An aircraft wheel cooling method and device

Abstract

The present invention discloses an aircraft wheel cooling method and device. The method includes: a temperature and pressure sensor collects the brake disc temperature and wheel hub air pressure in real time, and an embedded processing module calculates the temperature change rate and pressure difference to generate a set of initial thermodynamic parameters; based on the set of initial parameters, the thermal stress distribution value of the brake disc is solved, and a coolant flow rate and injection angle instruction set is generated in combination with a material threshold database; according to the instruction set, the opening degree of the piezoelectric valve of the annular pipeline is regulated, and the injection area is adjusted according to the flow velocity-surface area heat exchange efficiency, and the waste heat is recovered synchronously to optimize the power consumption; an infrared thermal imager obtains the temperature field distribution, compares it with the initial parameter tolerance range, and corrects the injection angle and flow rate instructions through a fuzzy PID algorithm to eliminate the temperature gradient error; based on the temperature drop rate, pressure stability and waste heat recovery efficiency, a neural network scoring model is used to generate an efficiency index, and the algorithm weight and piezoelectric valve response parameters are dynamically adjusted. The present invention effectively improves the cooling efficiency and reduces the risk of thermal stress.

Description

Technical Field

[0001] The present invention relates to the aircraft wheel thermal management technology in the field of aerospace engineering, combining technical fields such as intelligent algorithms in control science and engineering and waste heat recovery technology in energy and power engineering, and particularly relates to an aircraft wheel cooling method. Background Art

[0002] In modern aircraft braking systems, during the landing phase, the braking device undergoes a process of converting huge kinetic energy into heat energy, and its thermal management efficiency directly affects flight safety and component life. Traditional cooling technologies mostly adopt passive air-cooled structures or fixed-flow liquid-cooling schemes, which have three major technical bottlenecks: First, the cooling response lags behind the transient temperature rise, resulting in uneven thermal stress distribution on the brake disc and prone to micro-cracks; Second, the fixed injection mode is difficult to adapt to the thermal load distribution of different braking intensities, causing low utilization efficiency of the coolant; Third, the lack of a braking energy recovery mechanism leads to insufficient overall energy efficiency ratio of the system. Existing technologies generally lack the ability of multi-physical field collaborative control and cannot optimize the balance relationship between cooling parameters and energy recovery in real time, resulting in the brake disc material being in a critical state of thermal fatigue for a long time. Especially for continuous takeoff and landing conditions, the coupling effect of internal air pressure fluctuation and temperature gradient in the wheel hub will accelerate the aging of seals and increase the probability of tire burst. Statistical data from the International Air Transport Association shows that 21% of the unplanned maintenance events of landing gears are due to thermal management failures. Therefore, there is an urgent need to develop intelligent thermal regulation technology to achieve precise and efficient operation of the wheel cooling system through the deep integration of multi-modal sensing, dynamic fluid control, and energy recycling. Summary of the Invention

[0003] An aircraft wheel cooling method and device, comprising:

[0004] S1. A temperature and pressure sensor collects the temperature of the brake disc and the air pressure in the wheel hub in real time, and an embedded processing module calculates the temperature change rate and pressure difference to generate an initial thermodynamic parameter set.

[0005] S2. Based on the initial parameter set, solve the thermal stress distribution value of the brake disc, and combine it with the material threshold database to generate a cooling instruction set for the coolant flow rate and injection angle.

[0006] S3. According to the cooling instruction set, adjust the opening value of the piezoelectric valve of the annular pipeline, adjust the injection area according to the heat exchange efficiency of the flow velocity - surface area, and at the same time recover waste heat through a thermoelectric conversion device and optimize the system power consumption balance value.

[0007] S4. An infrared thermal imager obtains the temperature field distribution, compares it with the initial parameter tolerance range, and corrects the injection angle and flow rate instructions through a fuzzy PID algorithm to eliminate the temperature gradient error.

[0008] S5. Based on the temperature decay rate, pressure stability coefficient, and energy recovery efficiency value during the cooling process, use a neural network scoring model to generate an efficiency index, and dynamically adjust the algorithm weights and hardware response parameters.

[0009] An aircraft wheel cooling method as described above, wherein the semi-temperature and pressure sensor collects the brake disc temperature and hub air pressure in real time, and the embedded processing module calculates the temperature change rate and pressure difference to generate a set of initial thermodynamic parameters, including the following sub-steps:

[0010] Perform filtering and denoising processing on the temperature and air pressure data collected by the temperature and pressure sensor, and extract the effective signal segment;

[0011] Calculate the real-time change rate of the brake disc temperature as the temperature change rate based on the effective signal segment, and calculate the difference between the hub air pressure and the standard air pressure value as the pressure difference;

[0012] Associate and store the temperature change rate, pressure difference, and timestamp, and generate a set of initial thermodynamic parameters including temperature, air pressure, temperature change rate, and pressure difference.

[0013] An aircraft wheel cooling method as described above, wherein based on the set of initial parameters, solve the thermal stress distribution value of the brake disc, and combine with the material threshold database to generate a cooling instruction set for coolant flow rate and injection angle, including the following sub-steps:

[0014] Based on the set of initial thermodynamic parameters, simulate the thermal stress distribution of the brake disc by the finite element analysis method, and output the stress peak value and gradient value of each region;

[0015] According to the temperature resistance threshold and compressive resistance threshold in the material threshold database, match the coolant demand level corresponding to the current stress distribution value;

[0016] Combine the demand level with the heat exchange efficiency model to generate a cooling instruction set including coolant flow rate, injection angle, and duration.

[0017] An aircraft wheel cooling method as described above, wherein according to the cooling instruction set, adjust the opening value of the piezoelectric valve of the annular pipeline, adjust the injection area according to the flow velocity-surface area heat exchange efficiency, and at the same time recover the waste heat through the thermoelectric conversion device and optimize the system power consumption balance value, including the following sub-steps:

[0018] According to the flow rate value in the cooling instruction set, adjust the opening of the corresponding piezoelectric valve in the annular pipeline proportionally to control the coolant flow velocity;

[0019] Based on the flow velocity-surface area heat exchange efficiency model, dynamically adjust the injection area to cover the thermal stress peak area;

[0020] Convert the waste heat into electric energy through the thermoelectric conversion device, and calculate the system power consumption balance value in real time and optimize the energy distribution strategy.

