Laser packaging system and intelligent heat dissipation method therefor

Abstract

A laser packaging system and an intelligent heat dissipation method therefor. The system comprises: a laser array module (1) having a plurality of lasers (101) arranged at intervals, each laser (101) being provided with a heat sink (102), and the heat sink (102) being provided with a thermoelectric cooler (TEC) (103); a training module (2) used for training and verifying a deep learning model on the basis of an actually measured temperature of the laser (101) within a certain time period and historical operation state data, so as to output a trained deep learning model; a prediction module (3) used for inputting a measured first parameter and second parameter of the laser (101) into the trained deep learning model to generate a predicted temperature, the first parameter comprising power, and the second parameter comprising an operation time; and a temperature adjustment module (4) used for controlling, on the basis of the predicted temperature, the TEC (103) to adjust the temperature of the laser (101). The modules work in conjunction to predict the temperature of each laser (101) so as to drive the TEC (103) to dissipate heat of the laser. The whole system has a relatively simple structure, and no excessive sensors need to be provided, thereby greatly reducing the packaging difficulty and cost.

Description

Laser packaging system and intelligent heat dissipation method thereof Technical Field

[0001] The present application relates to the field of laser packaging technology, and in particular to a laser packaging system and an intelligent heat dissipation method thereof. Background Art

[0002] Semiconductor laser arrays are also called laser diode bars, or simply bars. Depending on the power requirements, a few to nearly a thousand laser diodes are grown on the same substrate using metal organic chemical vapor deposition (MOCVD) to form a one-dimensional array. If the multi-element array is divided into groups, each with 19 units, and the groups are separated by optical isolation technology, the light emitted by each group is incoherent, forming multi-aperture lasing, and the output power can be increased by increasing the number of integrations. However, semiconductor laser arrays have high power and their temperature will rise over long periods of operation. As the junction temperature rises, the wavelength of the semiconductor laser array will broaden, the threshold current will increase, the photoelectric conversion efficiency will decrease, the life span will decrease, and the reliability will decrease. Therefore, semiconductor laser packaging and heat dissipation technology are particularly important.

[0003] However, existing methods all detect the temperature rise of the laser by arranging a large number of sensors in the semiconductor laser array. The large number of sensors increases the difficulty of laser packaging, and the arrangement of a large number of sensors also increases the packaging cost.

[0004] Secondly, existing semiconductor laser arrays use water cooling for heat dissipation, but small water-cooling modules have poor heat dissipation, seriously affecting the normal operation of the semiconductor laser array. Large water-cooling modules are extremely difficult to package and are also not conducive to the use of semiconductor laser arrays. Furthermore, water-cooling modules also have risks such as micro-leakage and water leakage, and their safety and reliability are low. If the water-cooling module stops working, the entire semiconductor laser array will also stop working.

[0005] Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, the present application provides a laser packaging system and a heat dissipation method thereof.

[0007] The specific technical solutions are as follows:

[0008] Laser packaging system, including:

[0009] A laser array module having a plurality of lasers spaced apart and distributed in an array, each laser having a heat sink for conducting heat, and each heat sink being equipped with a TEC semiconductor cooler for dissipating heat from the laser;

[0010] a training module configured to obtain the measured temperature of the laser within a certain period of time and historical operating status data at corresponding moments to generate a database, and divide the database into training data and verification data; train a deep learning model based on the training data, and verify the trained deep learning model based on the verification data; and output the trained deep learning model if the difference between the predicted temperature outputted by the verification and the corresponding measured temperature is within a first preset interval;

[0011] a prediction module, configured to detect a first parameter and a second parameter of the laser in real time, and input the first parameter and the second parameter into a trained deep learning model to generate a predicted temperature corresponding to the laser; wherein the first parameter includes power and the second parameter includes operating time;

[0012] A temperature adjustment module is used to control the TEC semiconductor cooler to adjust the current temperature of the laser based on the predicted temperature.

[0013] In a specific embodiment, the plurality of lasers include a plurality of first lasers and a plurality of second lasers, and there are at least two second lasers between every two adjacent first lasers;

[0014] The system further includes: an allocation module for determining a measured temperature rise of the laser based on the measured temperature of the laser within a certain period of time, and if the measured temperature rise exceeds a second preset interval, using the laser as the first laser; if the measured temperature rise is within the second preset interval, using the laser as the second laser;

[0015] The first laser is independently equipped with one TEC semiconductor cooler, and at least two second lasers share one TEC semiconductor cooler.

