A power knife tower thermal management optimization method and system based on intelligent cooling

Abstract

The present application relates to the technical field of tool turret thermal management, and particularly relates to a power tool turret thermal management optimization method and system based on intelligent cooling, which comprises the following steps: acquiring temperature distribution data and associated working condition parameters of the power tool turret in real time, and establishing a thermal state feature vector representing the dynamic relationship between tool turret heat generation and heat dissipation capacity; constructing a cooling decision model, inputting the state vector into the cooling decision model to generate an initial control strategy; simulating the execution of the initial control strategy in a digital twin environment, and triggering a strategy reconstruction mechanism when the predicted temperature suppression rate is lower than a set threshold; according to the machining quality constraint and the equipment safety threshold, performing real-time dynamic optimization on the verified initial control strategy, and superimposing a thermal inertia delay compensation amount to generate a final control strategy. Through the present application, the temperature control failure problem caused by thermal inertia time lag is effectively solved, the precise space-time matching of cooling instructions and thermal load peaks is realized, and thermal deformation is significantly inhibited.

Description

Technical Field

[0001] This invention relates to the field of turret thermal management technology, and in particular to a method and system for optimizing thermal management of power turrets based on intelligent cooling. Background Technology

[0002] In the field of thermal management technology for CNC machine tool power turrets, traditional methods generally rely on temperature-triggered control strategies based on fixed thresholds. Such solutions monitor local temperature signals and initiate cooling operations when the temperature exceeds a preset limit. Although this can achieve basic temperature suppression, it has significant thermal inertia timing mismatch problems in dynamic machining scenarios.

[0003] Because the coupling effect of heat conduction delay and system response lag is not fully considered, the actual cooling effect and the heat generation process are difficult to precisely match on the time axis. This results in the cooling command taking effect significantly lagging behind the actual heat load changes. This timing mismatch not only causes alternating oscillations of overcooling or undercooling, but also exacerbates the thermal stress cycle of critical components, affecting equipment reliability in the long term. Especially in high-speed machining and complex contour cutting, the control delay caused by thermal inertia significantly amplifies the temperature fluctuation amplitude, directly causing workpiece thermal deformation exceeding tolerances and abnormal tool wear, severely restricting the realization of high-precision machining. Existing technologies lack a predictive compensation mechanism for thermal dynamic processes, making it difficult to fundamentally solve the essential contradiction of timing mismatch, becoming a key bottleneck restricting the improvement of thermal management performance of power turrets. Summary of the Invention

[0004] This invention provides a method and system for optimizing the thermal management of power turrets based on intelligent cooling, thereby effectively solving the problems pointed out in the background art.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A method for optimizing the thermal management of power turrets based on intelligent cooling includes:

[0007] Real-time acquisition of temperature distribution data and related operating parameters of the power turret; establishment of a thermal state feature vector characterizing the dynamic relationship between turret heat generation and heat dissipation capacity.

[0008] A cooling decision model is constructed based on historical cooling strategies. The state vector is input into the cooling decision model to generate an initial control strategy.

[0009] The initial control strategy is simulated and executed in a digital twin environment, and a strategy reconstruction mechanism is triggered when the predicted temperature suppression rate is lower than a set threshold.

[0010] Based on processing quality constraints and equipment safety thresholds, the verified initial control strategy is dynamically optimized in real time, and thermal inertia delay compensation is added to generate the final control strategy.

[0011] Furthermore, the method also includes: when an undefined processing scenario is detected, matching a baseline strategy based on the similarity evaluation results between the current state vector and historical strategies, and updating the cooling decision model through incremental learning.

[0012] Furthermore, real-time temperature distribution data and associated operating parameters of the power turret are acquired to establish a thermal state feature vector characterizing the dynamic relationship between turret heat generation and heat dissipation capacity, including:

[0013] Real-time acquisition of thermal dynamic correlation parameters;

[0014] Based on the aforementioned thermal dynamic correlation parameters, the real-time matching relationship between the heat generation characteristics and heat dissipation characteristics of the turret is analyzed.

