the interaction between different parameters. Initialize a variable to store the objective function value , the initial value is set to 0. For each state parameter (from 1 to ) Calculate the deviation term Calculate the penalty term , which reflects the degree of penalty for deviation. Plus the trend prediction term Initialize a variable to store the contribution of the interaction term, with the initial value set to 0. For the current state parameter , iterate over all other state parameters (from 1 to , but skip ), calculate the interaction term , , This reflects the state parameter and The contribution of the interaction term is added to the previously initialized variable. The contribution of a single state parameter (bias term, penalty term, trend prediction term) and the contribution of the interaction term are added to obtain the total contribution of the current state parameter to the objective function value. The total contribution of the current state parameter is multiplied by its weight. , and then accumulated to the objective function value Repeat the perturbation step to generate a new solution until all state parameters are calculated. Finally, the objective function value is It is the weighted sum of all state parameter contributions. This value reflects the overall deviation between the laser beam state parameters and the expected state.

[0106] Bias term in the objective function It can directly reflect the degree of deviation between each laser beam state parameter and the expected value. Through weighted summation, the overall deviation can be accurately quantified. Penalty coefficient and index , so that the degree of penalty for deviation can be flexibly adjusted according to actual needs. For some key parameters or parameters that need to strictly control deviation, the penalty can be increased by increasing the penalty coefficient or adjusting the index to ensure that these parameters can better approach the expected value. The addition of makes the objective function consider not only the deviation of the current state parameters, but also the future trend changes. This helps to achieve forward-looking control, that is, to adjust the laser beam state parameters in advance according to the predicted future trends, so as to better meet future needs. Interaction coefficient The introduction of enables the objective function to consider the mutual influence between different state parameters. In practical applications, there are complex interactions between different state parameters of the laser beam. By considering these interactions, the overall performance of the laser beam can be more comprehensively evaluated and optimized. The objective function combines the deviations, penalties, trend predictions and interactions of multiple state parameters to form a unified optimization goal. This makes the optimization process clearer and more efficient, and various optimization algorithms (such as gradient descent, genetic algorithms, etc.) can be used to quickly find the optimal solution, thereby improving the optimization efficiency. The various parameters in the objective function (such as weights, penalty coefficients, deviation index, etc.) can be customized according to specific application scenarios and requirements. This makes the objective function very flexible and adaptable, and can be applied to different types of laser beam state parameter optimization problems.

[0107] In another preferred embodiment of the present invention, the calculation formula of the acceptance probability is: ;

[0108] in, Indicates the current temperature; Indicates the dynamic adjustment coefficient; Indicates the current iteration number; Indicates the maximum number of iterations; Represents the historical average objective function value difference; Indicates the historical maximum objective function value difference; Represents the difference in objective function value between the new solution and the current solution; represents the adjustment parameter; Represents the threshold parameter.

[0109] In the embodiment of the present invention, the specific values ​​of all parameters in the formula are determined. These parameters include: the current temperature , dynamic adjustment coefficient , the current number of iterations , maximum number of iterations , the historical average objective function value difference , the historical maximum objective function value difference The difference in objective function value between the new solution and the current solution , adjustment parameters , threshold parameter .calculate , which represents the ratio of the current number of iterations to the maximum number of iterations. , which represents the ratio of the historical average objective function value difference to the historical maximum objective function value difference. , which represents the absolute value of the difference in the objective function value between the new solution and the current solution.

[0110] calculate , which represents the sum of the absolute value of the objective function value difference and the threshold parameter. Combining the intermediate terms into the formula, we can calculate , which is an adjustment coefficient used to dynamically adjust the denominator of the acceptance probability. , which is a tuning term used to further adjust the acceptance probability.

[0111] Substituting the combination into the formula for acceptance probability, we get The value is a probability value between 0 and 1, indicating the probability of accepting the new solution. This probability value is affected by many factors, including the current temperature, the number of iterations, the difference in historical objective function values, the difference between the new solution and the current solution, and the tuning parameters and threshold parameters.