[0016] In a specific embodiment, it also includes:

[0017] A temperature sensor, used to obtain the actual measured temperature of the laser at a preset time;

[0018] a calculation module, configured to calculate a difference between the measured temperature and the predicted temperature according to the measured temperature and the predicted temperature of the laser at a preset time;

[0019] An optimization module is configured to determine whether the difference between the measured temperature and the predicted temperature is within a third preset interval. If the difference between the measured temperature and the predicted temperature exceeds the third preset interval, the measured temperature of the laser and the historical operating status data at the corresponding moment are recorded to update a database, and the deep learning model is optimized based on the updated database.

[0020] In a specific embodiment, there are multiple temperature sensors, and the multiple temperature sensors are distributed within a distance range of 15% to 80% from the center to the edge of the laser array module; and the number of the temperature sensors gradually decreases along the direction from the center to the edge of the laser array module.

[0021] In a specific embodiment, it also includes:

[0022] A liquid cooling module, the liquid cooling module is arranged on a side of the laser close to the heat sink, and the liquid cooling module and the TEC semiconductor cooler are arranged in parallel or stacked;

[0023] The power module is used to drive the liquid cooling module to increase the circulation power when the power of the TEC semiconductor refrigerator is less than a threshold value.

[0024] In a specific embodiment, the first parameter further includes resistance, resonant frequency, emission wavelength or threshold current; the second parameter further includes the area of ​​the heat sink; the historical operating status data includes resistance, resonant frequency, emission wavelength, threshold current or operating time of the laser;

[0025] And / or, the temperature adjustment module includes:

[0026] A first determining module is configured to determine a predicted temperature rise of the laser within a certain period of time based on the predicted temperature.

[0027] An acquisition module is used to acquire the predicted temperature rise quantity based on the same TEC semiconductor cooler at the same time,

[0028] a second determining module, configured to use, if there are multiple predicted temperature rises, a maximum value of the predicted temperature rises as a final predicted temperature rise for adjusting the temperature of the TEC semiconductor cooler; and, if there is only one predicted temperature rise, use the predicted temperature rise as the final predicted temperature rise for adjusting the temperature of the TEC semiconductor cooler;

[0029] An adjustment module, configured to control the TEC semiconductor cooler to adjust the current temperature of the laser based on the final predicted temperature rise;

[0030] And / or, the adjustment module includes:

[0031] a first driving module, configured to drive the TEC semiconductor cooler to operate at 25% of the maximum power to restore the temperature of the laser when the final predicted temperature rise is within a first temperature rise range;

[0032] a second driving module, configured to drive the TEC semiconductor cooler to operate at 50% of the maximum power to restore the temperature of the laser when the final predicted temperature rise is within a second temperature rise range;

[0033] a third driving module, configured to drive the TEC semiconductor cooler to operate at 75% of the maximum power to restore the temperature of the laser when the final predicted temperature rise is within a third temperature rise range;

[0034] The reporting module is used to detect the current measured temperature of the laser and obtain current operating status data when the final predicted temperature rise exceeds the third temperature rise range, and update the measured temperature and the current operating status data to a database.

[0035] An intelligent heat dissipation method is applied to the laser packaging system as described above, and is characterized by comprising:

[0036] Obtaining the measured temperature of the laser within a certain period of time and historical operating status data at corresponding moments to generate a database, and dividing the database into training data and verification data; training a deep learning model based on the training data, and verifying the trained deep learning model based on the verification data; and outputting the trained deep learning model if the difference between the predicted temperature outputted by the verification and the corresponding measured temperature is within a first preset interval;

[0037] detecting a first parameter and a second parameter of the laser in real time, and inputting the first parameter and the second parameter into a trained deep learning model to generate a predicted temperature corresponding to the laser; wherein the first parameter includes power and the second parameter includes operating time;

[0038] The TEC semiconductor cooler is controlled based on the predicted temperature to adjust the current temperature of the laser.

[0039] In a specific embodiment, it also includes:

[0040] Obtaining the measured temperature of the laser at a preset time;

[0041] Calculating a difference between the measured temperature and the predicted temperature according to the measured temperature and the predicted temperature of the laser at a preset time;

[0042] Determine whether the difference between the measured temperature and the predicted temperature is within a third preset interval. If the difference between the measured temperature and the predicted temperature exceeds the third preset interval, record the measured temperature of the laser and the working status data at the corresponding moment to update the database, and optimize the deep learning model based on the updated database.