[0021] An aircraft wheel cooling method as described above, wherein an infrared thermal imager obtains the temperature field distribution, compares it with the initial parameter tolerance range, corrects the injection angle and flow rate command through a fuzzy PID algorithm, and eliminates the temperature gradient error, including the following sub-steps:

[0022] Obtain the temperature field distribution data of the brake disc through an infrared thermal imager, and extract the actual temperature gradient curve;

[0023] Compare the actual temperature gradient with the initial parameter tolerance range, calculate the deviation amount and input it into the fuzzy PID controller;

[0024] Adjust the coolant flow rate proportionality coefficient and injection angle compensation value according to the deviation amount until the temperature gradient error is less than the preset threshold.

[0025] An aircraft wheel cooling method as described above, wherein comparing the actual temperature gradient with the initial parameter tolerance range, calculating the deviation amount and inputting it into the fuzzy PID controller includes the following sub-steps:

[0026] When the temperature gradient deviation amount is greater than the tolerance upper limit, give priority to increasing the flow rate proportionality coefficient;

[0027] When the deviation amount is within the tolerance range but there are local over-limit situations, give priority to adjusting the injection angle compensation value;

[0028] When the deviation amount continuously falls below the tolerance lower limit, reduce the flow rate proportionality coefficient and enable the waste heat recovery priority.

[0029] An aircraft wheel cooling method as described above, wherein based on the temperature decay rate, pressure stability coefficient and energy recovery efficiency value during the cooling process, a neural network scoring model is used to generate an effectiveness index, and the algorithm weights and hardware response parameters are dynamically adjusted, including the following sub-steps:

[0030] Real-time collect the temperature decay rate, pressure stability coefficient and energy recovery efficiency value during the cooling process;

[0031] Input the above parameters into the pre-trained neural network scoring model and output the comprehensive effectiveness index;

[0032] Adjust the proportional-integral-derivative weights of the fuzzy PID algorithm according to the effectiveness index, and optimize the response speed of the piezoelectric valve and the thermoelectric conversion efficiency parameters.

[0033] 8. An aircraft wheel cooling device, characterized by comprising:

[0034] Data acquisition and preprocessing module: Temperature and pressure sensors real-time collect the brake disc temperature and wheel hub air pressure, and the embedded processing module calculates the temperature change rate and pressure difference to generate a set of thermodynamic initial parameters.

[0035] Thermal Stress Analysis and Instruction Generation Module: Based on the initial parameter set, calculate the thermal stress distribution values of the brake disc, and combine with the material threshold database to generate a cooling instruction set for the coolant flow rate and injection angle.

[0036] Cooling Execution and Waste Heat Recovery Module: According to the cooling instruction set, adjust the opening value of the piezoelectric valve in the annular pipeline, adjust the injection area based on the flow rate-surface area heat exchange efficiency, and at the same time recover waste heat through the thermoelectric conversion device and optimize the system power consumption balance value.

[0037] Temperature Field Feedback and Dynamic Correction Module: The infrared thermal imager obtains the temperature field distribution, compares it with the initial parameter tolerance range, and corrects the injection angle and flow rate instructions through the fuzzy PID algorithm to eliminate the temperature gradient error.

[0038] Performance Evaluation and Adaptive Optimization Module: Based on the temperature decay rate, pressure stability coefficient, and energy recovery efficiency value during the cooling process, use the neural network scoring model to generate a performance index, and dynamically adjust the algorithm weights and hardware response parameters.

[0039] A computer storage medium, characterized in that it includes: at least one memory and at least one processor;

[0040] The memory is used to store one or more program instructions;

[0041] The processor is used to run one or more program instructions to execute an aircraft wheel cooling method described in any one of the above.

[0042] The beneficial effects achieved by the present invention are as follows:

[0043] The present invention uses temperature and pressure sensors to real-time monitor the temperature of the brake disc and the air pressure of the wheel hub, combines with the embedded processing module to dynamically analyze the thermodynamic parameters, and establishes a high-precision initial data set to overcome the control error caused by data lag in traditional methods; based on the collaborative analysis of thermal stress distribution calculation and material threshold database, generate dynamic instructions for coolant flow rate and injection angle, realize the precise adaptation of cooling intensity and thermal load, optimize energy configuration while ensuring the material safety of the brake disc; use the technology of dynamic matching of piezoelectric valve opening and injection area, combined with the waste heat recovery mechanism, to construct a cooling and energy regeneration collaborative system, significantly improve the heat dissipation efficiency and reduce the overall energy consumption of the system; through closed-loop monitoring by the infrared thermal imager and dynamic correction by the fuzzy PID algorithm, effectively suppress the temperature gradient fluctuation, ensure the temperature uniformity of the brake disc surface, and avoid the deformation risk caused by local overheating; further introduce the neural network model, fuse multi-dimensional parameters to intelligently evaluate and dynamically optimize the cooling performance, enhance the adaptive ability of the system under extreme working conditions, and form an intelligent cooling system with continuous learning ability, comprehensively improving the reliability, stability and energy utilization efficiency of the braking system. Description of the Drawings

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0045] Figure 1 It is a flowchart of a method for cooling an aircraft wheel provided by an embodiment of the present application.

[0046] Figure 2 It is a schematic diagram of a device for cooling an aircraft wheel provided by an embodiment of the present application. Detailed implementation manners

[0047] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present invention.