[0015] The thermal state feature vector characterizing the thermal equilibrium state is generated based on the dynamic relationship between the heat generation feature and the heat dissipation feature.

[0016] Furthermore, a cooling decision-making model is constructed based on historical cooling strategies, including:

[0017] Data quality enhancement processing is performed on historical operating data to extract effective cooling strategy fragments and corresponding thermal state feature vectors;

[0018] Establish a multimodal model architecture that includes policy decision trees and deep learning networks;

[0019] Using the thermal state feature vector as input and the optimal cooling strategy as output target, the multimodal model architecture is trained under a reinforcement learning framework to generate an initial cooling decision model;

[0020] The cooling decision model is dynamically updated through an online incremental learning mechanism.

[0021] Furthermore, the cooling decision model is dynamically updated through an online incremental learning mechanism, including:

[0022] Calculate the matching degree between the newly added strategy fragment and the current working condition. The matching degree is evaluated based on the feature space distance of the state feature vector and the tool load similarity.

[0023] The weight allocation of the policy fragments is dynamically adjusted based on the matching degree, wherein policy fragments with a matching degree higher than a threshold are given priority in model parameter updates.

[0024] Furthermore, simulating the execution of the initial control strategy in a digital twin environment includes:

[0025] A multi-physics coupled digital twin is constructed based on the state vector and the three-dimensional thermodynamic model of the turret, wherein the digital twin maps the dynamic process of heat generation and heat dissipation of the turret in real time.

[0026] The initial control strategy is loaded into the digital twin, and after simulation execution, the predicted temperature distribution and thermal deformation of key components are output.

[0027] Furthermore, a policy reconfiguration mechanism is triggered when the predicted temperature suppression rate falls below a set threshold, including:

[0028] The predicted temperature suppression rate is calculated using the following formula: Where, η suppress To predict the temperature suppression rate, T actual T represents the current measured temperature. sim To simulate temperature, T target The target temperature threshold;

[0029] When the predicted temperature suppression rate is lower than the first set threshold, the strategy reconstruction mechanism is triggered;

[0030] The reconstruction strategy is re-imported into the digital twin environment for iterative verification until the predicted temperature suppression rate reaches the second set threshold for output.

[0031] Furthermore, based on processing quality constraints and equipment safety thresholds, the validated initial control strategy is dynamically optimized in real time, and thermal inertia delay compensation is added to generate a final control strategy, including:

[0032] The thermal variation deviation limit in the processing quality constraints is combined with the temperature rise rate of key components in the equipment safety threshold to generate a dynamic constraint space.

[0033] A temperature suppression-energy consumption balance model is constructed based on the thermal state feature vector, and a Pareto optimal strategy set is generated based on an evolutionary algorithm.

[0034] The Pareto optimal strategy set is subjected to timing compensation correction for thermal inertia and execution delay;

[0035] Based on the real-time operating conditions, a control strategy that meets the thermal stability margin requirement is selected from the corrected strategy set as the final control strategy.

[0036] A power turret thermal management optimization system based on intelligent cooling includes:

[0037] The state vector construction module acquires real-time temperature distribution data and related operating parameters of the power turret, and establishes a state vector characterizing the dynamic relationship between the turret's heat generation and heat dissipation capacity.

[0038] The initial strategy generation module constructs a cooling decision model based on historical cooling strategies, inputs the state vector into the cooling decision model, and generates an initial control strategy.

[0039] The reconstruction mechanism execution module simulates the execution of the initial control strategy in the digital twin environment, and triggers the strategy reconstruction mechanism when the predicted temperature suppression rate is lower than the set threshold.

[0040] The final strategy generation module dynamically optimizes the verified initial control strategy in real time based on processing quality constraints and equipment safety thresholds, and adds thermal inertia delay compensation to generate the final control strategy.

[0041] Furthermore, the state vector construction module includes:

[0042] The correlation parameter acquisition unit acquires thermal dynamic correlation parameters in real time.

[0043] The feature analysis unit analyzes the real-time matching relationship between the heat generation feature and the heat dissipation feature of the turret based on the thermal dynamic correlation parameters.