[0112] when When is smaller (i.e., the difference between the new solution and the current solution is smaller), the acceptance probability The larger the temperature, the more likely it is to accept new solutions. When it is higher, the probability of acceptance will also be larger, because high temperature means that the system is more willing to explore new solutions. As the number of iterations increases ( As the value of θ increases, the acceptance probability will gradually decrease, because the system gradually converges to the final solution. The difference in the historical average objective function value and the difference in the historical maximum objective function value will also affect the acceptance probability, reflecting the system's use of historical information. and threshold parameters Additional controls are provided for fine-tuning the calculation of the acceptance probability.

[0113] temperature , Dynamic adjustment coefficient , current iteration number and the maximum number of iterations The two functions work together to calculate the acceptance probability, so that the acceptance probability can be adjusted dynamically as the iteration process proceeds. This helps to maintain a high acceptance probability in the early stage of optimization to explore more solution space, and gradually reduce the acceptance probability in the later stage of optimization to converge to the final solution. Difference from the historical maximum objective function value , the formula can take into account information from past iterations. This helps the algorithm adjust the acceptance probability based on historical performance, thereby better guiding the subsequent search process. The difference in the objective function value between the new solution and the current solution Directly affects the calculation of the acceptance probability. When the new solution is obviously better than the current solution (i.e. is smaller), the acceptance probability will increase accordingly, making it easier to accept this new solution. and threshold parameters Provides fine control over the acceptance probability. By adjusting these two parameters, the calculation method of the acceptance probability can be further fine-tuned to meet the needs of different optimization problems. The calculation formula of the acceptance probability strikes a balance between exploration and exploitation. In the early stage of optimization, a higher acceptance probability encourages the algorithm to explore more solution space; while in the later stage of optimization, a lower acceptance probability prompts the algorithm to converge to the final solution.

[0114] In a preferred embodiment of the present invention, remote access to the cloud processing platform to view real-time data, historical records, analysis reports, and remotely adjust laser operating parameters or trigger early warning schemes may include:

[0115] Identify the type of real-time data that needs to be accessed, the scope of historical records, the format of the analysis report, and the laser operating parameters that need to be adjusted. Select a cloud processing platform and make sure it supports remote access, data storage, data analysis, and device control functions. Make sure you have a device that can remotely access the cloud, such as a computer, tablet, or smartphone, and install the necessary access software or applications. Log in to the control panel or management interface of the cloud processing platform using your credentials (such as username and password). Find the entry or tab for real-time data, historical records, and analysis reports on the control panel and click to enter. In the real-time data area, you should be able to see real-time information such as the current working status, parameter settings, output power, etc. of the laser. In the history record area, find and view past data records based on time range, laser number, or other filtering conditions. History records may include laser operation logs, fault records, maintenance records, etc. In the analysis report area, you can find analysis reports automatically generated by the platform or customized by the user.

[0116] These reports include laser performance analysis, energy efficiency analysis, fault prediction, etc. Find the entry or tab for laser parameter adjustment on the control panel and click to enter. In the parameter adjustment interface, you can see the current parameter settings of the laser. Find the entry or tab for warning settings on the control panel and click to enter. According to the operating status and parameters of the laser, set the conditions for triggering warnings, such as over-temperature, abnormal power, etc.

[0117] After the setup is complete, perform an early warning test to ensure that the early warning scheme works properly. Trigger a simulated early warning condition and check whether you can receive the expected early warning information. After completing all operations, make sure your changes have been saved and exit the cloud platform control panel safely. Regularly monitor the operating status and parameters of the laser, as well as the real-time data and analysis reports provided by the cloud platform.