[0043] In a specific embodiment, the step of “controlling the TEC semiconductor cooler to adjust the current temperature of the laser based on the predicted temperature” includes:

[0044] Determine the predicted temperature rise of the laser within a certain period of time based on the predicted temperature,

[0045] Based on the same TEC semiconductor cooler at the same time, the number of predicted temperature rises is obtained,

[0046] If the number of the predicted temperature rises is multiple, the maximum value of the predicted temperature rises is used as the final predicted temperature rise for adjusting the temperature of the TEC semiconductor cooler; if the number of the predicted temperature rises is one, the predicted temperature rise is used as the final predicted temperature rise for adjusting the temperature of the TEC semiconductor cooler;

[0047] The TEC semiconductor cooler is controlled based on the final predicted temperature rise to adjust the current temperature of the laser.

[0048] In a specific embodiment, the “controlling the TEC semiconductor cooler to adjust the current temperature of the laser based on the final predicted temperature rise” includes:

[0049] When the final predicted temperature rise is within the first temperature rise range, driving the TEC semiconductor cooler to operate at 25% of the maximum power to restore the temperature of the laser;

[0050] When the final predicted temperature rise is within the second temperature rise range, driving the TEC semiconductor cooler to operate at 50% of the maximum power to restore the temperature of the laser;

[0051] When the final predicted temperature rise is within the third temperature rise range, driving the TEC semiconductor cooler to operate at 75% of the maximum power to restore the temperature of the laser;

[0052] When the final predicted temperature rise exceeds the third temperature rise range, the current measured temperature of the laser is detected and current operating status data is obtained, and the measured temperature and the current operating status data are updated to a database.

[0053] This application has at least the following beneficial effects:

[0054] The present application provides a laser packaging system and an intelligent heat dissipation method thereof, the system comprising: a laser array module, comprising a plurality of lasers arranged at intervals and distributed in an array, each laser having a heat sink for conducting heat, and each heat sink being equipped with a TEC semiconductor cooler for dissipating heat from the laser; a training module, for obtaining the measured temperature of the laser within a certain period of time and the historical operating status data at the corresponding moment to generate a database, and dividing the database into training data and verification data; training a deep learning model based on the training data, and verifying the trained deep learning model based on the verification data; outputting the trained deep learning model if the difference between the predicted temperature output by the verification and the corresponding measured temperature is within a first preset interval; a prediction module, for detecting a first parameter and a second parameter of the laser in real time, and inputting the first parameter and the second parameter into the trained deep learning model to generate a predicted temperature corresponding to the laser; wherein the first parameter includes power and the second parameter includes operating time; and a temperature control module, for controlling the TEC semiconductor cooler to adjust the current temperature of the laser based on the predicted temperature. This application uses a TEC semiconductor cooler to dissipate heat for the laser, solving the safety hazards of micro-leakage and water leakage in the existing water-cooling technology, making the system safer and more reliable; and adopts the cooperation of training module and prediction module to predict the predicted temperature of the laser through a deep learning model, thereby replacing the detection of traditional temperature sensors. The entire system structure is relatively simple, without the need to set up too many sensors, which greatly reduces the packaging difficulty and packaging cost, while also having a good heat dissipation effect.

[0055] Furthermore, the multiple lasers include multiple first lasers and multiple second lasers, with at least two second lasers between each two adjacent first lasers; the system also includes: an allocation module for determining the measured temperature rise of the laser based on the measured temperature of the laser within a certain period of time, if the measured temperature rise exceeds the second preset interval, the laser is used as the first laser; if the measured temperature rise is within the second preset interval, the laser is used as the second laser; the first laser is individually configured with a TEC semiconductor cooler, and at least two second lasers share a TEC semiconductor cooler. This application distinguishes the first laser and the second laser by the size of the measured temperature rise of the laser. The first laser is a laser that generates more heat, so a separate TEC semiconductor cooler is required for heat dissipation; the second laser is a laser that generates less heat, so at least two adjacent lasers can be set on the same TEC semiconductor cooler for heat dissipation. This design method not only reduces the number of TEC semiconductor coolers and reduces production costs, but also has a good heat dissipation effect, making the heat dissipation more uniform and efficient.