[0048] Embodiment 1

[0049] As Figure 1 shown, a method for cooling an aircraft wheel in an embodiment of the present application includes:

[0050] Step S1: The temperature and pressure sensor collects the temperature of the brake disc and the air pressure in the wheel hub in real time. The embedded processing module calculates the temperature change rate and the pressure difference, and generates a set of initial thermodynamic parameters, which specifically includes the following sub-steps:

[0051] Step S11: Perform filtering and denoising processing on the temperature and pressure data collected by the temperature and pressure sensor, and extract the effective signal segment;

[0052] The embedded system first performs composite noise reduction processing on the original data of the temperature and pressure sensor. For the random noise existing in the temperature signal, a moving average filter is used for time series smoothing, and high-frequency disturbances are suppressed through a dynamically adjusted time window. For the possible sudden outliers in the pressure signal, a median filter algorithm is enabled to maintain the integrity of the signal waveform and eliminate transient interference terms. To further improve the signal quality, the system performs multi-resolution decomposition on the preprocessed data, separates the noise components in different frequency bands using a specific wavelet basis function, and weakens the residual interference through a threshold processing method. After the noise reduction is completed, the signal enters the effective segment identification stage. Based on the change point detection algorithm, the statistical characteristics of the signal change trend are analyzed, and the transient interference during the sensor startup stage and the invalid data during the system anomaly period are excluded by combining the sliding hypothesis testing method. Finally, the stable signal interval with physical significance is locked.

[0053] Step S12: Calculate the real-time change rate of the brake disc temperature as the temperature change rate based on the valid signal segment, and calculate the difference between the wheel hub air pressure and the standard air pressure value as the pressure difference;

[0054] The calculation of the temperature change rate adopts an improved hybrid algorithm of difference and fitting. First, perform a time-series difference operation based on the temperature values of continuous sampling points. When it is detected that the temperature fluctuation exceeds the preset range, automatically switch to the linear fitting mode:

[0055] Among them, represents the optimal parameter value obtained by minimizing the objective function; represents finding the value that minimizes the subsequent expression; represents the total number of data points participating in the linear fitting; represents the discrete time index within the time window, with a value of , used to index the data points participating in the fitting; represents the time series data; represents the intercept term in the linear fitting model; represents the slope term in the linear fitting model; represents the time interval between adjacent time points.

[0056] Balance the calculation accuracy and real-time requirements by adaptively adjusting the time window range. In the pressure difference calculation link, the system obtains the vehicle load parameters in real time through the in-vehicle bus, dynamically matches the multi-dimensional pressure calibration curve corresponding to the wheel hub model, and at the same time introduces the air pressure compensation coefficient based on the ambient temperature to correct the original air pressure value to a comparable value under standard working conditions.

[0057] All intermediate calculation results need to pass a two-way verification mechanism: on the one hand, set the theoretical boundary value of parameter change based on the thermodynamics characteristics of the material:

[0058] Among them, represents the optimal parameter value obtained by minimizing the objective function; its value range satisfies ; represents the theoretical minimum value; represents the theoretical maximum value; represents the thermal conductivity of the material, reflecting the ability of the material to conduct heat; represents the area related to heat transfer; represents the density of the brake disc material; represents the specific heat capacity; represents the volume of the brake disc; represents the maximum heat.

[0059] On the other hand, a dynamic threshold model is constructed in combination with historical data, and the parallel review of the backup algorithm is initiated for abnormal calculation results to ensure the physical rationality of parameter output and the reliability of the system.

[0060] Step S13: Correlate and store the temperature change rate, pressure difference, and timestamp to generate a set of initial thermodynamic parameters including temperature, air pressure, temperature change rate, and pressure difference.

[0061] The embedded system establishes a unified time reference through a hardware clock source and uses an event-triggered mechanism under a multi-threaded processing architecture to achieve precise alignment of sensor data, calculation parameters, and timestamps. The system constructs a multi-dimensional data structure to integrate raw observations, derived parameters, and environmental context information. The storage layer adopts a mechanism combining hierarchical buffering and persistent storage, maintaining a rolling update of the latest parameter set in the real-time data area, and at the same time transferring structured data batches to non-volatile memory at fixed intervals.

[0062] The data encapsulation process introduces a cyclic redundancy check mechanism to ensure the integrity of information during transmission and establishes a fast retrieval channel through an address mapping table. To support subsequent fault diagnosis and life prediction functions, the system automatically performs data feature extraction and statistic calculation to form an enhanced set of thermodynamic parameters containing information such as trend features and extreme value distributions.

[0063] Step S2: Based on the set of initial parameters, calculate the thermal stress distribution value of the brake disc, and combine it with the material threshold database to generate a cooling instruction set for the coolant flow rate and injection angle, specifically including the following sub-steps:

[0064] Step S21: Based on the set of initial thermodynamic parameters, simulate the thermal stress distribution of the brake disc through the finite element analysis method, and output the stress peak and gradient values of each region.

[0065] Construct a three-dimensional transient heat conduction model of the brake disc based on the set of initial thermodynamic parameters, and its control equation is:

[0066]

[0067] Among them, represents the density of the brake disc material; represents the specific heat; represents the temperature partial derivative with respect to time ; represents the Hamiltonian operator, which represents the divergence operation of a vector in three-dimensional space; represents the reference value of the thermal conductivity; represents the temperature coefficient; represents the temperature gradient; represents the braking friction coefficient, reflecting the friction characteristics between the friction surfaces during braking. represents the braking pressure; represents the contact speed, the relative movement speed between the brake disc and the friction component; represents the contact area.

[0068] Adopt the adaptive mesh generation technology to locally refine the braking contact area, dynamically balance the calculation accuracy and real-time requirements. The setting of boundary conditions integrates the real-time monitoring data of the temperature field and the dynamic parameters of the wheel hub air pressure, and characterizes the thermal conductivity and thermal expansion characteristics of the brake disc with temperature change through a nonlinear material model.

[0069] Introduce an implicit time integration algorithm in the finite element solution process, iteratively solve the distribution of the thermo-mechanical coupling field within each calculation step, and output a spatial distribution matrix containing equivalent stress, principal stress direction, and gradient change rate. For key areas, the system automatically extracts the stress peak value and its spatial change rate, and generates a two-dimensional thermal stress characteristic diagram marked in polar coordinates to provide a quantitative basis for subsequent decisions.