[0044] The feature vector characterization unit generates the thermal state feature vector characterizing the thermal equilibrium state based on the dynamic relationship between the heat generation feature quantity and the heat dissipation feature quantity.

[0045] The technical solution of this invention can achieve the following technical effects:

[0046] It effectively solves the problem of temperature control failure caused by thermal inertia time lag, achieves precise spatiotemporal matching between cooling command and peak heat load, and significantly suppresses thermal deformation. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a flowchart illustrating an optimization method for thermal management of a power turret based on intelligent cooling.

[0049] Figure 2 A flowchart illustrating the process of establishing a thermal state feature vector;

[0050] Figure 3 A flowchart illustrating the process of building a cooling decision model;

[0051] Figure 4 A flowchart illustrating the process of dynamically updating the cooling decision model;

[0052] Figure 5 A schematic diagram illustrating the process of simulating the execution of the initial control strategy in a digital twin environment;

[0053] Figure 6 A flowchart illustrating the process of generating the final control strategy. Detailed Implementation

[0054] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0056] Example 1

[0057] like Figure 1 As shown, this invention provides a method for optimizing the thermal management of a power turret based on intelligent cooling, the method comprising:

[0058] S1: Real-time acquisition of temperature distribution data and related operating parameters of the power turret, and establishment of a thermal state feature vector characterizing the dynamic relationship between turret heat generation and heat dissipation capacity;

[0059] Specifically, a high-fidelity thermal state sensing system is constructed to overcome the limitations of traditional single-point temperature control: by integrating temperature field distribution monitoring with multi-dimensional operating parameters, a quantitative characteristic index reflecting the dynamic balance between heat generation and heat dissipation is established, providing a holographic decision-making basis for intelligent temperature control strategies. This process realizes a leap from discrete temperature monitoring to dynamic modeling of thermodynamic systems.

[0060] S2: Construct a cooling decision model based on historical cooling strategies, input the state vector into the cooling decision model, and generate an initial control strategy;

[0061] Specifically, based on historical cooling operation records in massive machining scenarios (such as coolant flow rate adjustment timing, spindle power matching data, etc.), the association rules between operating condition features and optimal control parameters are extracted; a time-series deep learning algorithm is used, with the thermal state feature vector generated in step S1 as input, to dynamically learn the optimal control strategy generation mode under different thermodynamic states; the real-time feature vector is input into the trained decision model, and the output is an initial control command set containing parameters such as cooling intensity, action phase, and duration, which serves as the benchmark strategy for subsequent dynamic optimization.

[0062] S3: Simulate the execution of the initial control strategy in the digital twin environment, and trigger the strategy reconstruction mechanism when the predicted temperature suppression rate is lower than the set threshold;

[0063] Specifically, a virtual verification barrier is constructed, and the initial control strategy is pre-evaluated in a digital twin environment to avoid the risks of trial and error on physical equipment using traditional methods. A high-precision thermodynamic model is used to simulate the temperature field evolution process after the strategy is executed, and the temperature suppression effectiveness is dynamically predicted. When the simulation results show that the suppression rate does not reach the threshold, the strategy reconstruction mechanism is automatically triggered, forming a closed-loop optimization link of "simulation verification → defect identification → strategy iteration". This step essentially establishes a safety isolation layer for the control strategy, eliminating inefficient or high-risk strategies in advance before physical execution, and ensuring the reliability and control accuracy of the thermal management system from the source.

[0064] S4: Based on processing quality constraints and equipment safety thresholds, the verified initial control strategy is dynamically optimized in real time, and thermal inertia delay compensation is added to generate the final control strategy.

[0065] Specifically, by integrating processing accuracy requirements (such as deformation tolerance) and equipment physical limits (such as maximum load of the cooling system), the validated initial strategy is dynamically optimized across multiple objectives. At the same time, thermal inertia delay compensation is introduced to correct the timing phase of the control commands, ultimately generating a globally optimal strategy that takes into account processing quality, equipment safety, and thermal dynamic characteristics. This step, through a dual mechanism of "constraint boundary definition + time delay compensation embedding," overcomes the limitations of traditional static strategies and single-objective optimization, ensuring the accurate implementation and robust execution of control commands in real thermodynamic systems.