[0056] Furthermore, the present application also includes: a temperature sensor for obtaining the measured temperature of the laser at a preset moment; a calculation module for calculating the difference between the measured temperature and the predicted temperature based on the measured temperature at the preset moment of the laser and the predicted temperature; an optimization module for determining whether the difference between the measured temperature and the predicted temperature is within a third preset interval. If the difference between the measured temperature and the predicted temperature exceeds the third preset interval, the measured temperature of the laser and the historical operating status data at the corresponding moment are recorded to update the database, and the deep learning model is optimized based on the updated database. The present application verifies the predicted temperature of the prediction module in real time through a temperature sensor, so as to facilitate the subsequent timely optimization of the deep learning model, thereby improving the reliability of the predicted temperature in the system, reducing the deviation between the predicted temperature and the actual temperature, and significantly improving the reliability of the entire system. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0058] FIG1 is a schematic diagram of a laser array module provided in Example 1;

[0059] FIG2 is a first schematic diagram of the laser provided in Example 1;

[0060] FIG3 is a second schematic diagram of the laser provided in Example 1;

[0061] FIG4 is a first schematic diagram of the system provided in Example 1;

[0062] FIG5 is a second schematic diagram of the system provided in Example 1;

[0063] FIG6 is a third schematic diagram of the system provided in Example 1;

[0064] FIG7 is a flow chart of the intelligent heat dissipation method provided in Example 2;

[0065] FIG8 is a schematic diagram of the temperature sensor distribution provided in Example 1.

[0066] Reference numerals:

[0067] 1-Laser array module; 2-Training module; 3-Prediction module; 4-Temperature control module;

[0068] 5-distribution module; 6-temperature sensor; 7-calculation module; 8-optimization module;

[0069] 9-Liquid cooling module; 10-Power module; 11-Placement area;

[0070] 101-laser; 102-heat sink; 103-TEC semiconductor cooler;

[0071] 1011-first laser; 1012-second laser;

[0072] 401-first determination module; 402-acquisition module; 403-second determination module; 404-adjustment module;

[0073] 4041 - first driving module; 4042 - second driving module; 4043 - third driving module; 4044 - reporting module. DETAILED DESCRIPTION

[0074] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0075] Example 1

[0076] As shown in FIG1 , the present application provides a laser 101 packaging system and an intelligent heat dissipation method thereof. The system comprises: a laser array module 1 having a plurality of lasers 101 arranged at intervals and distributed in an array, each laser 101 having a heat sink 102 for conducting heat, and each heat sink 102 being equipped with a TEC semiconductor cooler 103 for dissipating heat from the laser 101; a training module 2 for obtaining the measured temperature of the laser 101 within a certain period of time and the historical operating status data at the corresponding time to generate a database, and dividing the database into training data and verification data; and performing deep learning model training based on the training data. The trained deep learning model is trained and verified based on the verification data; if the difference between the predicted temperature output by the verification and the corresponding measured temperature is within a first preset interval, the trained deep learning model is output; the prediction module 3 is used to detect the first parameter and the second parameter of the laser 101 in real time, and input the first parameter and the second parameter into the trained deep learning model to generate a predicted temperature corresponding to the laser 101; wherein the first parameter includes power and the second parameter includes operating time; the temperature adjustment module 4 is used to control the TEC semiconductor cooler 103 to adjust the current temperature of the laser 101 based on the predicted temperature.

[0077] This application uses a TEC semiconductor cooler 103 to dissipate heat from the laser 101, resolving the issues with existing water-cooled heat dissipation technologies, such as large size, inconvenience, and potential safety hazards such as micro-leakage and water leakage during use, making the system safer and more reliable. Furthermore, the system utilizes a training module 2 and a prediction module 3 to coordinate and predict the temperature of the laser 101 using a deep learning model, thereby replacing the traditional temperature sensor 6. The overall system structure is relatively simple, eliminating the need for excessive sensors and significantly reducing packaging difficulty and cost. Furthermore, the system exhibits superior heat dissipation, enabling the laser array module 1 to have a good photoelectric conversion rate and extending the service life of the laser array module 1.

[0078] As shown in Figures 1-3, the multiple lasers 101 of the present application can be multiple different types of lasers 101, or multiple lasers 101 of the same type, as needed. Different types of lasers 101 correspond to different powers.

[0079] In one embodiment, multiple lasers 101 are of the same type, each laser 101 having a heat sink 102, and each heat sink 102 correspondingly configured with a TEC semiconductor cooler 103. In other words, each laser 101 is individually configured with a TEC semiconductor cooler 103. By driving each TEC semiconductor cooler 103, the temperature rise of each laser 101 can be adjusted. This configuration provides more precise temperature control and better heat dissipation.

