Air-cooled laser adaptive control method and system based on health state assessment

Abstract

The application provides a kind of air-cooled laser adaptive control method and system based on health state evaluation, the method synchronously acquires laser internal operating parameters, environmental parameters and acoustic vibration signals, decouples to calculate thermal resistance, photoelectric efficiency and fan performance health sub-score, generates comprehensive health index by D-S evidence theory fusion, and calculates dew risk index combined with Magnus formula.Each sub-score is input into a digital twin model to correct thermal resistance and heat transfer coefficient, and the future working condition prediction value is output.The current value and prediction trend are used as double indexes to match five types of control modes through multi-dimensional strategy mapping, calculate prospective compensation, and adjust the heat dissipation capacity, output power and heating amount in linkage.The application solves the problems of thermal management response lag, passive dew prevention and control and degradation traceability difficulty, realizes adaptive optimization throughout the life cycle, significantly prolongs the fault-free operation time of the equipment, and reduces the operation and maintenance cost.

Description

Technical Field

[0001] This invention relates to the field of laser technology, and in particular to an adaptive control method and system for air-cooled lasers based on health status assessment. Background Technology

[0002] Air-cooled lasers are widely used in industrial processing, medical aesthetics, and other fields due to their simple structure and low cost. However, their heat dissipation capacity is limited by the physical limits of the air-cooling system, and they are highly susceptible to condensation failures caused by ambient temperature and humidity. Existing technologies typically use fixed temperature thresholds to control fan speed or rely solely on passive responses based on data from a single sensor. This results in lagging thermal management response and an inability to distinguish between device aging and decreased heat dissipation capacity. Furthermore, traditional condensation protection methods often rely on single dew point monitoring, ignoring the coldest point drift caused by increased internal thermal resistance of the laser, and lacking forward-looking prediction of future degradation trends, leading to rigid control strategies. In addition, existing health assessment methods often depend on human experience or single parameter thresholds, making it difficult to adaptively adjust assessment weights throughout the equipment's lifecycle. They are also prone to misjudgments when sensors fail occasionally, failing to meet the long-term stable operation requirements of high-power lasers under complex conditions. Summary of the Invention

[0003] In view of the shortcomings of the prior art described above, the purpose of this invention is to provide an adaptive control method and system for air-cooled lasers based on health status assessment, which solves the problem of thermal management response lag caused by the reliance on a single parameter threshold for passive control of air-cooled lasers in the prior art.

[0004] To achieve the above and other related objectives, the present invention provides the following technical solution:

[0005] An adaptive control method for wind-cooled lasers based on health status assessment includes the following steps:

[0006] S1, synchronously acquires the laser's internal operating parameters, external environmental parameters, and acoustic vibration parameters;

[0007] S2, based on the internal operating parameters, the thermal resistance health sub-score and photoelectric efficiency health sub-score are decoupled and calculated through a preset health assessment model; the acoustic vibration parameters are decomposed in the time-frequency domain to extract the energy proportion of the characteristic frequency of the fan bearing and the passing frequency of the blades, and the fan efficiency health sub-score is calculated; the health sub-scores are fused to generate a comprehensive health index characterizing the degree of degradation of the laser body; based on the external environmental parameters, the condensation risk index characterizing the possibility of condensation is calculated.

[0008] S3, input the internal operating parameters, external environmental parameters and each health sub-score into the pre-constructed digital twin model, and output the predicted value of the comprehensive health index and the predicted value of the condensation risk index within the future time window;

[0009] S4, using the current value and predicted value change trend of the comprehensive health index as the first index, and the current value and predicted value change trend of the condensation risk index as the second index, the target control mode and forward-looking compensation amount are matched and determined through the preset multi-dimensional strategy mapping relationship. The target control mode includes standard temperature control mode, anti-condensation priority mode, enhanced heat dissipation mode, compromise mode and safe shutdown mode.

[0010] S5, in response to the target control mode and the forward compensation amount, generates a cooperative control command to dynamically adjust the heat dissipation capacity of the air-cooling system, the output power of the laser, and the heating amount of the air duct heating device.

[0011] In one embodiment of the present invention, the step of fusing the various health sub-scores to generate a comprehensive health index specifically includes:

[0012] The thermal resistance health sub-score, photoelectric efficiency health sub-score, and fan efficiency health sub-score are considered as three independent sources of evidence.

[0013] The real-time operating status of the sensors corresponding to each evidence source is obtained. The operating status includes signal noise level, sampling stability and historical drift record. The original health sub-score value is corrected based on the real-time operating status to obtain the corrected health sub-score.

[0014] When the real-time working state meets the preset reliability requirements, based on the preset mapping relationship, the corrected health sub-score is converted into a basic confidence assignment value characterizing the laser in a healthy state, a sub-healthy state, and a degraded state.

[0015] When the real-time working state does not meet the preset reliability requirements, the evidence source data is determined to be unreliable, and the basic credibility allocation value of the evidence source is reassigned as: m(Θ) = 1, m(healthy) = 0, m(sub-healthy) = 0, m(deterioration) = 0, where Θ is the identification framework and Θ = {healthy, sub-healthy, deterioration}.

[0016] Based on the cumulative runtime of the air-cooled laser, the fusion weight of the basic credibility allocation value of each evidence source in the fusion process is dynamically adjusted, and the basic credibility allocation value after weight adjustment is determined as the final credibility allocation value of the evidence source.

[0017] Using the combination rules based on DS evidence theory, the final credibility allocation values ​​of the three evidence sources are fused to obtain the fused comprehensive credibility distribution;

[0018] Based on the integrated credibility distribution, the integrated health index is generated according to preset weight coefficients, where healthy state, sub-healthy state and degenerative state correspond to different weight coefficients.

[0019] In one embodiment of the present invention, the step of dynamically adjusting the fusion weights of each evidence source based on the cumulative runtime of the wind-cooled laser specifically includes:

[0020] When the cumulative running time is less than the first preset time, a first weight is assigned to the evidence source corresponding to the photoelectric efficiency health sub-score, and the first weight is greater than the sum of the weights of the other two evidence sources;

[0021] When the cumulative runtime is greater than or equal to the first preset runtime and less than the second preset runtime, a second weight is assigned to the evidence source corresponding to the thermal resistance health sub-score, and the second weight is greater than the sum of the weights of the other two evidence sources;

[0022] When the cumulative runtime is greater than or equal to the second preset runtime, equal weights are assigned to each evidence source.

[0023] The first preset duration is less than the second preset duration.

[0024] In one embodiment of the present invention, the calculation of the condensation risk index characterizing the likelihood of condensation based on the external environmental parameters in step S2 specifically includes:

[0025] Dew point temperature is calculated using the Magnus formula based on ambient temperature and relative humidity. ;

[0026] Obtain the lowest internal temperature of the laser The lowest internal temperature of the laser is obtained in real time by multiple temperature sensors distributed on the surface of the pump source, gain medium and heat sink;

[0027] Calculate the condensation risk margin ,when When indicates the existence of a safety margin, This indicates that the dew point temperature has been reached or exceeded.

[0028] Set a safe temperature difference threshold and dangerous temperature difference threshold ,and ;

[0029] when At that time, the condensation risk index ;when At that time, the condensation risk index ;when At that time, the condensation risk index .

[0030] In one embodiment of the present invention, step S3 includes:

[0031] A digital twin model of an air-cooled laser is constructed, comprising a thermodynamic sub-model, a photoelectric conversion sub-model, a fan performance sub-model, and an environmental heat exchange sub-model.

[0032] The standard thermal resistance curve, photoelectric conversion efficiency benchmark curve, fan performance curve, and environmental heat exchange coefficient of the laser in its new state are used as the initial parameters of the digital twin model.