[0066] This invention effectively solves the problem of temperature control failure caused by thermal inertia time lag, achieves precise spatiotemporal matching between cooling commands and peak heat load, and significantly suppresses thermally induced deformation.

[0067] As a preferred embodiment of the above, the method further includes: when an undefined processing scenario is detected, matching a benchmark strategy based on the similarity evaluation results between the current state vector and historical strategies, and updating the cooling decision model through incremental learning.

[0068] Specifically, when an undefined processing scenario is detected, a highly correlated reference strategy is extracted and weighted based on a dynamic similarity assessment between the thermal state feature vector and the historical strategy library to generate a baseline control strategy adapted to the new working conditions. Subsequently, the effectiveness of the strategy is verified through a digital twin environment, and its safety is confirmed by thermodynamic simulation. The optimized strategy parameters and scenario features are then used as an incremental dataset to trigger an online learning mechanism for the decision model. An elastic weight solidification algorithm is employed to update the network parameters, embedding the decision logic of the new scenario while retaining the historical knowledge base. This process, through a closed-loop architecture of "scenario matching → hybrid strategy generation → safety verification → model iteration," achieves adaptive evolution of the strategy under unknown working conditions, effectively overcoming the risk of strategy mismatch in novel processing tasks using traditional methods.

[0069] As a preferred embodiment of the above, such as Figure 2 As shown, step S1 involves acquiring real-time temperature distribution data and related operating parameters of the power turret, and establishing a thermal state feature vector characterizing the dynamic relationship between turret heat generation and heat dissipation capacity, including:

[0070] S11: Real-time acquisition of thermal dynamic correlation parameters;

[0071] S12: Analyze the real-time matching relationship between the heat generation characteristics and heat dissipation characteristics of the turret based on thermal dynamic correlation parameters;

[0072] S13: Generate a thermal state feature vector representing the thermal equilibrium state based on the dynamic relationship between the heat generation feature and the heat dissipation feature.

[0073] Specifically, firstly, real-time temperature field distribution data (including spatial temperature gradient and time-varying fluctuation characteristics) and related operating parameters (covering key physical quantities such as spindle power, cutting force spectrum, and cooling medium flow rate) of the power turret are collected. Based on sensor data synchronous calibration technology, time-series deviations are eliminated to construct a raw dataset with spatiotemporal consistency. Subsequently, a heat source-heat dissipation decoupling analysis method is adopted. The main component contribution of electromagnetic loss and mechanical friction heat generation is quantified through a frequency domain loss separation algorithm. Combined with an unsteady heat transfer model, the dynamic coupling relationship between coolant convection heat transfer and structural heat conduction is analyzed, and a real-time matching degree evaluation model of heat generation rate and heat dissipation capacity is established. Finally, based on time-varying matching degree index, temperature field spatial heterogeneity parameters, and thermal inertia delay characteristics, a multi-dimensional thermal state feature vector is constructed. This vector integrates quantitative indicators such as heat accumulation trend, heat dissipation response phase difference, and spatial thermal imbalance degree to form a standardized state descriptor that can dynamically reflect the evolution of the turret's thermal balance, providing high-resolution decision input for the subsequent generation of intelligent temperature control strategies.

[0074] As a preferred embodiment of the above, such as Figure 3 As shown, a cooling decision-making model is constructed based on historical cooling strategies, including:

[0075] A10: Perform data quality enhancement processing on historical operating data to extract effective cooling strategy segments and corresponding thermal state feature vectors;

[0076] A20: Establish a multimodal model architecture that includes policy decision trees and deep learning networks;

[0077] A30: Using thermal state feature vectors as input and optimal cooling strategy as output target, a multimodal model architecture is trained under a reinforcement learning framework to generate an initial cooling decision model;

[0078] A40: The cooling decision model is dynamically updated through an online incremental learning mechanism.