[0080] In another embodiment, multiple lasers 101 are configured as different types, with some lasers 101 sharing a TEC semiconductor cooler 103, and other lasers 101 being individually configured with a TEC semiconductor cooler 103. This configuration can reduce the use of TEC semiconductor coolers 103 and reduce the required cost.

[0081] In addition to classifying different lasers 101 according to their types, different lasers 101 can also be distinguished according to the temperature rise of the lasers 101. Specifically:

[0082] As shown in FIG1 and FIG4 , the plurality of lasers 101 include a plurality of first lasers 1011 and a plurality of second lasers 1012 , and there are at least two second lasers 1012 between every two adjacent first lasers 1011 ;

[0083] The system further includes: an allocation module 5 for determining a measured temperature rise of the laser 101 based on the measured temperature of the laser 101 within a certain period of time, and if the measured temperature rise exceeds a second preset interval, the laser 101 is used as a first laser 1011; if the measured temperature rise is within the second preset interval, the laser 101 is used as a second laser 1012;

[0084] The first laser 1011 is independently configured with a TEC semiconductor cooler 103 , and at least two second lasers 1012 share one TEC semiconductor cooler 103 .

[0085] The present application distinguishes the first laser 1011 and the second laser 1012 by the measured temperature rise of the laser 101. The first laser 1011 is a laser 101 that generates more heat, so a separate TEC semiconductor cooler 103 is required for heat dissipation. The second laser 1012 is a laser 101 that generates less heat, so at least two adjacent lasers 101 can be set on the same TEC semiconductor cooler 103 for heat dissipation. This design method not only reduces the number of TEC semiconductor coolers 103 and reduces production costs, but also has a good heat dissipation effect, making the heat dissipation more uniform and efficient.

[0086] As shown in Figure 5, this application also includes:

[0087] Temperature sensor 6, used to obtain the measured temperature of the laser 101 at a preset time;

[0088] Calculation module 7, used to calculate the difference between the measured temperature and the predicted temperature according to the measured temperature and the predicted temperature of the laser 101 at a preset time;

[0089] The optimization module 8 is used to determine whether the difference between the measured temperature and the predicted temperature is within a third preset interval. If the difference between the measured temperature and the predicted temperature exceeds the third preset interval, the measured temperature of the laser 101 and the historical operating status data at the corresponding time are recorded to update the database, and the deep learning model is optimized based on the updated database.

[0090] The function of the temperature sensor 6 of the present application is different from that of the prior art for measuring temperature to adjust the temperature of the laser 101. In the present application, the temperature sensor 6 only verifies the predicted temperature of the prediction module 3. If the difference between the measured temperature and the predicted temperature is within the third preset interval, the predicted temperature can continue to be used to adjust the temperature of the laser 101. If the difference between the measured temperature and the predicted temperature exceeds the third preset interval, it is necessary to record the measured temperature of the laser 101 and the working status data at the corresponding moment to update the database, and optimize the deep learning model based on the updated database. Under the joint action of the temperature sensor 6, the calculation module 7 and the optimization module 8, the optimization of the deep learning model is realized, the reliability of the predicted temperature in the system is improved, the deviation between the predicted temperature and the actual temperature is reduced, and the reliability of the entire system is significantly improved.

[0091] As shown in FIG1-6 , the temperature sensor 6 can be set to one or more.

[0092] If there is only one temperature sensor 6 , preferably, the temperature sensor 6 is disposed at the center of the laser array module 1 or at the maximum heating area of ​​the laser array module 1 .

[0093] As shown in FIG8 , when there are multiple temperature sensors 6 , the multiple temperature sensors 6 are distributed within a distance range of 15% to 80% from the center to the edge of the laser array module 1 ; and the number of temperature sensors 6 gradually decreases along the direction from the center to the edge of the laser array module 1 .

[0094] Specifically, for the laser 101 array, the closer the laser 101 is to the center, the greater the heat it generates, and the greater the temperature rise change will be. Therefore, it is necessary to drive the TEC semiconductor cooler 103 more accurately and timely to adjust the temperature of the laser 101 in the middle area. When setting the temperature sensor 6, the present application distributes multiple temperature sensors 6 within a distance range of 15% to 80% from the center to the edge of the laser array module 1, so that the temperature sensor 6 can promptly determine the difference between the predicted temperature of the laser 101 in the middle area and the actual temperature, so as to obtain the information of the predicted temperature anomaly in time, and optimize the deep learning model in time to ensure the accuracy and reliability of the predicted data (wherein, the distance range of 15% to 80% from the center to the edge of the laser array module 1 is the placement area 11 as shown in Figure 8).