[0118] In a preferred embodiment of the present invention, based on the real-time data analysis results and historical data, a prediction model is established using time series to predict future trends, which may include:

[0119] Integrate real-time data analysis results and historical data, including laser beam power, wavelength, line width and working environment data, as well as abnormal point identification information obtained through box plot analysis, to obtain an integrated data set;

[0120] Arrange the integrated data set in chronological order to construct time series data;

[0121] Determine an exponential smoothing model and use time series data to train the forecasting model. During the training process, the exponential smoothing model learns the time dependency and trend in the data to predict future data.

[0122] Use the exponential smoothing model as the forecasting model and train the exponential smoothing model using time series data;

[0123] During the training process, the exponential smoothing model predicts future data based on the changing patterns of historical data.

[0124] In an embodiment of the present invention, the power data of the laser beam is collected in real time from the laser monitoring system, and the corresponding timestamp is recorded. The historical power data is extracted from the database or archive, which also includes the timestamp.

[0125] Use methods such as box plot analysis to identify outliers in the data. Outliers may be caused by equipment failure, measurement errors, or special events. For outliers, you can choose to remove, correct, or retain (if they represent real but rare events). Integrate real-time data and historical data into a data set, ensuring that the data is arranged in chronological order. The data set should contain timestamps, laser beam power, and possible other relevant parameters (such as wavelength, line width, working environment data, etc.). Arrange the integrated data set in chronological order to form time series data. Each time point corresponds to a record of one or more laser beam parameters. Select a suitable exponential smoothing model based on the characteristics of the data and the prediction requirements. For the prediction of laser beam power, a simple exponential smoothing model may be sufficient. If there is a trend in the data, the Holt linear trend model can be considered.

[0126] The exponential smoothing model is trained using the time series data. During the training process, the model learns the time dependencies and trends in the data. For the simple exponential smoothing model, the smoothing coefficient needs to be determined, which determines how much the historical data affects the current forecast. By trying different values, the value that minimizes the forecast error is selected. Use a portion of the data (such as the last few days of data) as a validation set to evaluate the forecast performance of the model. The model can be evaluated by calculating the error between the predicted value and the actual value (such as mean square error, mean absolute error). Use the trained exponential smoothing model to predict future data. Depending on the type and parameters of the model, the model will predict the laser beam power for a period of time in the future based on the historical data and trends.

[0127] Consider a time series of laser beam power in watts at hourly intervals, which has been aggregated and preprocessed.

[0128]

[0129] Select the simple exponential smoothing model for prediction and determine the smoothing coefficient α=0.5 (this value is obtained through trial and verification). Use the trained model to predict the laser beam power in the next 24 hours. When predicting, the model will calculate the predicted value for each future time point based on the trend and smoothing coefficient in the historical data. The prediction results are:

[0130]

[0131] Through this process, the future laser beam power can be predicted based on time series data and exponential smoothing models. The prediction results can help understand the operating status and trends of the laser and provide decision support for equipment maintenance, performance optimization or production planning.

[0132] By integrating real-time data and historical data, the model can more comprehensively understand the changing trends of the power, wavelength, line width and working environment data of the laser beam. The addition of abnormal point identification information helps the model to capture abnormal fluctuations in the data more accurately. The exponential smoothing model can more accurately predict future data changes by learning the time dependence and trends in the data. Accurate prediction results can help companies allocate resources more reasonably, such as adjusting the operating parameters of the laser, arranging maintenance plans, optimizing production plans, etc. This helps to reduce resource waste and improve resource utilization efficiency, thereby reducing costs. The future trend information provided by the prediction model can provide decision makers with strong data support to help them make decisions faster and more accurately. This helps to shorten the decision-making cycle, improve decision-making efficiency, and enable companies to respond to market changes more quickly. Through the prediction model, companies can identify potential risks and opportunities in advance, such as laser performance degradation, changes in market demand, etc. This helps companies take timely measures to deal with risks or seize opportunities, thereby enhancing risk management capabilities. The prediction model can not only be used to predict future trends, but also to analyze historical data to find out the key factors affecting laser beam performance. This helps companies make continuous improvements on these factors, enhance the performance and stability of lasers, and thus improve product quality and customer satisfaction.