[0079] Specifically, historical operating data undergoes time-series alignment and noise suppression. Effective cooling strategy fragments are extracted using a window sliding segmentation method. Simultaneously, generative adversarial networks (GANs) are used to augment sparse operating condition samples, constructing a matching dataset containing thermal state feature vectors and cooling control commands. A hybrid architecture model is designed, where a policy decision tree captures discrete control logic rules, and a 3D convolutional neural network learns continuous control parameter mappings. Both are fused using an attention mechanism for multimodal feature fusion. A reinforcement learning framework is constructed based on a deep deterministic policy gradient algorithm. The thermal state feature vector is input into the policy evaluation network to generate multi-objective control parameters such as cooling intensity and timing. A value network evaluates the policy's merits and calculates Bellman error backpropagation to update network parameters. An elastic weight solidification algorithm is deployed to achieve online model updates. After new policy data is validated, the Fisher information matrix is ​​calculated to determine the retention strength of important weights. Knowledge transfer is completed while controlling model parameter offsets, ensuring the decision model's continuous adaptability in scenarios such as equipment aging and process changes.

[0080] As a preferred embodiment of the above, such as Figure 4 As shown, the cooling decision model is dynamically updated through an online incremental learning mechanism, including:

[0081] A41: Calculate the matching degree between the newly added strategy fragment and the current working condition. The matching degree is evaluated based on the feature space distance of the state feature vector and the similarity of the tool load.

[0082] A42: The weight allocation of policy fragments is dynamically adjusted based on the matching degree, with policy fragments with a matching degree higher than the threshold taking priority in model parameter updates.

[0083] Specifically, step A41 is first performed to evaluate the value of the newly added strategy data—mapping the thermal state feature vector to a high-dimensional feature space constructed from historical working conditions, using an improved Mahalanobis distance to calculate the matching degree with the current working condition, and simultaneously evaluating the load pattern similarity based on the time-frequency energy distribution characteristics of the cutting force signal using a dynamic time warping algorithm. The two are then fused using a sigmoid function to generate a 0-1 standardized matching degree score, with a dynamic threshold set. Next, step A42, the elastic update stage, processes the data according to the matching degree classification: high-matching-degree samples are directly input into the model and assigned 3 times the training weight, while medium-matching-degree samples... After adding Gaussian noise to the feature space for enhancement, the samples were used in training. 12% of the low-matching samples were retained as decision boundary reinforcement data after anomaly detection. Finally, based on the core mechanism of the EWC algorithm, the Fisher information matrix of historical policy parameters was calculated to identify key weights. A KL divergence constraint term was superimposed on the loss function. When updating the network parameters through gradient descent, the offset of high-importance weights was strictly limited to ensure that the model maintains the stability of the historical core decision logic when absorbing new policies. At the same time, after every 50 incremental updates, the model was dynamically adjusted according to the difference in accuracy between the new and old test sets to form a closed-loop adaptive optimization system.

[0084] As a preferred embodiment of the above, such as Figure 5 As shown, the initial control strategy is simulated and executed in the digital twin environment, including:

[0085] B10: Construct a multi-physics coupled digital twin based on state vectors and a 3D thermodynamic model of the turret, in which the digital twin maps the dynamic process of heat generation and heat dissipation of the turret in real time;

[0086] B20: Load the initial control strategy into the digital twin, and output the predicted temperature distribution and thermal deformation of key components after simulation execution.

[0087] Specifically, based on real-time thermal state feature vectors and the 3D geometric topology of the turret, a multi-physics coupled twin simulation environment is constructed. High-resolution mesh discretization is performed using the finite element method, integrating the electromagnetic heat generation model and the heat dissipation fluid dynamics model. The heat flux density distribution, cooling medium convection efficiency, and structural thermoelastic deformation are simultaneously analyzed through a coupled solver. The initial control strategy is transformed into twin boundary conditions, and transient thermodynamic simulation is performed. The temperature field evolution trend and thermal deformation prediction data of key components are output. Based on the comprehensive assessment model of thermal imbalance risk, the strategy effectiveness index is generated. By comparing with preset safety thresholds, the feasibility of the strategy is automatically determined, providing pre-verification support for the deployment of control strategies for physical equipment.