[0070] Step S22: According to the temperature resistance threshold and compressive strength threshold in the material threshold database, match the coolant demand level corresponding to the current stress distribution value;

[0071] The system has a built-in hierarchical material threshold library, and a dynamic performance map containing parameters such as phase transition temperature, yield strength attenuation curve, and fatigue crack growth rate is established for each type of brake disc material. The stress matching process uses the spatial mapping technology to compare the stress distribution field output by the finite element with the three-dimensional failure surface in the material database point by point, and identifies the risk areas exceeding the threshold through tensor operations. The determination of the cooling demand level introduces a fuzzy logic algorithm, comprehensively considering characteristic parameters such as the proportion of stress peak exceeding the limit, the steepness of gradient change, and the area of high-temperature areas, and outputs a decision vector containing multi-level instructions such as emergency cooling, maintaining heat dissipation, and preventive adjustment. At the same time, the system monitors the cumulative damage amount of material performance in real time. When it detects that the brake disc has experienced multiple thermal cycles, it automatically loads the aging correction coefficient to dynamically compensate the threshold to ensure that the judgment logic conforms to the actual service state of the material.

[0072] Step S23: Combine the demand level with the heat exchange efficiency model to generate a cooling instruction set containing coolant flow rate, injection angle, and duration;

[0073] The heat exchange efficiency model pre-sets the flow field distribution characteristics at different injection angles through computational fluid dynamics simulation, and establishes a non-linear relationship matrix of flow rate - coverage area - heat dissipation rate. The cooling strategy is formulated using a constrained optimization algorithm. On the premise of meeting the stress suppression requirements, the minimum coolant consumption and the shortest response time are used as the objective functions to solve the optimal combination of flow rate distribution and injection angle. For the multi-region collaborative cooling scenario, the system implements hierarchical injection control, preferentially implementing high-precision directional cooling on the peak stress region, and at the same time reducing the overall temperature baseline through a wide-area coverage mode.

[0074] The encapsulation of the instruction set follows the vehicle bus protocol specification. Each control cycle outputs a structured instruction packet containing parameters such as the flow rate set value, nozzle deflection angle, and pulse modulation duty cycle, and conducts dynamic feasibility verification through a hardware-in-the-loop simulation module to ensure that the instructions are executable within the working boundaries of the hydraulic actuator. The entire process implements a closed-loop feedback mechanism, and the cooling effect is evaluated in real time through the thermodynamic parameter set of the next cycle, and the response coefficient of the control strategy is dynamically adjusted.

[0075] Step S3: According to the cooling instruction set, adjust the opening value of the piezoelectric valve in the annular pipeline, adjust the injection area according to the flow velocity - surface area heat exchange efficiency, and at the same time recover waste heat through a thermoelectric conversion device and optimize the system power consumption balance value, specifically including the following sub-steps:

[0076] Step S31: According to the flow rate value in the cooling instruction set, adjust the opening of the corresponding piezoelectric valve in the annular pipeline in proportion to control the coolant flow rate;

[0077] Convert the target flow rate value in the cooling instruction into a drive voltage signal, and precisely adjust the bending deformation of the valve plate in the annular pipeline through the inverse piezoelectric effect of the piezoelectric ceramic. The control algorithm adopts a feedforward-feedback composite strategy:

[0078]

[0079] Among them, The control quantity obtained after calculation by the feedforward-feedback composite control algorithm; Represents the function operation rule for calculating the feedforward term value according to the parameters in the parentheses; Represents related to the temperature Related viscosity coefficient; Represents the pipe length; Represents the pipe diameter; Represents the target mass flow rate; Represents the proportional coefficient. In feedback control, it determines the strength of the proportional relationship between the controller output and the error signal; Represents the error value; Represents the integral coefficient. In feedback control, the integral link integrates the error, Determine the strength of the integral action, mainly used to eliminate the steady-state error of the system; Represents the integral of the error, which is the integral operation of the error over time.

[0080] Based on the coolant viscosity-temperature characteristic curve, preset the reference opening value, and at the same time, form a closed-loop correction by real-time detecting the actual flow rate through an ultrasonic flowmeter. For the hysteretic nonlinear characteristics of the piezoelectric valve, the system loads a pre-calibrated hysteresis compensation model and adopts a dynamic parameter estimation method based on the Preisach operator to eliminate the phase deviation in the opening adjustment process. When multiple valves are coordinated for control, a decoupling algorithm is implemented, and a pressure equalizer is used to suppress the hydraulic coupling interference between pipelines, ensuring that the flow rate distribution of each injection unit and the command value error are controlled within the allowable range.

[0081] Step S32: Dynamically adjust the injection area based on the flow velocity-surface area heat exchange efficiency model to cover the peak thermal stress area;

[0082] The heat exchange efficiency model establishes a dimensionless correlation formula with the Reynolds number and Nusselt number as the core through the preset nozzle flow field simulation data, and calculates the convective heat transfer coefficient of the coolant jet on the brake disc surface at different flow velocities in real time. The dynamic area matching algorithm rasterizes the thermal stress distribution map output by the finite element method, uses image morphology methods to identify the geometric center and contour features of the high-stress area, and maps the physical position to the injection unit array of the annular pipeline through coordinate transformation. The control strategy combines the greedy algorithm and predictive control, preferentially activates the nozzle group covering the peak stress area, and at the same time adjusts the opening gradient of adjacent nozzles in advance based on the heat diffusion rate prediction model to form a moving cooling window following the migration of the hot spot. During the execution process, the actual cooling effect is feedback through an infrared thermal imager, and the surface emissivity parameter in the model is dynamically updated to compensate for the heat transfer attenuation caused by brake dust pollution.

[0083] Step S33: Convert the waste heat into electrical energy through a thermoelectric conversion device, calculate the system power consumption balance value in real time, and optimize the energy distribution strategy;

[0084] The thermoelectric conversion device adopts a segmented bismuth telluride-based thermoelectric module layout. According to the radial temperature gradient distribution characteristics of the brake disc, large-sized modules are configured in series in the high-temperature area to increase the output voltage, and small modules are connected in parallel in the low-temperature area to reduce the internal resistance loss. The power management unit implements the maximum power point tracking algorithm, dynamically adjusts the load impedance matching through the perturbation observation method, and at the same time uses a supercapacitor bank to suppress the power fluctuation caused by the sudden change of the braking condition. The system power consumption balance model introduces a dynamic weight factor, and automatically assigns the usage priority of the recovered electric energy according to the state of the vehicle power supply - when the battery has a low power, it preferentially supplies the cooling system pump, otherwise the electric energy is fed back to the vehicle power grid. The optimization algorithm solves the optimal solution of energy distribution in real time through the Lyapunov optimization framework, minimizes the net power consumption of the system while ensuring the cooling performance, and at the same time predicts the waste heat generation trend in the next few seconds through a Kalman filter, and pre-adjusts the working point of the thermoelectric module to improve the overall energy conversion efficiency.