[0033] Input the current internal operating parameters, external environmental parameters, and health sub-scores into the digital twin model;

[0034] The digital twin model outputs a comprehensive health index prediction and a condensation risk index prediction within a future time window based on current parameters and preset operating condition change trends. The preset operating condition change trends include preset output power change curves, fan speed change curves, and ambient temperature change curves of the laser within the future time window.

[0035] In one embodiment of the present invention, the matching rule for the multi-dimensional strategy mapping relationship in step S4 specifically includes: defining the change in health status. ,in The current overall health index, The predicted value of the comprehensive health index; the change in condensation risk is defined. ,in This represents the current condensation risk index. This is the predicted value of the condensation risk index;

[0036] Set the following thresholds, The threshold for health status. The threshold for sub-health status, and ; The low-risk threshold; This is the lower limit threshold for medium risk. It is the upper limit threshold of medium risk, and ; The tolerance threshold for declining health status To set a tolerance threshold for the increased risk of condensation, The threshold for triggering a safe shutdown;

[0037] when ,and , , At that time, it matches the standard temperature control mode;

[0038] when ,and or Match the anti-condensation priority mode;

[0039] when ,and , At the same time, match the enhanced heat dissipation mode;

[0040] when ,and At that time, the compromise mode is matched;

[0041] when ,or At that time, match the safe shutdown mode.

[0042] In one embodiment of the present invention, when matching a standard temperature control mode, the forward compensation amount is zero;

[0043] When matching the anti-condensation priority mode, when At that time, the forward-looking compensation amount includes the preheating amount of the duct heating device; when At that time, the forward-looking compensation amount includes the preheating amount of the duct heating device and the pre-reduction amount of the output power, wherein The threshold for predicting condensation risk;

[0044] When matching enhanced heat dissipation mode, when At that time, the forward-looking compensation amount includes the pre-increment of fan speed; when At that time, the forward compensation amount includes the pre-increment of fan speed and the pre-decrease of output power, wherein Predict trigger thresholds for health status;

[0045] When matching compromise modes, when and At that time, the forward-looking compensation amount includes a weighted combination of the fan speed pre-increment and the preheating amount of the duct heating device; when and In this context, the forward-looking compensation amount includes a weighted combination of the pre-increment of fan speed, the pre-decrease of output power, and the pre-heating amount of the duct heating device; in the compromise mode, the weighted combination includes the heat dissipation-related weights. =( ) / ( ), weighting related to condensation prevention ( - ),Will and After normalization, the final weights are obtained. , The forward-looking compensation amount of the weighted combination = preset maximum pre-increment of fan × +Preset maximum preheating heat × When the forward-looking compensation amount also includes the output power pre-drop amount, the output power pre-drop amount is executed in full according to the preset maximum power pre-drop amount, and is superimposed with the forward-looking compensation amount of the above weighted combination.

[0046] In one embodiment of the present invention, the multi-dimensional strategy mapping relationship includes an initial mapping table and historical correction records;

[0047] After each execution of the target control mode, the actual control effect evaluation index is recorded. The evaluation index includes at least the following:

[0048] The first deviation between the actual rate of change and the predicted rate of change of the overall health index after control. ;

[0049] The second deviation between the actual and predicted rates of change of the controlled condensation risk index ;

[0050] Overall evaluation value ,in and The preset deviation weighting coefficient;

[0051] When the comprehensive evaluation value When the cumulative offset exceeds the preset update threshold, the self-correction process of the mapping table is triggered.

[0052] The self-correction process is based on the comprehensive evaluation value from the most recent N control records. For the target control mode corresponding to the highest control record, adjust the mode matching priority under the corresponding index condition, promote the target control mode with the highest comprehensive evaluation value to the first priority, and at the same time, demote the other target control modes in order of comprehensive evaluation value from high to low.

[0053] In one embodiment of the present invention, degradation pattern recognition is performed based on the relative magnitudes of the thermal resistance health sub-score, the photoelectric efficiency health sub-score, and the fan performance health sub-score:

[0054] When the thermal resistance health sub-score is less than both the photoelectric efficiency health sub-score and the fan performance health sub-score, the dominant degradation mode is determined to be the heat dissipation system degradation, and the first maintenance recommendation to clean the heat dissipation channel is generated.

[0055] When the photoelectric efficiency health sub-score is less than both the thermal resistance health sub-score and the fan efficiency health sub-score, the dominant degradation mode is determined to be pump source or optical device degradation, and a second maintenance recommendation to check the pump source or optical device is generated.

[0056] When the fan performance health sub-score is lower than both the thermal resistance health sub-score and the photoelectric efficiency health sub-score, the dominant degradation mode is determined to be fan mechanical degradation, and a third maintenance recommendation to replace the fan is generated.

[0057] When the thermal resistance health sub-score, photoelectric efficiency health sub-score, and fan performance health sub-score are all below their respective preset severe thresholds, it is judged as a comprehensive degradation, and a fourth maintenance suggestion for whole-machine overhaul is generated.

[0058] In another embodiment of the present invention, an adaptive control system for a wind-cooled laser based on health status assessment is provided, comprising:

[0059] The data acquisition module is configured to synchronously acquire the laser's internal operating parameters, external environmental parameters, and acoustic vibration signals.

[0060] The health assessment module is connected to the data acquisition module and the strategy matching module respectively. It is configured to calculate the health sub-scores of thermal resistance, photoelectric efficiency and fan performance based on the internal operating parameters, integrate them to generate a comprehensive health index, and calculate the condensation risk index based on external environmental parameters.

[0061] The digital twin prediction module is connected to the data acquisition module and is configured to construct a thermodynamic and photoelectric conversion model to output the predicted values ​​of the comprehensive health index and the condensation risk index within a future time window.

[0062] The strategy matching module is connected to the health assessment module and the digital twin prediction module respectively. It is configured to match the target control mode through a multi-dimensional mapping relationship using the current index value and the predicted trend of change as indexes, and calculate the forward compensation amount.

[0063] The execution control module is connected to the strategy matching module and is configured to generate cooperative instructions in response to the target control mode to adjust the air cooling capacity, laser output power and air duct heating amount in a coordinated manner.

[0064] As described above, the adaptive control method and system for air-cooled lasers based on health status assessment of the present invention has the following beneficial effects: The present invention achieves for the first time accurate source tracing and robust assessment of laser thermal management degradation, optical device aging, and air-cooling power attenuation by decoupling the calculation of three types of health sub-scores (thermal resistance, photoelectric efficiency, and fan efficiency) and fusing them with DS evidence, thus solving the problem that traditional methods cannot distinguish the root causes of degradation; Based on the digital twin model, health sub-scores are introduced as correction inputs, combined with a dual-index decision mechanism of "current state + future trend", enabling the control strategy to shift from passive response to active prediction, effectively avoiding thermal failure and condensation damage; Multi-dimensional strategy mapping and forward-looking compensation mechanism properly solve the contradiction between heat dissipation and anti-condensation in high humidity environments, and with the self-correction logic based on historical control effects, the adaptability and stability of the system throughout its entire life cycle are significantly improved. Attached Figure Description

[0065] Figure 1The flowchart shown is a process of the method of the present invention. Detailed Implementation

[0066] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. It should be noted that, unless otherwise specified, the following embodiments and features described herein can be combined with each other.