[0095] As shown in Figure 3, this application also includes:

[0096] Liquid cooling module 9, the liquid cooling module 9 is arranged on the side of the laser 101 close to the heat sink 102, and the liquid cooling module 9 and the TEC semiconductor cooler 103 are arranged in parallel or stacked;

[0097] The power module 10 is used to drive the liquid cooling module 9 to increase the circulation power when the power of the TEC semiconductor cooler 103 is less than a threshold.

[0098] The laser 101 of the present application can dissipate heat through the TEC semiconductor cooler 103, and can also dissipate heat through the liquid cooling module 9. By combining the liquid cooling module 9 and the TEC semiconductor cooler 103, the heat dissipation effect of the system is further enhanced. At the same time, the liquid cooling module 9 can increase the circulation power when the TEC semiconductor cooler 103 is abnormal, thereby ensuring that the system always has a good heat dissipation effect, preventing the entire laser 101 array from overheating when the TEC semiconductor cooler 103 stops working, thereby affecting normal operation, so that the entire system has good stability and reduces safety hazards.

[0099] As shown in FIG1-6 , the first parameter further includes resistance, resonant frequency, emission wavelength, or threshold current; the second parameter further includes the area of ​​the heat sink 102; the historical operating status data includes the resistance, resonant frequency, emission wavelength, threshold current, or operating time of the laser 101;

[0100] And / or, the temperature adjustment module 4 includes:

[0101] The first determination module 401 is configured to determine a predicted temperature rise of the laser 101 within a certain period of time based on the predicted temperature.

[0102] The acquisition module 402 is used to acquire the number of predicted temperature rises based on the same TEC semiconductor cooler 103 at the same time.

[0103] The second determining module 403 is configured to use the maximum value of the predicted temperature rise as the final predicted temperature rise for adjusting the temperature of the TEC semiconductor cooler 103 if the number of predicted temperature rises is multiple; and use the predicted temperature rise as the final predicted temperature rise for adjusting the temperature of the TEC semiconductor cooler 103 if the number of predicted temperature rises is one;

[0104] An adjustment module 404 is configured to control the TEC semiconductor cooler 103 to adjust the current temperature of the laser 101 based on the final predicted temperature rise;

[0105] And / or, the adjustment module 404 includes:

[0106] The first driving module 4041 is configured to drive the TEC semiconductor cooler 103 to operate at 25% of the maximum power to restore the temperature of the laser 101 when the final predicted temperature rise is within the first temperature rise range;

[0107] The second driving module 4042 is configured to drive the TEC semiconductor cooler 103 to operate at 50% of the maximum power to restore the temperature of the laser 101 when the final predicted temperature rise is within the second temperature rise range;

[0108] The third driving module 4043 is configured to drive the TEC semiconductor cooler 103 to operate at 75% of the maximum power to restore the temperature of the laser 101 when the final predicted temperature rise is within the third temperature rise range;

[0109] The reporting module 4044 is configured to detect the current measured temperature of the laser 101 and obtain current operating status data when the final predicted temperature rise exceeds the third temperature rise range, and update the measured temperature and current operating status data to the database.

[0110] Example 2

[0111] As shown in Figure 7, the intelligent heat dissipation method is applied to the laser packaging system, including:

[0112] S1: Obtain the measured temperature of the laser within a certain period of time and the historical operating status data at the corresponding time to generate a database, and divide the database into training data and verification data; train the deep learning model based on the training data, and verify the trained deep learning model based on the verification data; if the difference between the predicted temperature output and the corresponding measured temperature is within a first preset interval, output the trained deep learning model;

[0113] S2: Detecting a first parameter and a second parameter of the laser in real time, and inputting the first parameter and the second parameter into a trained deep learning model to generate a predicted temperature corresponding to the laser; wherein the first parameter includes power and the second parameter includes operating time;

[0114] S3: Based on the predicted temperature, the TEC semiconductor cooler is controlled to adjust the current temperature of the laser.

[0115] This application trains a deep learning model and uses the first and second laser parameters to predict the laser's temperature, replacing traditional temperature sensor detection and making the entire system more intelligent. This heat dissipation method also achieves excellent heat dissipation, resulting in a high photoelectric conversion rate for the laser array module and a longer service life.