[0133] In a preferred embodiment of the present invention, according to the analysis results of the cloud processing platform, the working state of the laser emission module is automatically adjusted, including adjusting the laser power and stabilizing the wavelength, which may include:

[0134] A set of initial solutions is randomly generated in the solution space of laser power and wavelength, each solution represents a set of set values ​​of laser power and wavelength;

[0135] Based on the analysis results of the cloud processing platform, a fitness function is defined to evaluate the performance of each individual and obtain the fitness evaluation result;

[0136] According to the fitness evaluation results, the individual is determined as the parent, and a crossover operation is performed on the determined parent to generate a new individual by combining the genes of the parent.

[0137] Perform mutation operations on new individuals and repeat the selection, crossover and mutation operations until the preset number of iterations is reached to determine the final parameters;

[0138] Based on the final parameters, the cloud processing platform generates control instructions, including the adjusted laser power and wavelength values, and sends the control instructions to the feedback control system module;

[0139] The feedback control system module receives control instructions from the cloud processing platform and automatically adjusts the laser power and wavelength to the final set values ​​by controlling the power supply and temperature controller components in the laser emission module according to the control instructions.

[0140] In an embodiment of the present invention, a set of initial solutions is randomly generated in the solution space of laser power and wavelength. These solutions can be expressed in vector form, each vector containing the set values ​​of laser power and wavelength. According to the analysis results of the cloud processing platform, a fitness function is defined to evaluate the performance of each individual. This function can be constructed based on multiple indicators such as laser cutting accuracy, speed, energy consumption, etc. For example, the fitness function can be defined as the deviation between the cutting accuracy and the ideal accuracy, the gap between the cutting speed and the target speed, and the weighted sum of energy consumption. Each initial solution is evaluated using the fitness function to obtain a fitness evaluation result. According to the fitness evaluation result, an individual is selected as a parent. A crossover operation is performed on the determined parent generation, and a new individual is generated by combining the genes of the parent generation (i.e., the set values ​​of laser power and wavelength). The crossover operation can be performed in the form of single-point crossover, multi-point crossover, or uniform crossover. A mutation operation is performed on the newly generated individual to introduce diversity. The mutation operation can randomly adjust the set values ​​of the laser power or wavelength, or adjust both at the same time. The selection, crossover, and mutation operations are repeated until a preset number of iterations is reached or other stop conditions are met (such as the fitness reaching a certain threshold). In each iteration, the fitness function is used to evaluate new individuals, and individuals with excellent performance are selected as the parents of the next iteration. After multiple iterations, the final parameters, that is, the final laser power and wavelength setting values, are determined. The cloud processing platform generates control instructions based on the final parameters, including the adjusted laser power and wavelength values. The cloud processing platform sends the control instructions to the feedback control system module. The feedback control system module receives the control instructions and automatically adjusts the laser power and wavelength to the final set values ​​by controlling the power supply and temperature controller components in the laser emission module according to the instructions.

[0141] By automatically adjusting the laser power and wavelength, it can be ensured that the laser cutting machine can achieve the best cutting accuracy and efficiency for different materials and thicknesses. This helps to reduce the scrap rate and improve production efficiency. Optimizing the laser power and wavelength can reduce energy consumption during the cutting process. This not only helps to save resources, but also reduces operating costs. The automatic adjustment function enables the laser cutting machine to adapt to the needs of different cutting tasks more quickly. This enhances the flexibility and versatility of the system and improves the utilization rate of the equipment. The automated adjustment process reduces the need for manual intervention and reduces the labor intensity of operators. At the same time, this also reduces cutting quality problems caused by human operating errors. By improving cutting accuracy, efficiency and reducing energy consumption, the overall performance of the laser cutting machine has been improved. This helps companies gain a better competitive advantage in the market and enhance the added value of their products.