[0067] Please see Figure 1 This invention provides an adaptive control method for an air-cooled laser based on health status assessment, comprising the following steps: S1, synchronously acquiring the laser's internal operating parameters, external environmental parameters, and acoustic vibration parameters; S2, based on the internal operating parameters, decoupling and calculating the thermal resistance health sub-score and photoelectric efficiency health sub-score using a preset health assessment model; performing time-frequency domain decomposition on the acoustic vibration parameters to extract the energy proportion of the fan bearing characteristic frequency and the blade passing frequency, and calculating the fan efficiency health sub-score; fusing the various health sub-scores to generate a comprehensive health index characterizing the degree of laser degradation; calculating a condensation risk index characterizing the possibility of condensation based on the external environmental parameters; S3, combining the internal operating parameters, external environmental parameters, and various health sub-scores... The system inputs a pre-constructed digital twin model and outputs a predicted value for the comprehensive health index and a predicted value for the condensation risk index within a future time window; S4, using the current value and predicted trend of the comprehensive health index as the first index and the current value and predicted trend of the condensation risk index as the second index, the system matches and determines the target control mode and forward-looking compensation amount through a pre-set multi-dimensional strategy mapping relationship. The target control mode includes a standard temperature control mode, an anti-condensation priority mode, an enhanced heat dissipation mode, a compromise mode, and a safe shutdown mode; S5, in response to the target control mode and forward-looking compensation amount, the system generates a coordinated control command to dynamically adjust the heat dissipation capacity of the air-cooled system, the output power of the laser, and the heating amount of the air duct heating device.

[0068] This method addresses the industry pain points of air-cooled lasers, such as delayed thermal management response, passive condensation control, and difficulty in tracing degradation status. First, in stage S1, it simultaneously collects internal laser operating parameters, external environmental parameters, and acoustic vibration parameters to provide multi-source data support for subsequent status assessment. Stage S2 conducts a multi-dimensional decoupled health status assessment: the thermal resistance health sub-score is obtained through the following process: based on the pump source temperature, heat sink temperature, and drive current from the internal operating parameters, a real-time thermal resistance curve is fitted at different output powers. This curve is then compared across the entire range with a standard thermal resistance curve in a brand-new laser state. The zero-power drift intercept, power-thermal resistance drift slope, and full-power drift area are extracted. Based on these morphological characteristics, two degradation modes—aging of the thermal interface material and decay of the convective heat transfer coefficient—are distinguished before calculation. The photoelectric efficiency health sub-score is obtained through the following process: based on the input electrical power and output optical power from the internal operating parameters... The real-time photoelectric conversion efficiency at different power levels is calculated, and this efficiency curve is compared with the photoelectric conversion efficiency benchmark curve under new conditions. The efficiency attenuation across the entire power range is extracted, and the photoelectric efficiency health sub-score is calculated by combining the efficiency temperature correction coefficient corresponding to the temperature data in the internal operating parameters. The fan performance health sub-score is obtained through the following process: time-frequency domain decomposition of the collected acoustic vibration parameters is performed to extract the amplitude of the characteristic frequency of the fan bearing and the passing frequency of the blades. The proportion of the energy of the above characteristic frequencies to the total vibration energy is calculated, and the offset between the real-time fan speed-airflow characteristics and the fan performance curve under new conditions is also calculated. The three types of health sub-scores are fused to generate a comprehensive health index characterizing the degree of laser body degradation, and a condensation risk index characterizing the possibility of condensation is calculated based on the ambient temperature and humidity. In stage S3, the three types of health sub-scores, along with real-time operating and environmental parameters, are input into a pre-constructed digital twin model. Based on the health sub-scores, key parameters such as thermal resistance and convective heat transfer coefficient within the model are corrected, and the predicted values ​​of the comprehensive health index and condensation risk index within the future time window are output, achieving forward prediction of degradation trends. In stage S4, the current values ​​and predicted trends of the comprehensive health index and condensation risk index are used as dual indices. A pre-defined multi-dimensional strategy mapping relationship is used to match the corresponding target control mode, and a forward-looking compensation amount is calculated to achieve precise decision-making based on both the current state and future trends. In stage S5, the decision results are responded to, generating collaborative control commands to adjust the heat dissipation capacity of the air-cooling system, the output power of the laser, and the heating amount of the duct heating device, thus completing adaptive closed-loop control.

[0069] This method, through decoupled calculation of three health sub-scores—thermal resistance, photoelectric efficiency, and fan performance—achieves for the first time precise tracing of thermal management degradation, photoelectric conversion degradation, and air-cooling power degradation in air-cooled lasers, solving the problem of control strategy inaccuracies caused by the inability of traditional methods to distinguish the root causes of degradation. Based on a DS evidence fusion mechanism for sensor reliability verification, it can automatically identify abnormal states such as sensor drift and failure, maximizing the quality of uncertainty from unreliable data sources, effectively avoiding evaluation bias caused by single-sensor errors, and improving evaluation robustness in complex industrial environments. The digital twin model introduces health sub-scores as core correction inputs, enabling… The model's state is mapped to the actual degradation state of the equipment in real time, and the prediction results accurately reflect the future operating performance considering its own aging characteristics. The prediction accuracy is much higher than that of traditional statistical trend extrapolation methods. The multi-dimensional strategy matching mechanism, which uses the current value of the dual exponent and the predicted trend as indexes, combined with the forward-looking compensation strategy, properly solves the contradiction between the increase in heat dissipation demand and the prevention and control of condensation risk, and avoids thermal shock or condensation damage to optical components caused by passive decondensation. At the same time, the self-correction mechanism can dynamically optimize the strategy mapping priority according to the historical control effect, adapt to the degradation characteristics of the equipment throughout its entire life cycle, and has excellent long-term control stability, which can effectively extend the fault-free operation time of the laser.

[0070] The specific steps for fusing the various health sub-scores to generate a comprehensive health index include: treating the thermal resistance health sub-score, photoelectric efficiency health sub-score, and fan efficiency health sub-score as three independent evidence sources; acquiring the real-time operating status of the sensors corresponding to each evidence source, including signal noise level, sampling stability, and historical drift records; and correcting the original health sub-score values ​​based on the real-time operating status to obtain corrected health sub-scores; when the real-time operating status meets preset reliability requirements, converting the corrected health sub-scores into basic credibility allocation values ​​characterizing the laser's healthy, sub-healthy, and degraded states based on a preset mapping relationship; when the real-time operating status does not meet preset reliability requirements, determining that the evidence source data is unreliable, and reassigning the basic credibility allocation value of the evidence source to: m(Θ) = 1, m(healthy) = 0, m(sub-healthy) = 0, m(degraded) = 0, where Θ is the recognition frame and Θ = {Healthy, Sub-healthy, Degenerative}; Based on the cumulative runtime of the air-cooled laser, the fusion weight of the basic credibility allocation value of each evidence source in the fusion process is dynamically adjusted, and the basic credibility allocation value after weight adjustment is determined as the final credibility allocation value of the evidence source; Using the combination rules based on DS evidence theory, the final credibility allocation values ​​of the three evidence sources are fused to obtain the fused comprehensive credibility distribution; Based on the fused comprehensive credibility distribution, the comprehensive health index is generated according to the preset weight coefficients, where the healthy state, sub-healthy state, and degenerative state correspond to different weight coefficients.

[0071] In the evidence preprocessing layer, the system treats the three health sub-scores—thermal resistance, photoelectric efficiency, and fan efficiency—as independent evidence sources and acquires the real-time operating status of the corresponding sensors for each evidence source—including signal noise level, sampling stability, and historical drift records. Based on this status information, the system corrects the original health sub-scores: when the sensor is working reliably, the corrected score is mapped to a basic credibility assignment on the three propositions of "healthy," "sub-healthy," and "degraded"; when the sensor reliability does not meet the requirements, the evidence source is determined to be completely unreliable, and the basic credibility assignment value of the evidence source is reassigned to m(Θ)=1 (i.e., completely uncertain), so that the evidence source is "neutralized and abstained" in subsequent fusion, avoiding erroneous data from contaminating the overall evaluation results.