[0085] Step S4: The infrared thermal imager obtains the temperature field distribution, compares it with the initial parameter tolerance range, and corrects the injection angle and flow rate command through the fuzzy PID algorithm to eliminate the temperature gradient error, which specifically includes the following sub-steps:

[0086] Step S41: The infrared thermal imager obtains the brake disc temperature field distribution data and extracts the actual temperature gradient curve;

[0087] The high-resolution infrared thermal imager uses non-uniformity correction technology to eliminate the optical system aberration, and improves the temperature measurement accuracy through real-time calibration with a blackbody radiation source. When collecting data, the brake disc rotation phase signal is obtained synchronously, and the motion compensation algorithm is used to eliminate the image blur caused by the centrifugal effect. The preprocessed temperature field matrix identifies the boundary of the heat dissipation groove through the edge detection algorithm, divides the radial and circumferential detection areas, calculates the temperature gradient modulus value in each direction using the second-order partial differential equation, and generates a gradient intensity cloud map distributed in polar coordinates. For local high-temperature areas, the system expands the detection range through morphological dilation operation, combines the historical temperature distribution characteristics to identify the migration trajectory of abnormal hot spots, and constructs a spatio-temporal feature vector including gradient amplitude, direction and change rate.

[0088] Step S42: Compare the actual temperature gradient with the initial parameter tolerance range, calculate the deviation amount and input it into the fuzzy PID controller;

[0089] The system performs pattern matching between the measured gradient feature vector and the tolerance interval preset in the initial parameter set, and quantifies the deviation degree through the membership function. The rule base design of the fuzzy PID controller adopts a three-dimensional input structure: the gradient deviation amount, the deviation change rate and the historical cumulative error are used as the inference inputs together. For three types of working conditions, the system implements different adjustment strategies:

[0090] (1) When the temperature gradient deviation amount is greater than the tolerance upper limit, the flow rate proportionality coefficient is preferentially increased;

[0091] When the global gradient deviation is detected to break through the upper tolerance limit, a strong intervention mode is triggered, and an exponential weighting mechanism is introduced in the proportional coefficient adjustment link to dynamically increase the flow regulation weight according to the area ratio of the over-limit area.

[0092] (2) When the deviation is within the tolerance range but there is local over-limit, the injection angle compensation value is preferentially adjusted;

[0093] For the working conditions with local over-limit but overall controllable, the system starts the spatial decoupling algorithm. Through the mapping relationship between the thermal imager coordinates and the nozzle array, the optimal injection angle compensation vector is calculated. The geometric relationship between the injection coverage area and the hot spot position is described by the Jacobian matrix, and directional compensation cooling is implemented.

[0094] (3) When the deviation continuously falls below the lower tolerance limit, the flow proportional coefficient is reduced and the waste heat recovery priority is enabled.

[0095] When the system enters the low-deviation steady state, the energy efficiency optimization mode is activated. While reducing the base flow, the working voltage of the thermoelectric conversion device is increased, and more waste heat is converted into storable electric energy through dynamic impedance matching.

[0096] Step S43: Adjust the coolant flow proportional coefficient and the injection angle compensation value according to the deviation until the temperature gradient error is less than the preset threshold;

[0097] The adjustment of the flow proportional coefficient adopts the variable universe fuzzy strategy, and the range of the adjustment amplitude universe is dynamically stretched and shrunk according to the real-time working conditions. A larger adjustment step size is allowed in the stage of severe fluctuation to quickly suppress the deviation, and it switches to the fine adjustment mode when approaching the steady state. The calculation of the injection angle compensation value introduces a prediction-correction mechanism. Based on the jet coverage model preset by computational fluid dynamics, the flow field distribution at different deflection angles is predicted, and then the nozzle pointing angle is corrected by the actual cooling effect feedback of the infrared thermal imager. The jet angle prediction is expressed by the following formula:

[0098]

[0099] Among them, represents the predicted jet angle at time ; represents the current time; represents the integral variable, which is used to cumulatively calculate the relevant quantities in the time period from to during the integral operation; represents the time lag; represents the kernel function, which reflects the influence weight of different times on the prediction result of the current time ; represents the complementary error function; represents the temperature value at time ; represents the reference temperature, which serves as the reference benchmark value for temperature; represents the thermal diffusivity.

[0100] During the closed-loop control process, the system continuously monitors the decay rate of the gradient error and automatically injects a damping factor to smooth the adjustment process when detecting an overshoot risk. After the parameter tuning is completed, the updated control command is sent to the annular pipeline actuator through the timestamp synchronization mechanism, and at the same time, the adjusted system state is written into the operation log of the non-volatile memory to provide an initial parameter optimization benchmark for subsequent control cycles. The entire adjustment process implements multiple protection mechanisms. When the single adjustment amplitude exceeds the safety threshold or the continuous adjustment fails to meet the convergence standard, the fault diagnosis process is started and the backup control strategy is switched.