[0116] Smart cooling methods also include:

[0117] Obtain the measured temperature of the laser at a preset time;

[0118] Calculate the difference between the measured temperature and the predicted temperature according to the measured temperature and the predicted temperature at the preset time of the laser;

[0119] Determine whether the difference between the measured temperature and the predicted temperature is within a third preset interval. If the difference between the measured temperature and the predicted temperature exceeds the third preset interval, record the measured temperature of the laser and the working status data at the corresponding moment to update the database, and optimize the deep learning model based on the updated database.

[0120] This application verifies the implementation of the trained deep learning model by comparing the measured temperature with the predicted temperature, improves the reliability of the predicted temperature in the system, reduces the deviation between the predicted temperature and the actual temperature, and significantly improves the reliability of the entire system.

[0121] “Controlling the TEC semiconductor cooler to adjust the current temperature of the laser based on the predicted temperature” includes:

[0122] Determine the predicted temperature rise of the laser within a certain period of time based on the predicted temperature,

[0123] Based on the same TEC semiconductor cooler at the same time to obtain the number of predicted temperature rises,

[0124] If the number of predicted temperature rises is multiple, the maximum value of the predicted temperature rise is used as the final predicted temperature rise for adjusting the temperature of the TEC semiconductor cooler; if the number of predicted temperature rises is one, the predicted temperature rise is used as the final predicted temperature rise for adjusting the temperature of the TEC semiconductor cooler;

[0125] Based on the final predicted temperature rise, the TEC semiconductor cooler is controlled to adjust the current temperature of the laser.

[0126] By obtaining the maximum value of the predicted temperature rise on the same TEC semiconductor cooler, adjusting the final predicted temperature rise of the TEC semiconductor cooler based on the maximum value can better ensure the heat dissipation of the laser, thereby improving the reliability and safety of the system.

[0127] “Controlling the TEC semiconductor cooler to adjust the current temperature of the laser based on the final predicted temperature rise” includes:

[0128] When the final predicted temperature rise is within the first temperature rise range, the TEC semiconductor cooler is driven to operate at 25% of the maximum power to restore the temperature of the laser;

[0129] When the final predicted temperature rise is within the second temperature rise range, the TEC semiconductor cooler is driven to operate at 50% of the maximum power to restore the temperature of the laser;

[0130] When the final predicted temperature rise is in the third temperature rise range, the TEC semiconductor cooler is driven to operate at 75% of the maximum power to restore the temperature of the laser;

[0131] When the final predicted temperature rise exceeds the third temperature rise range, the current measured temperature of the laser is detected and the current operating status data is obtained, and the measured temperature and the current operating status data are updated to the database.

[0132] By driving the TEC semiconductor cooler at different powers with different predicted temperature rises, the heat dissipation effect of the entire laser array is improved and more uniform. Moreover, when the predicted temperature rise exceeds the third temperature rise range, the actual temperature of the laser at this time is detected and the working status data is obtained. The measured temperature and working status data are updated to the database, so that the deep learning model can be optimized based on the updated database.

[0133] Those skilled in the art will appreciate that the modules or steps of the present application described above can be implemented using a general-purpose computing device. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Alternatively, they can be implemented using program code executable by a computer device, so that they can be stored in a storage device and executed by the computing device. Alternatively, they can be fabricated into separate integrated circuit modules, or multiple modules or steps can be fabricated into a single integrated circuit module for implementation. Thus, the present application is not limited to any specific combination of hardware and software.

Claims

1. Laser packaging system, characterized in that, Including: A laser array module, having a plurality of lasers arranged at intervals and distributed in an array, each of the lasers having a heat sink for conducting heat, and each of the heat sinks being configured with a TEC semiconductor cooler for dissipating heat from the laser; A training module, configured to obtain the measured temperature of the lasers within a certain period of time and the historical operating state data at corresponding moments to generate a database, and divide the database into training data and validation data; Train a deep learning model based on the training data, and validate the trained deep learning model based on the validation data; If the difference between the predicted temperature output by the validation and the corresponding measured temperature is within a first preset interval, output the trained deep learning model; A prediction module, configured to detect the first parameter and the second parameter of the laser in real time, and input the first parameter and the second parameter into the trained deep learning model to generate a predicted temperature corresponding to the laser; wherein, the first parameter includes power, and the second parameter includes operating time; A temperature adjustment module, configured to control the TEC semiconductor cooler based on the predicted temperature to adjust the current temperature of the laser.