[0072] In the dynamic fusion layer, the system dynamically adjusts the fusion weights of each evidence source based on the different stages of the cumulative runtime of the wind-cooled laser. In the early stages of operation, photoelectric efficiency best reflects early degradation, so it is given a dominant weight. In the middle stages, thermal resistance degradation becomes the primary issue, so thermal resistance is given a dominant weight. In the final stages, all three aspects degrade significantly, so equal-weighted fusion is used. The weighted values ​​of each evidence source are assigned as the final credibility distribution. These are then input into the combination rules of the DS evidence theory for orthogonal summation to obtain the fused comprehensive credibility distribution. Finally, according to the preset weight coefficients corresponding to the three states of healthy, sub-healthy, and degraded, the comprehensive credibility distribution is compressed into a single comprehensive health index for subsequent strategy matching and decision control.

[0073] Compared to traditional fixed-weighted average or single-sensor determination methods, this fusion scheme has several advantages: First, it introduces a sensor reliability verification mechanism, which avoids evaluation bias caused by sensor drift and failure through uncertainty quality allocation. Even in the event of a single sensor failure, it can still output stable results based on other reliable evidence sources. Compared to fault-tolerant schemes that directly discard abnormal data, this design avoids interference from erroneous data and retains the possibility of recovery from abnormal evidence sources, significantly improving the evaluation robustness in complex industrial environments. Second, it uses a dynamic weight adjustment mechanism based on cumulative runtime, which conforms to the degradation pattern of the laser's entire life cycle and solves the problem that fixed weights cannot adapt to the dominant degradation characteristics at different stages, making the fusion result more consistent with the actual health status of the equipment. Third, it uses DS evidence theory to process multi-source uncertainty information, which can effectively fuse heterogeneous health sub-score data and quantify the confidence of different degradation states. Compared to methods such as fuzzy comprehensive evaluation, it has higher distinguishability of edge states, providing a more reliable decision-making basis for the accurate matching of subsequent control modes.

[0074] The step of dynamically adjusting the fusion weight of each evidence source based on the cumulative runtime of the air-cooled laser specifically includes: when the cumulative runtime is less than a first preset runtime, assigning a first weight to the evidence source corresponding to the photoelectric efficiency health sub-score, the first weight being greater than the sum of the weights of the other two evidence sources; when the cumulative runtime is greater than or equal to the first preset runtime and less than a second preset runtime, assigning a second weight to the evidence source corresponding to the thermal resistance health sub-score, the second weight being greater than the sum of the weights of the other two evidence sources; when the cumulative runtime is greater than or equal to the second preset runtime, assigning equal weights to each evidence source; wherein the first preset runtime is less than the second preset runtime.

[0075] The core of this technical solution lies in establishing a dynamic weight allocation mechanism based on the degradation patterns throughout the laser's entire lifecycle. In the early stages of laser operation (cumulative runtime less than a first preset duration), the degradation of photoelectric conversion efficiency in the pump source and optical components is the first sign of performance decline. At this time, the thermal interface material and fan bearings are still in a stable performance period. Therefore, a dominant weight (greater than the sum of the other two) is assigned to the photoelectric efficiency health sub-score, ensuring the evaluation results maintain the highest sensitivity to early optical degradation, achieving "early detection and early intervention." In the middle stages of laser operation (cumulative runtime greater than or equal to the first preset duration and less than a second preset duration), the aging of the thermal interface material and the increase in thermal resistance caused by micro-deformation of the heat sink contact surface gradually become the main contradiction restricting the laser's output performance. At this point, the dominant weight is transferred to the thermal resistance health sub-score, focusing the evaluation on the most critical thermal management degradation risk at this stage. At the end of the laser's operation (with a cumulative operating time greater than or equal to the second preset duration), thermal resistance, photoelectric efficiency, and fan efficiency all show significant degradation, and the degradation mechanisms are coupled with each other. At this point, a single source of evidence can no longer represent the overall health status. Therefore, the three factors are given equal weight to reflect the comprehensive degradation degree of the equipment's "systemic aging" in the most objective way.

[0076] Compared to traditional fixed-weight fusion mechanisms, this weight adjustment scheme offers significant advantages in adaptability and evaluation accuracy. Firstly, it precisely matches the degradation characteristics throughout the laser's entire lifecycle, avoiding missed detections of minor photoelectric efficiency degradation due to excessively high weights for thermal resistance and fans in the new machine stage, and misjudgments of degradation modes due to fixed weights in the older machine stage. This greatly improves the temporal consistency and state identification accuracy of the comprehensive health index. Secondly, by setting a rule that "the weight of the dominant evidence source is greater than the sum of the other two," a reasonable decision bias is formed in the identification of core degradation modes at the current stage, effectively suppressing noise interference from secondary degradation factors. Simultaneously, the equal-weight design for balancing degradation stages avoids evaluation bias caused by excessively high weights for a single indicator, providing a more reliable decision-making basis for the accurate matching of subsequent control modes and the scientific formulation of operation and maintenance strategies. Furthermore, this weight adjustment logic can be implemented solely based on the device's existing cumulative runtime data, requiring no additional sensors, resulting in low engineering implementation costs and applicability to health management scenarios for air-cooled lasers of various power levels.

[0077] The calculation of the condensation risk index, which characterizes the likelihood of condensation, based on the external environmental parameters in step S2 specifically includes: calculating the dew point temperature using the Magnus formula based on the ambient temperature and relative humidity. ; Obtain the lowest internal temperature of the laser The lowest internal temperature of the laser is obtained in real time by multiple temperature sensors distributed on the surfaces of the pump source, gain medium, and heat sink; the condensation risk margin is calculated. ,when When indicates the existence of a safety margin, This indicates that the dew point temperature has been reached or exceeded; a safe temperature difference threshold is set. and dangerous temperature difference threshold ,and ;when At that time, the condensation risk index ;when At that time, the condensation risk index ;when At that time, the condensation risk index .

[0078] This step follows the external environmental parameter processing flow of stage S2. First, based on the ambient temperature and relative humidity, the dew point temperature is calculated using the internationally recognized Magnus high-precision dew point calculation formula. Compared to the traditional lookup table method, this method offers higher calculation accuracy and adaptability to different operating conditions. Subsequently, multiple temperature sensors distributed on the pump source, gain medium, and heat sink surface are used for synchronous sampling, and the minimum value of the multi-channel data is taken as the lowest internal temperature of the laser. —The above three types of components constitute the core optical region of the laser, which is also the area with the lowest internal temperature and most prone to condensation. Multi-sensor sampling effectively avoids the monitoring blind spots of single-point temperature measurement; then, the condensation risk margin is calculated. ,when This indicates that the internal temperature is above the dew point and there is a safety margin. This indicates that the internal temperature has reached or exceeded the dew point, posing a risk of condensation; further safety temperature difference thresholds should be set. dangerous temperature difference threshold (Both conditions are met) The risk margin is divided into three intervals, and continuous temperature differences are mapped to a 0-1 standardized condensation risk index D using a piecewise function: when When D=0, it means there is no risk of condensation. When D increases linearly as the margin decreases, When D=1, it represents high risk. The final generated D serves as the core quantitative input for the second index in stage S4, providing accurate condensation risk basis for subsequent control decisions.