[0142] When specifically applied, this includes:

[0143] Define the feasible range of laser power and wavelength, which constitutes our solution space. For example, the laser power may be between 10W and 100W, and the wavelength may be between 1060nm and 1070nm. Randomly generate a set of initial solutions in the solution space, each of which is a two-dimensional vector. Define fitness function: Based on the analysis results provided by the cloud processing platform, define a fitness function that comprehensively considers indicators such as laser cutting accuracy, speed, and energy consumption. Use the fitness function to evaluate each initial solution and obtain the fitness value of each solution. According to the fitness value, select individuals as parents. Perform crossover operations on the selected parents to generate new individuals. The crossover operation can be performed in the form of single-point crossover, multi-point crossover, or uniform crossover. For example, a single-point crossover may be to exchange the second half of two parent vectors at a randomly selected point to generate two new offspring. Perform mutation operations on the newly generated individuals to introduce diversity. Mutation can be to randomly fine-tune the set value of laser power or wavelength, or to adjust both at the same time. Repeat the selection, crossover, and mutation operations until the preset number of iterations is reached. In each iteration, the fitness function is used to evaluate new individuals, and individuals with excellent performance are selected as the parents of the next iteration. After multiple iterations, the setting values ​​of laser power and wavelength are used as the final parameters according to the individual fitness values ​​from the final population. The cloud processing platform generates control instructions based on the final parameters, which contain the laser power and wavelength values ​​that need to be adjusted. The cloud processing platform sends the control instructions to the feedback control system module through the communication network.

[0144] The feedback control system module receives control instructions from the cloud processing platform. According to the laser power and wavelength values ​​in the instructions, the feedback control system module automatically adjusts the laser power and wavelength to the final set values ​​by controlling the power supply and temperature controller components in the laser emission module. When the laser power and wavelength are adjusted to the final set values, the laser cutting machine can work according to the final parameters to achieve high-precision and high-efficiency cutting.

[0145] In another preferred embodiment of the present invention, the calculation formula of the fitness function is: ;

[0146] in, Representation solution The fitness value of represents the solution vector; Indicates The state parameters are input The actual value of Indicates The target value of each state parameter; Indicates The tolerance range of each state parameter; Indicates the number of trend items; Indicates the index of the trend item; Indicates The weight of the trend item; Indicates Trends in input The actual value of Indicates The target value of a trend.

[0147] In the embodiment of the present invention, it is ensured that the necessary parameters are available: the solution vector , the number of state parameters For each state parameter , actual value , target value , Tolerance range , weight , the number of trend items For each trend item : Actual value , target value , weight For each state parameter (from 1 to ), calculate the actual value With target value The difference, that is . Divide the difference by the tolerance , and get the normalized deviation . Squaring the normalized deviation yields . Multiply the squared deviation by the weight , and obtain the weighted deviation The weighted deviations of all state parameters are summed to obtain the deviation sum of the state parameters. (from 1 to ), calculate the actual value With target value The absolute value of the difference, that is . Multiply the absolute difference by the weight , and obtain the weighted deviation The weighted deviations of all trend items are summed to obtain the deviation sum of trend items. Add the deviation sum of the state parameters and the deviation sum of the trend term to obtain the total deviation sum. Finally, calculate the fitness value The calculated fitness value As output. This value represents the solution vector The larger the value, the closer the solution is to the ideal state.

[0148] This function can comprehensively consider multiple state parameters and trend terms to ensure that all relevant factors are taken into account when evaluating the quality of the solution. and The importance of each state parameter and trend item in the evaluation can be flexibly adjusted by the setting of The parameters of different dimensions or ranges are normalized in the form of , making them comparable. This helps to ensure that the contributions of various state parameters in the fitness function are balanced, and avoids excessive influence of some parameters on the fitness value due to too large or too small ranges. The deviation is calculated in the form of the sum of squares and absolute values, which not only takes into account the size of the deviation, but also ensures the accuracy of the calculation.