[0088] As a preferred embodiment of the above, when the predicted temperature suppression rate is lower than a set threshold, a strategy reconstruction mechanism is triggered, including:

[0089] C10: Calculate the predicted temperature suppression rate, using the following formula: Where, η suppress To predict the temperature suppression rate, T actual T represents the current measured temperature. sim To simulate temperature, T target The target temperature threshold;

[0090] C20: When the predicted temperature suppression rate is lower than the first set threshold, the strategy reconstruction mechanism is triggered;

[0091] C30: Re-import the reconstruction strategy into the digital twin environment for iterative verification until the predicted temperature suppression rate reaches the second set threshold for output.

[0092] Specifically, the predicted temperature suppression rate is calculated in real time. Based on the dynamic deviation between the simulated temperature data and the actual monitored temperature data from the digital twin, and combined with the preset target temperature value, the strategy execution effectiveness is evaluated through a suppression rate quantification model. This model is defined as the ratio of the relative deviation between the current actual temperature and the simulated predicted temperature to the target temperature difference. When the suppression rate is detected to be lower than the dynamically set primary warning threshold, the strategy reconstruction engine is activated (the reconstruction trigger condition is dynamically adjusted according to the real-time thermal inertia characteristics of the equipment and historical strategy failure modes). A multi-objective optimization algorithm based on reinforcement learning is launched to search for new strategy combinations that meet thermal balance constraints in the control parameter space. The reconstructed strategy then enters an iterative verification loop. Through rapid simulation and deduction in the digital twin environment, the predicted temperature suppression rate is continuously updated and calculated until it is consistently and stably higher than the secondary acceptance threshold (the secondary threshold is set to increase the preset safety margin compared to the primary threshold). Finally, an optimized strategy that meets the thermal management performance requirements is output and deployed to the physical entity.

[0093] As a preferred embodiment of the above, such as Figure 6 As shown, in step S4, based on processing quality constraints and equipment safety thresholds, the validated initial control strategy is dynamically optimized in real time, and thermal inertia delay compensation is added to generate the final control strategy, including:

[0094] S41: The thermal distortion tolerance limit in the processing quality constraints is combined with the temperature rise rate of key components in the equipment safety threshold to generate a dynamic constraint space;

[0095] S42: Construct a temperature suppression-energy consumption balance model based on thermal state feature vectors, and generate a Pareto optimal policy set based on evolutionary algorithm;

[0096] S43: Perform timing compensation correction for thermal inertia and execution delay on the Pareto optimal policy set;

[0097] S44: Select the control strategy that meets the thermal stability margin requirement from the corrected strategy set based on the real-time operating condition characteristics as the final control strategy.

[0098] Specifically, firstly, a dynamic constraint space is constructed based on key parameters within the thermal variation tolerance limit and equipment safety threshold in the machining quality constraints. An adaptive weighting mechanism enables real-time mapping and aggregation of constraint conditions, with the space boundary dynamically adjusted according to tool wear and load characteristics. Subsequently, a balance model between temperature suppression effectiveness and energy consumption cost is established based on real-time thermal characteristics. A multi-objective optimization algorithm searches for a Pareto optimal strategy set within the constraint space, generating a candidate strategy cluster encompassing cooling intensity gradients and spindle deceleration timing. To address thermal inertia delays and actuator response lags during strategy execution, a time-domain correction model is introduced to perform phase compensation and oscillation suppression on control commands, dynamically adjusting compensation coefficients based on historical delay characteristics. Finally, based on real-time operating conditions, the optimal solution that simultaneously satisfies instantaneous thermal balance requirements and cumulative thermal shock safety margins is selected from the compensated strategy set. After rapid verification in a digital twin environment, the final control command is output, ensuring thermal management accuracy and equipment operational stability under complex operating conditions.

[0099] Example 2

[0100] Based on the same inventive concept as the intelligent cooling-based thermal management optimization method for power turrets in the foregoing embodiments, this invention also provides an intelligent cooling-based thermal management optimization system for power turrets, the system comprising:

[0101] The state vector construction module acquires real-time temperature distribution data and related operating parameters of the power turret, and establishes a state vector characterizing the dynamic relationship between the turret's heat generation and heat dissipation capacity.