[0079] Compared to traditional control logics that only monitor the environmental dew point or a single point of internal temperature, this condensation risk assessment scheme has several significant advantages: Firstly, by sampling the lowest temperature at multiple locations and combining it with high-precision calculations using the Magnus formula, it accurately locates the most condensation-prone risk points inside the laser, effectively avoiding missed detections such as "condensation has occurred inside the laser even though the environment has not reached the dew point" or "localized condensation has not been detected." Secondly, by mapping continuous temperature differences into a standardized 0-1 risk index, it preserves the continuity of risk changes and achieves quantitative alignment with the comprehensive health index, facilitating subsequent multi-module collaborative processing such as dual-index strategy matching and digital twin prediction, thus solving the pain point of difficulty in unified decision-making for heterogeneous parameters. Furthermore, the design of the three-level risk range takes into account both safety and energy efficiency performance. Redundant heating is not activated at times, reducing ineffective energy consumption; the intermediate transition range adjusts the D value through linear gradient to achieve proactive pre-intervention, avoiding the thermal shock to optical devices caused by passive high-power heating after condensation occurs; in the high-risk range, D=1 triggers the highest level of protection to ensure the safety of core optical components, and the entire calculation logic can be implemented using only existing temperature and humidity sensors without additional hardware costs, adapting to the condensation control needs of air-cooled lasers of various power levels. Step S3 includes: constructing a digital twin model of the air-cooled laser, the digital twin model including a thermodynamic sub-model, a photoelectric conversion sub-model, a fan performance sub-model, and an environmental heat exchange sub-model; using the standard thermal resistance curve, photoelectric conversion efficiency benchmark curve, fan performance curve, and environmental heat exchange coefficient of the laser in a brand-new state as the initial parameters of the digital twin model; inputting the current internal operating parameters, external environmental parameters, and various health sub-scores into the digital twin model; based on the current parameters and the preset operating condition change trend, the digital twin model outputs the predicted value of the comprehensive health index and the predicted value of the condensation risk index within a future time window, the preset operating condition change trend including the preset output power change curve, fan speed change curve, and environmental temperature change curve of the laser within the future time window.

[0080] At the model construction level, the digital twin model integrates four mutually coupled physical sub-models: the thermodynamic sub-model is responsible for describing the heat generation, conduction, convection and dissipation processes between the pump source, gain medium, heat sink and heat sink, and calculates the temperature distribution and evolution trend of key points inside the laser based on the current thermal resistance state and heat dissipation conditions; the photoelectric conversion sub-model describes the decay law of the electrical energy to light energy conversion efficiency with operating conditions and degradation degree based on the relationship between pump drive current, real-time photoelectric conversion efficiency and output optical power; the fan performance sub-model calculates the real-time air volume and air pressure based on the current fan speed, air duct resistance characteristics and bearing degradation state; and the environmental heat exchange sub-model describes the dynamic interaction between the external environmental temperature and humidity, the air state at the air duct inlet and the internal thermal balance of the laser. The four sub-models are coupled with each other (the temperature output by the thermodynamic sub-model affects the photoelectric conversion efficiency, and the heat loss calculated by the photoelectric conversion sub-model has a reaction effect on the thermodynamic boundary; the fan performance sub-model determines the convective heat transfer intensity, and the environmental heat exchange sub-model sets the external boundary conditions), together forming a complete multiphysics simulation framework for air-cooled lasers with thermal-electrical-fluid-loop characteristics.

[0081] At the predictive execution level, the system first uses the standard thermal resistance curve of the laser in its new state, the benchmark curve of photoelectric conversion efficiency, the fan performance curve, and the environmental heat exchange coefficient as the initial parameters of the model. Then, it inputs the currently collected internal operating parameters (temperature, power, current, speed, etc.), external environmental parameters (temperature and humidity), and the health sub-scores calculated in step S2 into the model. Using this as the initial state, the digital twin model, combined with preset future operating condition trends (including preset output power change curves, fan speed change curves, and ambient temperature change curves within the future time window), recursively calculates through the coupling of the four sub-models to output the predicted values ​​of the comprehensive health index and the condensation risk index within the future time window, providing a quantitative basis for the forward-looking decision-making in step S4.

[0082] Compared to traditional pure data-driven prediction methods or fixed-parameter offline simulation methods, this digital twin prediction scheme has significant technical advantages: First, it introduces three types of health sub-scores as model correction inputs, breaking the limitation of traditional simulation models that "never age," allowing the digital twin state to be updated synchronously with the actual degradation of the equipment, and the prediction results accurately reflect the future operating performance under the current aging state; Second, it incorporates the fan speed change curve into the operating condition trend input for the first time, fully considering the feedback effect of control actions on the system state, solving the pain point of traditional prediction ignoring control behavior and only simulating the "laissez-faire" state, and the prediction results are highly consistent with the actual control scenario; Third, the four physical sub-models realize multi-physics field coupling calculation, and the prediction results of comprehensive health and condensation risk are no longer isolated single-dimensional indicators, but rather consider the linkage results of the interaction of multiple factors such as heat, electricity, current, and environment, providing a highly reliable forward-looking decision-making basis for the subsequent dual-index strategy matching in the S4 stage. Furthermore, this model is built based on physical mechanisms and does not require training with massive amounts of historical operating data. It can be activated immediately upon commissioning of new machines, has strong generalization capabilities, and can adapt to the prediction needs of wind-cooled lasers with different power levels and operating conditions, providing core technical support for health management throughout the entire life cycle.

[0083] The matching rules for the multi-dimensional strategy mapping relationship in step S4 specifically include: defining the change in health status. ,in The current overall health index, The predicted value of the comprehensive health index; the change in condensation risk is defined. ,in This represents the current condensation risk index. To predict the condensation risk index, the following thresholds are set: The threshold for health status. The threshold for sub-health status, and ; The low-risk threshold; This is the lower limit threshold for medium risk. It is the upper limit threshold of medium risk, and ; The tolerance threshold for declining health status To set a tolerance threshold for the increased risk of condensation, The threshold for triggering a safe shutdown; when ,and , , When, it matches the standard temperature control mode; when ,and or Matching the anti-condensation priority mode; when ,and , When, match the enhanced heat dissipation mode; when ,and When, match the compromise mode; when ,or At that time, match the safe shutdown mode.

[0084] This multi-dimensional strategy mapping scheme has several groundbreaking advantages over traditional single-indicator and static threshold control logic: First, it pioneers a dual-index decision-making mechanism of "current state + future trend," which considers both the current operating level of the equipment and predicts its future evolution direction. This solves the pain point of traditional control relying solely on the current snapshot and having a delayed response. For example, it can intervene in advance when the current health level is acceptable but a sharp drop is predicted, thus preventing failures. Second, the three-layer threshold system and five-modal design achieve full-scenario coverage without blind spots. All combinations of operating conditions, from healthy to deteriorating health and from low to high condensation risk, have corresponding strategies, and the boundaries between modes are clear. The absence of logical overlap avoids control command conflicts; third, the introduction of tolerance thresholds creates an anti-disturbance buffer, preventing frequent mode switching when health and condensation risk fluctuate slightly around the threshold, significantly improving system stability; fourth, the "OR" logic design for safe shutdown takes into account both absolute degradation of the current state and sudden deterioration of future trends, providing dual protection for equipment safety and eliminating the risk of missed judgments from a single threshold; fifth, the targeted design of compromise modes effectively solves the common industry contradiction of "enhanced heat dissipation easily leads to condensation, while anti-condensation easily leads to overheating" in high humidity environments, expanding the applicability of lasers in high humidity and variable operating conditions.

[0085] When matching the standard temperature control mode, the forward-looking compensation is zero; when matching the anti-condensation priority mode, when At that time, the forward-looking compensation amount includes the preheating amount of the duct heating device; when At that time, the forward-looking compensation amount includes the preheating amount of the duct heating device and the pre-reduction amount of the output power, wherein The threshold for predicting condensation risk; when matching the enhanced heat dissipation mode, when At that time, the forward-looking compensation amount includes the pre-increment of fan speed; when At that time, the forward compensation amount includes the pre-increment of fan speed and the pre-decrease of output power, wherein Predict the trigger threshold for health status; when matching the trade-off modality, when and At that time, the forward-looking compensation amount includes a weighted combination of the fan speed pre-increment and the preheating amount of the duct heating device; when and In this context, the forward-looking compensation amount includes a weighted combination of the pre-increment of fan speed, the pre-decrease of output power, and the pre-heating amount of the duct heating device; in the compromise mode, the weighted combination includes the heat dissipation-related weights. =( ) / ( ), weighting related to condensation prevention ( - ),Will and After normalization, the final weights are obtained. , The forward-looking compensation amount of the weighted combination = preset maximum pre-increment of fan × +Preset maximum preheating heat × When the forward-looking compensation amount also includes the output power pre-drop amount, the output power pre-drop amount is executed in full according to the preset maximum power pre-drop amount, and is superimposed with the forward-looking compensation amount of the above weighted combination.