[0101] Step S5: Based on the temperature decay rate, pressure stability coefficient, and energy recovery efficiency value during the cooling process, use the neural network scoring model to generate an efficiency index and dynamically adjust the algorithm weights and hardware response parameters, which specifically includes the following sub-steps:

[0102] Step S51: Real-time collect the temperature decay rate, pressure stability coefficient, and energy recovery efficiency value during the cooling process;

[0103] The calculation of the temperature decay rate uses the weighted second derivative of the temperature change amount within the sliding window:

[0104]

[0105] where represents the second derivative of the temperature weighted by the sliding window at time ; represents the half-width of the sliding window, that is, the range of the number of data points considered forward and backward from the current time ; represents the summation index variable used to traverse the terms corresponding to different time offsets within the sliding window, and its value range is from to , representing different time intervals relative to the current time ; represents the standard deviation, which is a parameter in the weight function and controls the decay rate of the exponential part in the weight function; sinc(*) is the sinc function, which is a sine basis function; represents a parameter used in the sinc function and affects the shape of the weight function; represents the temperature value at time ; represents the temperature value at time the temperature value, i.e., relative to the current moment , the time offset is the temperature at the moment; represents the temperature value at the moment , i.e., relative to the current moment , the time offset is the temperature at the moment; represents the time interval, i.e., the time difference between two adjacent sampling moments;

[0106] The high-frequency noise interference on trend judgment is eliminated by the Savitzky-Golay filter; the pressure stability coefficient is calculated from the real-time data of the annular pipeline pressure sensor array. The main frequency characteristics of the pressure fluctuation are extracted by the principal component analysis method, and the normal distribution deviation degree of the pressure distribution is determined by combining the Kolmogorov-Smirnov test; the energy recovery efficiency value is obtained by the four-quadrant measurement method. The hot-end temperature difference, output voltage and load current of the thermoelectric conversion device are synchronously collected, and the real-time energy conversion efficiency is calculated through the dynamic impedance matching model. All the original data streams are synchronized at the millisecond level through the timestamp alignment mechanism and filtered by the rationality verification module to filter out transient outliers.

[0107] Step S52: Input the above parameters into the pre-trained neural network scoring model to output the comprehensive efficiency index;

[0108] The pre-trained three-channel convolutional neural network receives the normalized time-series feature matrix. The first channel processes the third-order difference spectrogram of the temperature decay rate, the second channel analyzes the time-frequency distribution characteristics of the pressure stability coefficient, and the third channel analyzes the load characteristic curve of the energy recovery efficiency. The feature fusion layer uses the attention mechanism to dynamically allocate the contribution weights of each parameter, and the fully connected layer outputs a three-dimensional scoring vector including the thermal management efficiency, system stability and energy efficiency. The neural network model is continuously optimized through transfer learning. When deployed on the embedded side, the knowledge distillation technology is used to compress the model scale. At the same time, an online verification module is loaded to monitor the input data distribution offset. When it is detected that the working condition exceeds the coverage range of the training data, it automatically switches to the analogy reasoning mode based on fuzzy similarity to ensure the reliability of the evaluation result.

[0109] Step S53: Adjust the proportional-integral-derivative weights of the fuzzy PID algorithm according to the efficiency index, and optimize the response speed of the piezoelectric valve and the parameters of the thermoelectric conversion efficiency;

[0110] The adjustment of the fuzzy PID weights adopts a reinforcement learning framework. The performance index is used as the input of the reward function, and the membership function shapes of the fuzzy rule tables for the proportional, integral, and derivative terms are updated online through the policy gradient method. The optimization of the piezoelectric valve response speed is achieved by adjusting the rising slope of the driving signal. The system establishes a dynamic response model of the piezoelectric ceramic, calculates the optimal step response parameters based on the current valve body temperature and driving history, and uses a pre-distortion compensation algorithm to offset the influence of the hysteresis effect. To improve the thermoelectric conversion efficiency, dual-mode control is implemented. In the high-performance index interval, an aggressive search mode of the maximum power point tracking algorithm is enabled, and in the low-performance interval, it switches to a conservative tracking strategy based on the historical optimal parameters. All parameter adjustments are verified by the Lyapunov stability criterion, and a parameter version tree is established in the non-volatile memory to support fast rollback and fault tracing when the system performance is abnormal.

[0111] Embodiment 2

[0112] As Figure 2 shown, Embodiment 2 of the present application provides an aircraft wheel cooling device, including:

[0113] Data acquisition and preprocessing module 21: The temperature and pressure sensors collect the brake disc temperature and the hub air pressure in real time. The embedded processing module calculates the temperature change rate and the pressure difference, and generates an initial thermodynamic parameter set, including the following sub-modules:

[0114] Data preprocessing sub-module 211: Filters and denoises the temperature and air pressure data collected by the temperature and pressure sensors, and extracts the effective signal segments;

[0115] The temperature and pressure sensors collect the original signals of the brake disc temperature and the hub air pressure in real time through the embedded hardware interface. After the signals are converted by analog-to-digital conversion, they are transmitted to the preprocessing module. For the noise interference in the signals, first, the sliding window median filtering algorithm is used to eliminate the burst pulse noise, and the window length is adaptively adjusted according to the dynamic characteristics of the signals to balance the real-time performance and the denoising effect; the temperature signal is further processed by low-pass filtering to filter out the high-frequency noise introduced by electromagnetic interference or mechanical vibration, and the air pressure signal is suppressed by weighted moving average filtering to suppress short-term fluctuations and ensure the data smoothness. Subsequently, based on the preset effective range of the physical quantity, the filtered data is screened by thresholds, and the outliers exceeding the reasonable interval are removed; at the same time, combined with the dynamic change rate verification mechanism, if the change rate of the temperature or air pressure of consecutive sampling points exceeds the safety threshold, it is determined as a sensor failure or transient interference, and this section of data is marked as invalid and an alarm is triggered until the signal returns to the stable interval.

[0116] Feature calculation sub-module 212: Calculates the real-time change rate of the brake disc temperature as the temperature change rate according to the effective signal segments, and calculates the difference between the hub air pressure and the standard air pressure value as the pressure difference;

[0117] After the valid signal segment is input, perform a time series difference operation based on the temperature values of consecutive sampling points. When it is detected that the temperature fluctuation exceeds the preset range, automatically switch to the linear fitting mode:

[0118] Among them, represents the optimal parameter value obtained by minimizing the objective function; represents finding the value of that minimizes the subsequent expression; represents the total number of data points participating in the linear fitting; represents the discrete time index within the time window, and the value is and is used to index the data points participating in the fitting; represents the time series data; represents the intercept term in the linear fitting model; represents the slope term in the linear fitting model; represents the time interval between adjacent time points.