[0102] The initial strategy generation module constructs a cooling decision model based on historical cooling strategies, inputs the state vector into the cooling decision model, and generates the initial control strategy.

[0103] The reconfiguration mechanism execution module simulates the execution of the initial control strategy in the digital twin environment. When the predicted temperature suppression rate is lower than the set threshold, the strategy reconfiguration mechanism is triggered.

[0104] The final strategy generation module dynamically optimizes the verified initial control strategy in real time based on processing quality constraints and equipment safety thresholds, and adds thermal inertia delay compensation to generate the final control strategy.

[0105] The optimization system described above in this invention can effectively realize the thermal management optimization method of power turret based on intelligent cooling, and the technical effects it can achieve are as described in the above embodiments, and will not be repeated here.

[0106] As a preferred embodiment of the above, the state vector construction module includes:

[0107] The correlation parameter acquisition unit acquires thermal dynamic correlation parameters in real time.

[0108] The feature analysis unit analyzes the real-time matching relationship between the heat generation feature and the heat dissipation feature of the turret based on the thermal dynamic correlation parameters.

[0109] The feature vector representation unit generates a thermal state feature vector representing the thermal equilibrium state based on the dynamic relationship between the heat generation feature quantity and the heat dissipation feature quantity.

[0110] Similarly, the above-mentioned optimization schemes for the system can also achieve the optimization effects corresponding to the methods in Embodiment 1, which will not be repeated here.

[0111] Although this application has been described in conjunction with specific features and embodiments, it will be apparent that various modifications and combinations can be made therein without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of the application as defined herein, and are to be considered as covering any and all modifications, variations, combinations, or equivalents within the scope of this application.

[0112] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A method for optimizing the thermal management of a power turret based on intelligent cooling, characterized in that, include: Real-time acquisition of temperature distribution data and related operating parameters of the power turret; establishment of a thermal state feature vector characterizing the dynamic relationship between turret heat generation and heat dissipation capacity. A cooling decision model is constructed based on historical cooling strategies. The thermal state feature vector is input into the cooling decision model to generate an initial control strategy. The initial control strategy is simulated and executed in a digital twin environment. When the predicted temperature suppression rate is lower than a set threshold, a strategy reconstruction mechanism is triggered, including: The predicted temperature suppression rate is calculated using the following formula: in, To predict the temperature suppression rate, This is the current measured temperature. To simulate temperature, The target temperature threshold; When the predicted temperature suppression rate is lower than the first set threshold, the strategy reconstruction mechanism is triggered; The reconstruction strategy is re-imported into the digital twin environment for iterative verification until the predicted temperature suppression rate reaches the second set threshold for output. Based on processing quality constraints and equipment safety thresholds, the validated initial control strategy is dynamically optimized in real time, and thermal inertia delay compensation is added to generate a final control strategy, including: The thermal variation deviation limit in the processing quality constraints is combined with the temperature rise rate of key components in the equipment safety threshold to generate a dynamic constraint space. A temperature suppression-energy consumption balance model is constructed based on the thermal state feature vector, and a Pareto optimal strategy set is generated based on an evolutionary algorithm. The Pareto optimal strategy set is subjected to timing compensation correction for thermal inertia and execution delay; Based on the real-time operating conditions, a control strategy that meets the thermal stability margin requirement is selected from the corrected strategy set as the final control strategy.

2. The method for optimizing thermal management of power turrets based on intelligent cooling according to claim 1, characterized in that, The method further includes: when an undefined processing scenario is detected, matching a benchmark strategy based on the similarity evaluation results between the current thermal state feature vector and historical strategies, and updating the cooling decision model through incremental learning.

3. The method for optimizing thermal management of power turrets based on intelligent cooling according to claim 1, characterized in that, Real-time acquisition of temperature distribution data and related operating parameters of the power turret; establishment of a thermal state feature vector characterizing the dynamic relationship between turret heat generation and heat dissipation capacity, including: Real-time acquisition of thermal dynamic correlation parameters; Based on the aforementioned thermal dynamic correlation parameters, the real-time matching relationship between the heat generation characteristics and heat dissipation characteristics of the turret is analyzed. The thermal state feature vector characterizing the thermal equilibrium state is generated based on the dynamic relationship between the heat generation feature and the heat dissipation feature.