[0086] This step follows the multi-dimensional strategy matching results, quantifying the forward-looking compensation amount for each of the five control modes to achieve an active control closed loop of "prediction-pre-adjustment": In the standard temperature control mode, the system is in optimal operating condition and requires no pre-intervention, so the forward-looking compensation amount is zero; in the anti-condensation priority mode, when the increase in condensation risk does not exceed the trigger threshold, the internal temperature is increased only by preheating through the air duct heating device to expand the condensation risk margin and avoid the thermal shock of passive decondensation; when the increase in risk exceeds the threshold, the output power pre-reduction amount is superimposed to reduce internal heat generation and the introduction of humid air from the source, doubly blocking the risk of condensation; in the enhanced heat dissipation mode... In the lower mode, when the health status decline trend does not exceed the threshold, only the fan speed pre-increment is increased to improve heat dissipation capacity and delay degradation; when the health status drops sharply and exceeds the threshold, the output power pre-reduction is superimposed to reduce the heat load and avoid thermal failure; in the compromise mode, at the weight allocation level of the compromise mode, the system introduces a dynamic weight calculation mechanism with the current health status and condensation risk index as independent variables. When the output power pre-reduction is triggered, the preset maximum power pre-reduction is fully executed and superimposed with the aforementioned compensation amount, so that the system can intervene with maximum force and quickly block the degradation process when facing extreme conditions where both health status and condensation risk exceed the limit.

[0087] The multi-dimensional strategy mapping relationship includes an initial mapping table and historical correction records; after each execution of the target control mode, the actual control effect evaluation index is recorded, and the evaluation index includes at least the first deviation between the actual rate of change and the predicted rate of change of the comprehensive health index after control. The second deviation between the actual rate of change and the predicted rate of change of the controlled condensation risk index. Overall evaluation value ,in and The preset deviation weighting coefficient; when the comprehensive evaluation value When the cumulative offset exceeds a preset update threshold, a self-correction process for the mapping table is triggered; the self-correction process involves adjusting the comprehensive evaluation value from the most recent N control records. For the target control mode corresponding to the highest control record, adjust the mode matching priority under the corresponding index condition, promote the target control mode with the highest comprehensive evaluation value to the first priority, and at the same time, demote the other target control modes in order of comprehensive evaluation value from high to low.

[0088] This step is the self-optimizing closed-loop implementation of strategy matching in stage S4. It involves dynamic iteration based on a pre-stored multi-dimensional strategy initial mapping table and historical correction records: After each target control mode is executed, the actual rate of change of the comprehensive health index and condensation risk index after control is synchronously collected and compared with the corresponding predicted rate of change output in S3. The first deviation is calculated for each. Second deviation Substitute the two types of deviations into the formula for calculating the comprehensive evaluation value. This formula ensures that the closer the control effect is to the prediction, the closer E is to 1, and the better the control effect. The cumulative offset of the comprehensive evaluation value E is continuously calculated. When the cumulative offset exceeds the preset update threshold, the mapping table self-correction process is triggered: the most recent N control records are retrieved, the historical control case with the highest comprehensive evaluation value E is selected, and its corresponding target control mode is adjusted to the first priority under the current index condition. The remaining candidate control modes are degraded in order of comprehensive evaluation value from high to low. The updated mapping relationship is synchronously stored in the historical correction record for subsequent strategy matching.

[0089] This self-correcting scheme overcomes the limitations of traditional fixed-strategy mapping, which is "unchanging," and possesses several significant advantages: First, its self-optimization mechanism, based on its own historical control performance, can adapt to the manufacturing tolerances of different batches of lasers, the differences in operating conditions of different deployment environments, and the aging characteristics changes throughout the equipment's lifecycle, without relying on manual debugging or massive offline training data. This solves the problems of poor generalization of fixed strategies and high costs of manual debugging. Second, it adopts an exponential comprehensive evaluation value formula, which is far more sensitive to control deviations than linear evaluation methods. It can accurately distinguish minute differences in control performance and provide a highly reliable quantitative basis for priority adjustment. Third, the cumulative offset triggering mechanism sets up an anti-disturbance buffer, avoiding erroneous corrections caused by single, accidental fluctuations in operating conditions and ensuring the stability of strategy updates. Fourth, the priority adjustment logic based on "historical best cases" ensures that strategy iterations always converge towards better control performance, preventing strategy degradation.

[0090] Degradation patterns are identified based on the relative values ​​of the thermal resistance health sub-score, photoelectric efficiency health sub-score, and fan performance health sub-score: When the thermal resistance health sub-score is lower than both the photoelectric efficiency health sub-score and the fan performance health sub-score, the dominant degradation pattern is determined to be heat dissipation system degradation, and a first maintenance recommendation to clean the heat dissipation channels is generated; when the photoelectric efficiency health sub-score is lower than both the thermal resistance health sub-score and the fan performance health sub-score, the dominant degradation pattern is determined to be pump source or optical component degradation, and a second maintenance recommendation to check the pump source or optical component is generated; when the fan performance health sub-score is lower than both the thermal resistance health sub-score and the photoelectric efficiency health sub-score, the dominant degradation pattern is determined to be fan mechanical degradation, and a third maintenance recommendation to replace the fan is generated; when the thermal resistance health sub-score, photoelectric efficiency health sub-score, and fan performance health sub-score are all lower than their respective preset severe thresholds, it is determined to be comprehensive degradation, and a fourth maintenance recommendation to overhaul the entire machine is generated.

[0091] This step is based on the relative magnitudes of three health sub-scores—thermal resistance, photoelectric efficiency, and fan efficiency—obtained through decoupling in stage S2 to accurately identify degradation modes: First, the three sub-scores are compared pairwise to pinpoint the dominant degradation source of the current device through their relative numerical relationships. If the thermal resistance health sub-score is simultaneously lower than both the photoelectric efficiency and fan efficiency health sub-scores, it indicates the largest increase in thermal resistance and the most severe degradation of the heat dissipation path, thus determining the dominant degradation mode as heat dissipation system degradation. Based on this, the first maintenance recommendation is to clean the heat sink fins and remove dust from the ventilation ducts. If the photoelectric efficiency health sub-score is the smallest, it indicates the most significant degradation in photoelectric conversion performance, thus determining the dominant degradation mode as heat dissipation system degradation. The degradation mode is aging of core optical components such as pump sources or gain media, generating a second maintenance suggestion to check the collimation of the optical path and detect the electro-optic parameters of the pump source; if the fan efficiency health sub-score is the lowest, it indicates that the energy ratio of the vibration characteristic frequency is abnormal and the mechanical transmission performance is reduced, and the dominant degradation mode is determined to be mechanical degradation caused by fan bearing wear or blade dust accumulation, generating a third maintenance suggestion to replace the fan or lubricate the bearing; if all three sub-scores are lower than their respective preset severe thresholds, it indicates that the equipment has no single dominant degradation source and presents a comprehensive degradation characteristic of balanced aging of multiple components, generating a fourth maintenance suggestion to conduct a systematic overhaul of the entire laser.