[0149] In another preferred embodiment of the present invention, according to the fitness evaluation result, an individual is determined as a parent, and a crossover operation is performed on the determined parent to generate a new individual by combining the genes of the parent, including:

[0150] According to the fitness evaluation results, a fitness value list is obtained;

[0151] Based on the fitness value list, use roulette to determine the individual as the parent;

[0152] The parent individuals determined by roulette are paired, and for each pair of paired parents, a crossover operation is performed to generate two new individuals.

[0153] In an embodiment of the present invention, a set or list containing all individuals and their fitness values ​​is used. This set is traversed, the fitness value of each individual is extracted, a new fitness value list is formed, and a fitness value list is obtained, in which each element corresponds to the fitness value of an individual. According to the fitness value list, an individual is determined as a parent using a roulette wheel. According to the fitness value list, the sum of the fitness values ​​of all individuals is calculated, and a random number between 0 and the fitness sum is generated. Accumulation starts from the beginning of the fitness value list until the accumulated fitness value exceeds this random number, and the individual corresponding to this time is selected as the parent, and finally a parent individual list is obtained. The parent individuals determined by the roulette wheel are paired, and for each pair of paired parents, a crossover operation is performed to generate two new individuals. According to the parent individual list, the individuals in the parent individual list are randomly paired. For each pair of paired parent individuals, a crossover operation is performed, and one or more crossover points (random) are selected. The genes of the parent individual are divided into two parts at the crossover point. Exchange the split gene parts to form two new individuals and obtain a new list of individuals, where each new individual is generated by a crossover operation between two parent individuals.

[0154] Selecting parents through roulette wheel ensures that individuals with higher fitness have a greater chance of being selected, but at the same time leaves the possibility of being selected for individuals with lower fitness, thus maintaining the diversity of the population.

[0155] The crossover operation further increases the diversity of new individuals by combining genes from different parents, which helps to explore a wider solution space. The roulette selection mechanism makes individuals with higher fitness more likely to become parents, which helps to pass on excellent genes to the next generation, thereby accelerating the convergence of the population. The crossover operation has the potential to produce individuals that are better than the parents by combining excellent gene fragments, further promoting the evolution of the population. Since the roulette selection mechanism retains the possibility of selecting individuals with lower fitness and the randomness of the crossover operation, it helps to avoid the population from falling into a local solution too early, thereby increasing the chances of finding a global final solution. By adjusting parameters such as the probability distribution of roulette selection and the number of crossover points, it can easily adapt to optimization problems of different scales and complexities.

[0156] When specifically applied, this includes:

[0157] A collection or list containing all individuals and their fitness values. This collection may be an array, list, dictionary or other data structure, where each element contains an individual and its fitness value.

[0158] Traverse this collection, extract the fitness value for each element, and store these values ​​in a new list to obtain a list containing only fitness values, where each element in the list corresponds to an individual in the original collection. Sum all the values ​​in the fitness value list to get the sum of the fitness of all individuals. Use a random number generator to generate a floating point number between 0 and the sum of the fitness values. Initialize an accumulator to 0. Starting from the beginning of the fitness value list, accumulate the fitness values ​​into the accumulator one by one. When the value of the accumulator exceeds the generated random number for the first time, select the current corresponding individual as the parent.

[0159] Repeat the accumulation of values ​​from the beginning of the fitness value list, and generate a new random number between 0 and the sum of fitness each time. When the accumulated value exceeds this random number, the current individual is selected as the parent until the required number of parent individuals are selected, and a list containing the selected parent individuals is obtained. Use the random pairing method, that is, randomly shuffle the parent individual list, and then take out two individuals as a pair in turn. For each pair of paired parent individuals, determine one or more crossover points. The crossover point is usually selected randomly, but it can also be designed according to the specific problem.