4. The method for optimizing thermal management of power turrets based on intelligent cooling according to claim 1, characterized in that, A cooling decision-making model is constructed based on historical cooling strategies, including: Data quality enhancement processing is performed on historical operating data to extract effective cooling strategy fragments and corresponding thermal state feature vectors; Establish a multimodal model architecture that includes policy decision trees and deep learning networks; Using the thermal state feature vector as input and the optimal cooling strategy as output target, the multimodal model architecture is trained under a reinforcement learning framework to generate an initial cooling decision model; The cooling decision model is dynamically updated through an online incremental learning mechanism.

5. The method for optimizing thermal management of power turrets based on intelligent cooling according to claim 1, characterized in that, The cooling decision model is dynamically updated through an online incremental learning mechanism, including: Calculate the matching degree between the newly added strategy fragment and the current working condition. The matching degree is evaluated based on the feature space distance of the thermal state feature vector and the tool load similarity. The weight allocation of the policy fragments is dynamically adjusted based on the matching degree, wherein policy fragments with a matching degree higher than a threshold are given priority in model parameter updates.

6. The method for optimizing thermal management of power turrets based on intelligent cooling according to claim 1, characterized in that, Simulating the execution of the initial control strategy in a digital twin environment includes: A multi-physics coupled digital twin is constructed based on the thermal state feature vector and the three-dimensional thermodynamic model of the turret, wherein the digital twin maps the dynamic process of heat generation and heat dissipation of the turret in real time; The initial control strategy is loaded into the digital twin, and after simulation execution, the predicted temperature distribution and thermal deformation of key components are output.

7. A thermal management optimization system for power turrets based on intelligent cooling, characterized in that, include: The state vector construction module acquires real-time temperature distribution data and related operating parameters of the power turret, and establishes a thermal state feature vector characterizing the dynamic relationship between the turret's heat generation and heat dissipation capacity. The initial strategy generation module constructs a cooling decision model based on historical cooling strategies, inputs the thermal state feature vector into the cooling decision model, and generates an initial control strategy. The reconfiguration mechanism execution module simulates the execution of the initial control strategy in the digital twin environment. When the predicted temperature suppression rate falls below a set threshold, the strategy reconfiguration mechanism is triggered, including: The predicted temperature suppression rate is calculated using the following formula: in, To predict the temperature suppression rate, This is the current measured temperature. To simulate temperature, The target temperature threshold; When the predicted temperature suppression rate is lower than the first set threshold, the strategy reconstruction mechanism is triggered; The reconstruction strategy is re-imported into the digital twin environment for iterative verification until the predicted temperature suppression rate reaches the second set threshold for output. The final strategy generation module dynamically optimizes the validated initial control strategy in real time based on processing quality constraints and equipment safety thresholds, and adds thermal inertia delay compensation to generate the final control strategy, including: The thermal variation deviation limit in the processing quality constraints is combined with the temperature rise rate of key components in the equipment safety threshold to generate a dynamic constraint space. A temperature suppression-energy consumption balance model is constructed based on the thermal state feature vector, and a Pareto optimal strategy set is generated based on an evolutionary algorithm. The Pareto optimal strategy set is subjected to timing compensation correction for thermal inertia and execution delay; Based on the real-time operating conditions, a control strategy that meets the thermal stability margin requirement is selected from the corrected strategy set as the final control strategy.

8. The power turret thermal management optimization system based on intelligent cooling according to claim 7, characterized in that, The state vector construction module includes: The correlation parameter acquisition unit acquires thermal dynamic correlation parameters in real time. The feature analysis unit analyzes the real-time matching relationship between the heat generation feature and the heat dissipation feature of the turret based on the thermal dynamic correlation parameters. The feature vector characterization unit generates the thermal state feature vector characterizing the thermal equilibrium state based on the dynamic relationship between the heat generation feature quantity and the heat dissipation feature quantity.

Images