[0092] This invention directly identifies specific fault types through pairwise comparisons of three health sub-scores, providing maintenance teams with clear repair directions and significantly shortening fault diagnosis and repair decision-making time. The system generates targeted maintenance suggestions based on the identification results, ensuring that maintenance actions are focused and avoiding excessive repairs such as "total replacement and repair" due to unclear causes, thus significantly reducing maintenance costs.

[0093] In another embodiment of the present invention, an adaptive control system for an air-cooled laser based on health status assessment is provided, comprising: a data acquisition module configured to synchronously acquire the laser's internal operating parameters, external environmental parameters, and acoustic vibration signals;

[0094] The health assessment module, connected to both the data acquisition module and the strategy matching module, is configured to calculate health sub-scores for thermal resistance, photoelectric efficiency, and fan performance based on the internal operating parameters, fuse them to generate a comprehensive health index, and calculate a condensation risk index based on external environmental parameters. The digital twin prediction module, connected to the data acquisition module, is configured to construct thermodynamic and photoelectric conversion models, outputting predicted values ​​for the comprehensive health index and condensation risk index within a future time window. The strategy matching module, connected to both the health assessment module and the digital twin prediction module, is configured to use the current index value and predicted trend as an index to match the target control mode through a multi-dimensional mapping relationship and calculate a forward-looking compensation amount. The execution control module, connected to the strategy matching module, is configured to respond to the target control mode by generating coordinated commands to adjust the air-cooling heat dissipation capacity, laser output power, and duct heating amount in a coordinated manner.

[0095] This system serves as the hardware execution platform corresponding to the aforementioned methods. It achieves adaptive closed-loop control of the laser through the coordinated operation of five functional modules: the data acquisition module, acting as the perception layer, synchronously collects internal laser operating parameters (such as temperature, current, and optical power), external environmental parameters (such as temperature and humidity), and acoustic vibration signals through multiple types of sensors, providing high-real-time multi-source data input for upper-level decision-making; the health assessment module, as one of the core processing modules, receives the collected data and executes a multi-dimensional decoupling algorithm—calculating health sub-scores based on internal operating parameters such as thermal resistance, photoelectric efficiency, and fan efficiency, fusing them using DS evidence theory to generate a comprehensive health index, and simultaneously calculating a condensation risk index based on environmental temperature and humidity, thus accurately quantifying the laser's current health status and condensation risk; the digital twin prediction module is built-in... The thermodynamic and photoelectric conversion model, using the health sub-score as the correction input, combines current operating parameters with preset operating condition trends. Through multi-physics coupling simulation, it outputs the predicted values ​​of the comprehensive health index and condensation risk index within the future time window, achieving forward-looking prediction of degradation trends. The strategy matching module, as the decision-making center, uses the current index value and predicted change trend as dual indices to retrieve preset multi-dimensional mapping relationships to match the corresponding target control modes (such as standard temperature control, anti-condensation priority, etc.), and calculates the forward-looking compensation amount based on deviation analysis. The execution control module, as the execution layer, responds to the decision instructions of the strategy matching module, generates collaborative control instructions to adjust the heat dissipation capacity of the air-cooled system (such as fan speed), the output power of the laser, and the heating amount of the air duct heating device, completing the closed-loop adjustment from perception to control.

[0096] Compared to traditional discrete laser control systems, this system boasts significant advantages in integration and intelligence: First, its modular architecture achieves a closed-loop process encompassing perception, assessment, prediction, decision-making, and execution. Each module has a clearly defined function and standardized interfaces, facilitating system debugging and maintenance. Upgrading individual modules can also improve overall performance, significantly reducing lifecycle maintenance costs. Second, the collaborative work of the health assessment module and the digital twin prediction module overcomes the limitations of traditional systems that prioritize current conditions over future trends, shifting control strategies from passive response to proactive prediction and effectively preventing thermal failure and condensation damage. Third, the multi-dimensional mapping and forward-looking compensation mechanism of the strategy matching module effectively resolves the conflict between heat dissipation requirements and condensation prevention, improving the laser's stability under complex operating conditions. Fourth, the system can be directly implemented on existing air-cooled laser hardware platforms, requiring only embedded software upgrades, resulting in low modification costs and strong compatibility. This provides reliable hardware support for the unattended and long-term stable operation of high-power air-cooled lasers in industrial settings.

[0097] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. All equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this invention should still be covered by the claims of this invention.

Claims

1. An adaptive control method for wind-cooled lasers based on health status assessment, characterized in that, Includes the following steps: S1, synchronously acquires the laser's internal operating parameters, external environmental parameters, and acoustic vibration parameters; S2, based on the internal operating parameters, the thermal resistance health sub-score and photoelectric efficiency health sub-score are decoupled and calculated through a preset health assessment model; the acoustic vibration parameters are decomposed in the time-frequency domain to extract the energy proportion of the characteristic frequency of the fan bearing and the passing frequency of the blades, and the fan efficiency health sub-score is calculated. The health sub-scores are merged to generate a comprehensive health index characterizing the degree of degradation of the laser body; based on the external environmental parameters, a condensation risk index characterizing the possibility of condensation is calculated. S3, input the internal operating parameters, external environmental parameters and each health sub-score into the pre-constructed digital twin model, and output the predicted value of the comprehensive health index and the predicted value of the condensation risk index within the future time window; S4, using the current value and predicted value change trend of the comprehensive health index as the first index, and the current value and predicted value change trend of the condensation risk index as the second index, the target control mode and forward-looking compensation amount are matched and determined through the preset multi-dimensional strategy mapping relationship. The target control mode includes standard temperature control mode, anti-condensation priority mode, enhanced heat dissipation mode, compromise mode and safe shutdown mode. S5, in response to the target control mode and the forward compensation amount, generates a cooperative control command to dynamically adjust the heat dissipation capacity of the air-cooling system, the output power of the laser, and the heating amount of the air duct heating device.

2. The adaptive control method for wind-cooled lasers based on health status assessment according to claim 1, characterized in that, The process of merging the various health sub-scores to generate a comprehensive health index specifically includes: The thermal resistance health sub-score, photoelectric efficiency health sub-score, and fan efficiency health sub-score are considered as three independent sources of evidence. The real-time operating status of the sensors corresponding to each evidence source is obtained. The operating status includes signal noise level, sampling stability and historical drift record. The original health sub-score value is corrected based on the real-time operating status to obtain the corrected health sub-score. When the real-time working state meets the preset reliability requirements, based on the preset mapping relationship, the corrected health sub-score is converted into a basic confidence assignment value characterizing the laser in a healthy state, a sub-healthy state, and a degraded state. When the real-time working state does not meet the preset reliability requirements, the evidence source data is determined to be unreliable, and the basic credibility allocation value of the evidence source is reassigned as: m(Θ) = 1, m(healthy) = 0, m(sub-healthy) = 0, m(deterioration) = 0, where Θ is the identification framework and Θ = {healthy, sub-healthy, deterioration}. Based on the cumulative runtime of the air-cooled laser, the fusion weight of the basic credibility allocation value of each evidence source in the fusion process is dynamically adjusted, and the basic credibility allocation value after weight adjustment is determined as the final credibility allocation value of the evidence source. Using the combination rules based on DS evidence theory, the final credibility allocation values ​​of the three evidence sources are fused to obtain the fused comprehensive credibility distribution; Based on the integrated credibility distribution, the integrated health index is generated according to preset weight coefficients, where healthy state, sub-healthy state and degenerative state correspond to different weight coefficients.