2. The laser packaging system according to claim 1, wherein The plurality of lasers include a plurality of first lasers and a plurality of second lasers, and there are at least two second lasers between every two adjacent first lasers; The system further includes: a distribution module, configured to determine the actual temperature rise of the lasers based on the measured temperature of the lasers within a certain period of time, and if the actual temperature rise exceeds a second preset interval, use the laser as the first laser; if the actual temperature rise is within the second preset interval, use the laser as the second laser; Each of the first lasers is separately configured with a TEC semiconductor cooler, and at least two of the second lasers share one TEC semiconductor cooler.

3. The laser packaging system according to claim 1, wherein Further including: A temperature sensor, configured to obtain the measured temperature of the laser at a preset moment; A calculation module, configured to calculate the difference between the measured temperature and the predicted temperature according to the measured temperature of the laser at the preset moment and the predicted temperature; An optimization module, configured to determine whether the difference between the measured temperature and the predicted temperature is within a third preset interval, and if the difference between the measured temperature and the predicted temperature exceeds the third preset interval, record the measured Temperature and the historical operating state data at the corresponding moment to update the database, and optimize the deep learning model according to the updated database.

4. The laser packaging system according to claim 3, wherein, There are a plurality of the temperature sensors, and the plurality of temperature sensors are all distributed within a distance range of 15% to 80% from the center to the edge of the laser array module; and along the direction from the center to the edge of the laser array module, the number of the temperature sensors gradually decreases.

5. The laser packaging system according to claim 1, characterized in that, Further including: A liquid cooling module, the liquid cooling module is arranged on one side of the laser close to the heat sink, and the liquid cooling module is arranged in parallel or stacked with the TEC semiconductor cooler; A power module, configured to drive the liquid cooling module to increase the circulation power when the power of the TEC semiconductor cooler is less than a threshold.

6. The laser packaging system according to claim 1, wherein, The first parameter further includes resistance, resonance frequency, emission wavelength or threshold current; the second parameter further includes the area of the heat sink; the historical operating state data includes the resistance, resonance frequency, emission wavelength, threshold current or operating time of the laser; And / or, the temperature control module includes: A first determination module, configured to determine a predicted temperature rise of the laser within a certain period of time based on the predicted temperature; An acquisition module, configured to acquire the number of the predicted temperature rises at the same moment based on the same TEC semiconductor refrigerator; A second determination module, configured to, if the number of the predicted temperature rises is multiple, use the maximum value of the predicted temperature rises as the final predicted temperature rise for adjusting the temperature of the TEC semiconductor refrigerator; if the number of the predicted temperature rises is one, use the predicted temperature rise as the final predicted temperature rise for adjusting the temperature of the TEC semiconductor refrigerator; An adjustment module, configured to control the TEC semiconductor refrigerator based on the final predicted temperature rise to adjust the current temperature of the laser; And / or, the adjustment module includes: A first driving module, configured to, when the final predicted temperature rise is within a first temperature rise range, drive the TEC semiconductor refrigerator to work at 25% of the maximum power to restore the temperature of the laser; A second driving module, configured to, when the final predicted temperature rise is within a second temperature rise range, drive the TEC semiconductor refrigerator to work at 50% of the maximum power to restore the temperature of the laser; A third driving module, configured to, when the final predicted temperature rise is within a third temperature rise range, drive the TEC semiconductor refrigerator to work at 75% of the maximum power to restore the temperature of the laser; A reporting module, configured to, when the final predicted temperature rise exceeds the third temperature rise range, detect the current measured temperature of the laser and acquire the current operating state data, and update the measured temperature and the current operating state data to the database. Including:

7. An intelligent heat dissipation method, applied to the laser packaging system according to any one of claims 1-6, characterized in that, Acquire the measured temperature of the laser within a certain period of time and the historical operating state data at the corresponding moment to generate a database, and divide the database into training data and verification data; Train a deep learning model based on the training data, and verify the trained deep learning model based on the verification data; If the difference between the predicted temperature output by the verification and the corresponding measured temperature is within a first preset interval, output the trained deep learning model; Real-time detect the first parameter and the second parameter of the laser, and input the first parameter and the second parameter into the trained deep learning model to generate the predicted temperature corresponding to the laser; wherein, the first parameter includes power, and the second parameter includes operating time; Control the TEC semiconductor refrigerator based on the predicted temperature to adjust the current temperature of the laser. ​

Images