[0160] At the crossover point, the genes of the parent individuals (or the components of the solution) are split into two parts. The split gene parts are exchanged to form two new individuals. The genes of the new individuals combine part of the genes of the parent individuals to obtain a new list of individuals, where each new individual is generated by the crossover operation of two parent individuals.

[0161] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A cloud computing-based narrow linewidth 1200nm semiconductor laser data processing system, characterized in that: include: A laser emission module, used to emit a laser beam with a narrow line width of 1200nm using a semiconductor laser; The data acquisition module is used to connect with the laser emission module to obtain the state parameters of the laser beam in real time, including power, wavelength, line width and working environment data, and pre-process and digitally convert the signal output by the sensor; The cloud processing platform module is used to receive and process the data from the data acquisition module, realize real-time data analysis, anomaly detection, trend prediction and remote monitoring, and the analysis results of the cloud processing platform; User interface module, used to remotely access the cloud processing platform, view real-time data, historical records, analysis reports, and remotely adjust laser operating parameters or trigger early warning plans; Feedback control system module, used to automatically adjust the working state of the laser emission module according to the analysis results of the cloud processing platform, including adjusting the laser power and stabilizing the wavelength; Use key indicators and statistics to build box plots, analyze real-time data, automatically identify abnormal data points, and trigger alarm solutions to obtain real-time data analysis results, including: According to the box plot, determine whether the real-time data point falls within the normal range. If the data point falls within the abnormal value area of ​​the box plot or exceeds the range of the box plot, it is identified as an abnormal data point; Set the initial parameters of simulated annealing, including initial temperature, cooling rate, and termination temperature, and use real-time data and abnormal data points as input data for simulated annealing; By analyzing the box plot and identifying abnormal data points, the objective function value is calculated, and a new solution is generated based on the current solution through perturbation; Evaluate whether the data points corresponding to the new solution are closer to the normal range, and compare them with the objective function value of the current solution. According to the acceptance probability of simulated annealing, determine whether to accept the new solution. If the new solution is accepted, update the current solution to the new solution. The temperature is lowered according to the set cooling rate, and a new solution is generated by repeated perturbations. The current solution is evaluated and updated until the preset number of iterations is reached. The final solution is output and used as the laser operating parameter adjustment scheme. Adjust the laser operating parameters according to the final solution to obtain real-time data analysis results; The calculation formula of the objective function value is: ； in, represents the objective function value; A quantity representing the laser beam state parameters; Indicates The weight of each state parameter; Indicates Deviation of state parameters; Indicates The penalty coefficient of the state parameter; Indicates An index of the deviation of a state parameter; Indicates Trend prediction items of state parameters; Indicates The state parameters and The interaction coefficient between the state parameters; Indicates Deviation of state parameters; , Represents an index variable; The formula for calculating the probability of acceptance is: ； in, Indicates the current temperature; Indicates the dynamic adjustment coefficient; Indicates the current iteration number; Indicates the maximum number of iterations; Represents the historical average objective function value difference; Indicates the historical maximum objective function value difference; Represents the difference in objective function value between the new solution and the current solution; represents the adjustment parameter; Represents the threshold parameter.

2. The cloud computing-based narrow linewidth 1200nm semiconductor laser data processing system according to claim 1, characterized in that: Connect with the laser emission module to obtain the state parameters of the laser beam in real time, including power, wavelength, line width and working environment data, and pre-process and digitally convert the signal output by the sensor, including: Monitor the laser beam in real time and convert the monitored physical quantities, including light intensity and wavelength shift, into electrical signals; The signal conditioning unit pre-processes the electrical signal, including signal amplification, signal filtering and conditioning, to obtain a processed electrical signal; The processed electrical signal enters the analog-to-digital conversion unit to be converted into a digital signal.