3. The adaptive control method for wind-cooled lasers based on health status assessment according to claim 2, characterized in that, The step of dynamically adjusting the fusion weights of each evidence source based on the cumulative runtime of the wind-cooled laser specifically includes: When the cumulative running time is less than the first preset time, a first weight is assigned to the evidence source corresponding to the photoelectric efficiency health sub-score, and the first weight is greater than the sum of the weights of the other two evidence sources; When the cumulative runtime is greater than or equal to the first preset runtime and less than the second preset runtime, a second weight is assigned to the evidence source corresponding to the thermal resistance health sub-score, and the second weight is greater than the sum of the weights of the other two evidence sources; When the cumulative runtime is greater than or equal to the second preset runtime, equal weights are assigned to each evidence source. The first preset duration is less than the second preset duration.

4. The adaptive control method for wind-cooled lasers based on health status assessment according to claim 1, characterized in that, The calculation of the condensation risk index, which characterizes the likelihood of condensation, based on the external environmental parameters in step S2 specifically includes: Dew point temperature is calculated using the Magnus formula based on ambient temperature and relative humidity. ; Obtain the lowest internal temperature of the laser The lowest internal temperature of the laser is obtained in real time by multiple temperature sensors distributed on the surface of the pump source, gain medium and heat sink; Calculate the condensation risk margin ,when When indicates the existence of a safety margin, This indicates that the dew point temperature has been reached or exceeded. Set a safe temperature difference threshold and dangerous temperature difference threshold ,and ; when At that time, the condensation risk index ;when At that time, the condensation risk index ;when At that time, the condensation risk index .

5. The adaptive control method for wind-cooled lasers based on health status assessment according to claim 1, characterized in that, Step S3 includes: A digital twin model of an air-cooled laser is constructed, comprising a thermodynamic sub-model, a photoelectric conversion sub-model, a fan performance sub-model, and an environmental heat exchange sub-model. The standard thermal resistance curve, photoelectric conversion efficiency benchmark curve, fan performance curve, and environmental heat exchange coefficient of the laser in its new state are used as the initial parameters of the digital twin model. Input the current internal operating parameters, external environmental parameters, and health sub-scores into the digital twin model; The digital twin model outputs a comprehensive health index prediction and a condensation risk index prediction within a future time window based on current parameters and preset operating condition change trends. The preset operating condition change trends include preset output power change curves, fan speed change curves, and ambient temperature change curves of the laser within the future time window.

6. The adaptive control method for wind-cooled lasers based on health status assessment according to claim 1, characterized in that, The matching rules for the multi-dimensional strategy mapping relationship in step S4 specifically include: defining the change in health status. ,in The current overall health index, The predicted value of the comprehensive health index; the change in condensation risk is defined. ,in This represents the current condensation risk index. This is the predicted value of the condensation risk index; Set the following thresholds, The threshold for health status. The threshold for sub-health status, and ; The low-risk threshold; This is the lower limit threshold for medium risk. It is the upper limit threshold of medium risk, and ; The tolerance threshold for declining health status To set a tolerance threshold for the increased risk of condensation, The threshold for triggering a safe shutdown; when ,and , , At that time, it matches the standard temperature control mode; when ,and or Match the anti-condensation priority mode; when ,and , At the same time, match the enhanced heat dissipation mode; when ,and At that time, the compromise mode is matched; when ,or At that time, match the safe shutdown mode.

7. The adaptive control method for wind-cooled lasers based on health status assessment according to claim 6, characterized in that, When matching the standard temperature control mode, the forward compensation amount is zero; When matching the anti-condensation priority mode, when At that time, the forward-looking compensation amount includes the preheating amount of the duct heating device; when At that time, the forward-looking compensation amount includes the preheating amount of the duct heating device and the pre-reduction amount of the output power, wherein The threshold for predicting condensation risk; When matching enhanced heat dissipation mode, when At that time, the forward compensation amount includes the pre-increment of fan speed; when At that time, the forward compensation amount includes the pre-increment of fan speed and the pre-decrease of output power, wherein Predict trigger thresholds for health status; When matching compromise modes, when and At that time, the forward-looking compensation amount includes a weighted combination of the fan speed pre-increment and the preheating amount of the duct heating device; when and In this context, the forward-looking compensation amount includes a weighted combination of the pre-increment of fan speed, the pre-decrease of output power, and the pre-heating amount of the duct heating device; in the compromise mode, the weighted combination includes the heat dissipation-related weights. =( ) / ( ), related weights for preventing condensation ( - ),Will and After normalization, the final weights are obtained. , The forward-looking compensation amount of the weighted combination = preset maximum pre-increment of fan × +Preset maximum preheating heat × When the forward-looking compensation amount also includes the output power pre-drop amount, the output power pre-drop amount is executed in full according to the preset maximum power pre-drop amount, and is superimposed with the forward-looking compensation amount of the above weighted combination.

8. The adaptive control method for wind-cooled lasers based on health status assessment according to claim 7, characterized in that, The multidimensional strategy mapping relationship includes an initial mapping table and historical correction records; After each execution of the target control mode, the actual control effect evaluation index is recorded. The evaluation index includes at least the following: The first deviation between the actual rate of change and the predicted rate of change of the overall health index after control. ; The second deviation between the actual and predicted rates of change of the controlled condensation risk index ; Overall evaluation value ,in and The preset deviation weighting coefficient; When the comprehensive evaluation value When the cumulative offset exceeds the preset update threshold, the self-correction process of the mapping table is triggered. The self-correction process is based on the comprehensive evaluation value from the most recent N control records. For the target control mode corresponding to the highest control record, adjust the mode matching priority under the corresponding index condition, promote the target control mode with the highest comprehensive evaluation value to the first priority, and at the same time, demote the other target control modes in order of comprehensive evaluation value from high to low.

9. The adaptive control method for wind-cooled lasers based on health status assessment according to claim 1, characterized in that, Degradation pattern identification is performed based on the relative magnitudes of the thermal resistance health sub-score, photoelectric efficiency health sub-score, and fan performance health sub-score. When the thermal resistance health sub-score is less than both the photoelectric efficiency health sub-score and the fan performance health sub-score, the dominant degradation mode is determined to be the heat dissipation system degradation, and the first maintenance recommendation to clean the heat dissipation channel is generated. When the photoelectric efficiency health sub-score is less than both the thermal resistance health sub-score and the fan efficiency health sub-score, the dominant degradation mode is determined to be pump source or optical device degradation, and a second maintenance recommendation to check the pump source or optical device is generated. When the fan performance health sub-score is lower than both the thermal resistance health sub-score and the photoelectric efficiency health sub-score, the dominant degradation mode is determined to be fan mechanical degradation, and a third maintenance recommendation to replace the fan is generated. When the thermal resistance health sub-score, photoelectric efficiency health sub-score, and fan performance health sub-score are all below their respective preset severe thresholds, it is judged as a comprehensive degradation, and a fourth maintenance suggestion for whole-machine overhaul is generated.

10. A system for implementing the method according to any one of claims 1-9, characterized in that, include: The data acquisition module is configured to synchronously acquire the laser's internal operating parameters, external environmental parameters, and acoustic vibration signals. The health assessment module is connected to the data acquisition module and the strategy matching module respectively. It is configured to calculate the health sub-scores of thermal resistance, photoelectric efficiency and fan performance based on the internal operating parameters, integrate them to generate a comprehensive health index, and calculate the condensation risk index based on external environmental parameters. The digital twin prediction module is connected to the data acquisition module and is configured to construct a thermodynamic and photoelectric conversion model to output the predicted values ​​of the comprehensive health index and the condensation risk index within a future time window. The strategy matching module is connected to the health assessment module and the digital twin prediction module respectively. It is configured to match the target control mode through a multi-dimensional mapping relationship using the current index value and the predicted trend of change as indexes, and calculate the forward compensation amount. The execution control module is connected to the strategy matching module and is configured to generate cooperative instructions in response to the target control mode to adjust the air cooling capacity, laser output power and air duct heating amount in a coordinated manner.

Images