[0119] The differential pressure calculation adopts a dynamic reference correction strategy. The standard air pressure value is adjusted in real time according to the ambient temperature through a pre-calibrated temperature-pressure compensation table. The actual differential pressure is the difference between the current air pressure and the dynamic reference, ensuring that the calculation result reflects the air pressure deviation under the real working conditions. If the current air pressure data is marked as invalid due to preprocessing, the air pressure value of the previous valid cycle is used as a temporary reference to maintain the calculation continuity and synchronously trigger the sensor health diagnosis. All the calculation results of the features need to pass the logical verification. If data conflicts are detected, the output is frozen and the historical valid values are called, and at the same time, the system monitoring module is reported for fault tracing.

[0120] Data storage and management sub-module 213: Associatively store the temperature change rate, differential pressure, and time stamp, and generate a set of thermodynamic initial parameters including temperature, air pressure, temperature change rate, and differential pressure;

[0121] The feature calculation results are bound to high-precision timestamps and encapsulated into a structured data unit containing temperature, air pressure, temperature change rate, and pressure difference, and real-time storage management is implemented through a circular buffer. The buffer uses a circular queue structure to overwrite historical data within a fixed time window, automatically discarding the oldest data during writing to prevent memory overflow. At the same time, the data associated with abnormal events is marked with priorities to ensure that key information is retained until the persistent storage stage. The storage module periodically batches and compresses the buffer data and transfers it to non-volatile memory. The compression algorithm needs to balance efficiency and resource occupancy to adapt to the computing power of the embedded system; a checksum is appended before data writing to ensure integrity. If the check fails during reading, an attempt is made to recover from the buffer or backup area, and an error log is recorded for subsequent analysis. When the storage capacity is lower than the safety threshold, the sampling frequency is adaptively reduced or an alarm is triggered to prioritize the storage reliability of core data, forming a closed-loop logic from real-time processing to long-term storage management.

[0122] Thermal stress analysis and instruction generation module 22: Based on the initial parameter set, solve the thermal stress distribution values of the brake disc, and combine with the material threshold database to generate a cooling instruction set for coolant flow rate and injection angle, including the following sub-modules:

[0123] Thermodynamic modeling and simulation sub-module 221: Based on the initial thermodynamic parameter set, simulate the thermal stress distribution of the brake disc through the finite element analysis method, and output the stress peak and gradient values of each region;

[0124] Based on the initial thermodynamic parameter set, construct a three-dimensional transient heat conduction model of the brake disc, and its control equation is:

[0125]

[0126] Among them, represents the density of the brake disc material; represents the specific heat; represents the temperature the partial derivative with respect to time ; represents the Hamiltonian operator, which represents the divergence operation of a vector in three-dimensional space; represents the reference value of the thermal conductivity; represents the temperature coefficient; represents the temperature gradient; represents the braking friction coefficient, reflecting the friction characteristics between the friction surfaces during braking; represents the braking pressure; represents the contact speed, the relative movement speed between the brake disc and the friction component; represents the contact area.

[0127] First, non-uniform grids are divided according to the geometric structure and material properties of the brake disc. The grid density is increased in the potential areas of high-temperature gradients to improve the calculation accuracy. The boundary conditions are dynamically set based on the temperature field and air pressure data collected in real time. Among them, the heat flux density is derived from the temperature change rate and the friction power consumption model, and the convective heat dissipation coefficient is corrected by the aerodynamic formula related to the wheel hub air pressure and pressure difference. The solver uses an implicit iterative algorithm to calculate the transient heat conduction equation, and outputs the stress peak and gradient distribution diagrams of each area of the brake disc at fixed time steps. At the same time, the position and change trend of the maximum stress point are recorded. To balance the calculation efficiency and real-time requirements, the model runs in a simplified version in the embedded system, and the matrix operation is optimized by the hardware acceleration unit to ensure that the simulation results are output to the downstream module within a millisecond-level delay.

[0128] Decision rule matching sub-module 222: According to the temperature resistance threshold and compressive strength threshold in the material threshold database, match the coolant demand level corresponding to the current stress distribution value;

[0129] The thermo-stress distribution values output by the simulation are matched with the material threshold database in real time. The database pre-stores parameters such as the temperature resistance threshold, compressive strength threshold, and fatigue life curve of the brake disc material. The matching logic is divided into multiple levels of judgment: First, locate the risk area according to the stress peak. If the peak exceeds the instantaneous tensile strength threshold of the material, the highest-level cooling demand is directly triggered; if the peak is within the safe range but the gradient value exceeds the fatigue damage critical value, the cooling level is dynamically evaluated according to the gradient amplitude and duration. The cooling demand levels are divided into four levels: emergency, high, medium, and low. Each level corresponds to a preset coolant flow rate reference value and angle adjustment range. An environment adaptation strategy is introduced during the matching process; if there is no current material matching item in the database, an interpolation estimation mode based on similar material properties is enabled, and an artificial verification request is triggered to ensure the reliability of the decision.

[0130] Control strategy generation sub-module 223: Combine the demand level and the heat exchange efficiency model to generate a cooling instruction set including coolant flow rate, injection angle, and duration;

[0131] Based on the cooling demand level and the heat exchange efficiency model, an executable cooling instruction set is generated. The heat exchange efficiency model integrates the physical properties of the coolant, pipeline characteristics, and injection dynamics, and calculates the optimal flow rate and angle combination in real time through a non-linear equation set. The control instruction set includes the flow rate setting value, injection angle, and duration, and limits the frequent switching of the instructions through an anti-saturation algorithm to avoid overloading the actuator. After the instruction is generated, a redundant verification mechanism is embedded: if the current actuator state conflicts with the instruction, switch to the backup strategy, and at the same time report to the system monitoring module for fault tracing to ensure the safety and executability of the control instruction.