3. The cloud computing-based narrow linewidth 1200nm semiconductor laser data processing system according to claim 2, characterized in that: Receive and process data from the data acquisition module to achieve real-time data analysis, anomaly detection, trend prediction and remote monitoring. The analysis results of the cloud processing platform include: The cloud processing platform receives the laser beam status parameter data transmitted by the data acquisition module in real time, including power, wavelength, line width and working environment data; Preprocessing the laser beam state parameter data to obtain processed data; Based on the processed data, the cloud processing platform calculates key indicators and statistics in real time, including standard deviation, maximum value, and minimum value, to reflect the current state of the laser beam; Use key indicators and statistics to build box plots, analyze real-time data, automatically identify abnormal data points, and trigger alarm solutions to obtain real-time data analysis results; Based on real-time data analysis results and historical data, using time series, a forecasting model is established to predict future trends; Through real-time data analysis, anomaly detection and trend prediction, the cloud processing platform obtains analysis results, including key indicator statistics, anomaly point identification and future trend prediction.

4. The cloud computing-based narrow linewidth 1200nm semiconductor laser data processing system according to claim 3, characterized in that: Based on real-time data analysis results and historical data, using time series, a forecasting model is established to predict future trends, including: Integrate real-time data analysis results and historical data, including laser beam power, wavelength, line width and working environment data, as well as abnormal point identification information obtained through box plot analysis, to obtain an integrated data set; Arrange the integrated data set in chronological order to construct time series data; Determine an exponential smoothing model and use time series data to train the forecasting model. During the training process, the exponential smoothing model learns the time dependency and trend in the data to predict future data. Use the exponential smoothing model as the forecasting model and train the exponential smoothing model using time series data; During the training process, the exponential smoothing model predicts future data based on the changing patterns of historical data.

5. The cloud computing-based narrow linewidth 1200nm semiconductor laser data processing system according to claim 4, characterized in that: According to the analysis results of the cloud processing platform, the working state of the laser emission module is automatically adjusted, including adjusting the laser power and stabilizing the wavelength, including: A set of initial solutions is randomly generated in the solution space of laser power and wavelength, each solution represents a set of set values ​​of laser power and wavelength; Based on the analysis results of the cloud processing platform, a fitness function is defined to evaluate the performance of each individual and obtain the fitness evaluation result; According to the fitness evaluation results, the individual is determined as the parent, and a crossover operation is performed on the determined parent to generate a new individual by combining the genes of the parent. Perform mutation operations on new individuals and repeat the selection, crossover and mutation operations until the preset number of iterations is reached to determine the final parameters; Based on the final parameters, the cloud processing platform generates control instructions, including the adjusted laser power and wavelength values, and sends the control instructions to the feedback control system module; The feedback control system module receives control instructions from the cloud processing platform and automatically adjusts the laser power and wavelength to the final set values ​​by controlling the power supply and temperature controller components in the laser emission module according to the control instructions.

6. The cloud computing-based narrow linewidth 1200nm semiconductor laser data processing system according to claim 5, characterized in that: The calculation formula of the fitness function is: ； in, Representation solution The fitness value of represents the solution vector; Indicates The state parameters are input The actual value of Indicates The target value of each state parameter; Indicates The tolerance range of each state parameter; Indicates the number of trend items; Indicates the index of the trend item; Indicates The weight of the trend item; Indicates Trends in input The actual value of Indicates The target value of a trend.

7. The cloud computing-based narrow linewidth 1200nm semiconductor laser data processing system according to claim 6, characterized in that: According to the fitness evaluation results, the individual is determined as the parent, and a crossover operation is performed on the determined parent to generate a new individual by combining the genes of the parent, including: According to the fitness evaluation results, a fitness value list is obtained; Based on the fitness value list, use roulette to determine the individual as the parent; The parent individuals determined by roulette are paired, and for each pair of paired parents, a crossover operation is performed to generate two new individuals.

Images