[0132] Cooling Execution and Waste Heat Recovery Module 23: According to the cooling instruction set, adjust the opening value of the piezoelectric valves in the annular pipeline, adjust the spraying area based on the flow velocity - surface area heat exchange efficiency, and at the same time recover waste heat through the thermoelectric conversion device and optimize the system power consumption balance value, including the following sub - modules:

[0133] Actuator Control Sub - module 231: According to the flow rate value in the cooling instruction set, adjust the opening of the corresponding piezoelectric valve in the annular pipeline proportionally to control the coolant flow velocity;

[0134] The flow rate setting value in the cooling instruction set is converted into a piezoelectric valve opening adjustment signal through the embedded control unit. The piezoelectric valve adopts a closed - loop control strategy. Based on the deviation between the real - time feedback of the flow sensor and the instruction value, the feed - forward - feedback composite strategy is adopted through the control algorithm:

[0135]

[0136] Among them, The control quantity obtained after calculation by the feed - forward - feedback composite control algorithm; Represents the function operation rule for calculating the feed - forward term value according to the parameters in the brackets; Represents related to the temperature The viscosity coefficient; Represents the pipe length; Represents the pipe diameter; Represents the target mass flow rate; Represents the proportional coefficient. In feedback control, it determines the strength of the proportional relationship between the controller output and the error signal; Represents the error value; Represents the integral coefficient. In feedback control, the integral link performs integral operation on the error, Determines the strength of the integral action, mainly used to eliminate the steady - state error of the system; Represents the integral of the error, which is the integral operation of the error over time.

[0137] Make the actual coolant flow velocity accurately match the target value. The relationship between the piezoelectric valve opening and the flow rate is compensated by a pre - calibrated non - linear curve to ensure high - resolution adjustment in the low - flow section and fast response in the high - flow section. The valve control logic integrates a fault protection mechanism: If the flow sensor fails or the valve feedback is abnormal, switch to the open - loop mode, perform conservative control according to the preset voltage - flow mapping table, and at the same time trigger an alarm and record the fault code. When multiple valves work together, a time - division multiplexing strategy is adopted to avoid instantaneous current overload, and hardware interlock is used to ensure the timing safety of the valve actions.

[0138] Dynamic Optimization Sub - module 232: Based on the flow velocity - surface area heat exchange efficiency model, dynamically adjust the spraying area to cover the peak area of thermal stress;

[0139] Based on the flow velocity - surface area heat exchange efficiency model, the optimal injection area is calculated in real - time. The model inputs include the peak position of thermal stress, the current flow velocity of the coolant, and the nozzle layout parameters, and the outputs are the activation priorities and angle adjustment amounts of each nozzle. The optimization objective is to maximize the cooling efficiency of the thermal stress peak area under limited flow rate. The greedy algorithm is used to dynamically allocate the flow rate: the nozzles covering the high - stress area are preferentially activated, and the injection angle is finely adjusted according to the stress gradient direction. The nozzle angle is adjusted by a stepper motor or a piezoelectric actuator, and the angle resolution needs to match the spatial accuracy of the thermal stress distribution. The optimization process embeds an inertia constraint to limit the mechanical wear caused by frequent switching of the nozzle angle, and predicts the adjustment trend of the injection area through learning from historical data to reduce the real - time calculation load.

[0140] Energy management sub - module 233: Convert waste heat into electrical energy through a thermoelectric conversion device, calculate the power consumption balance value of the system in real - time, and optimize the energy distribution strategy;

[0141] The thermoelectric conversion device is arranged closely to the high - temperature area of the brake disc and uses the Seebeck effect to convert waste heat into electrical energy. The conversion efficiency is calibrated in real - time through the temperature difference, and the output voltage is stabilized by a converter and stored in the supercapacitor bank as the auxiliary power supply of the system. The power consumption balance value is calculated by monitoring the ratio of the main power supply to the recovered electrical energy in real - time, and the power consumption strategy of the cooling system is dynamically adjusted: during high waste heat recovery periods, the coolant flow rate is increased to the upper limit to accelerate heat dissipation, and at the same time, the redundant electrical energy is fed back to the main power supply; during low waste heat periods, it switches to the low - power mode to limit the energy consumption of non - critical loads. The energy distribution strategy is implemented through a fuzzy logic controller, comprehensively considering the remaining battery power, braking intensity, and thermal stress risk level, and giving priority to ensuring the stability of core functions. If the total system power consumption exceeds the safety threshold, a degradation protocol is triggered, and secondary functions are gradually shut down and an energy warning is reported.

[0142] Temperature field feedback and dynamic correction module 24: The infrared thermal imager obtains the temperature field distribution, compares it with the initial parameter tolerance range, and corrects the injection angle and flow rate commands through the fuzzy PID algorithm to eliminate the temperature gradient error, including the following sub - modules:

[0143] Temperature field monitoring and feature extraction sub - module 241: Obtain the brake disc temperature field distribution data through the infrared thermal imager and extract the actual temperature gradient curve;

[0144] The infrared thermal imager captures the temperature field distribution data on the surface of the brake disc in real time through high-frame-rate scanning. After the original data is corrected for non-uniformity and compensated for bad pixels, it is converted into a temperature matrix. The feature extraction algorithm is based on matrix gradient calculation, generating two-dimensional temperature gradient curves along the radial and circumferential directions of the brake disc, and locating the high-temperature aggregation areas and gradient mutation points. To improve the anti-interference ability, the data fusion module synchronously accesses the discrete temperature values of the temperature and pressure sensors, and calibrates the local measurement error of the infrared thermal imager through the weighted average method. The extracted temperature gradient curve is bound to the time stamp and output to the downstream control module, while an anomaly detection mechanism is embedded: if it is detected that the temperature field distribution seriously deviates from the physical heat transfer model, it is determined that the sensor fails or there is optical contamination, triggering the self-cleaning process or switching to the backup sensor.

[0145] Intelligent control strategy sub-module 242: Compare the actual temperature gradient with the tolerance range of the initial parameters, calculate the deviation amount and input it into the fuzzy PID controller;

[0146] The actual temperature gradient curve is compared with the tolerance range preset by the initial parameter set of thermodynamics, and the deviation amount calculates the global error and the area of the local over-limit region through the space integration method. The fuzzy PID controller dynamically adjusts the control parameters according to the amplitude, duration and spatial distribution characteristics of the deviation amount:

[0147] (1). When the temperature gradient deviation amount is greater than the upper tolerance limit, give priority to increasing the flow rate proportionality coefficient;

[0148] When the overall deviation of the temperature gradient exceeds the upper tolerance limit, the fuzzy rule base gives priority to increasing the flow rate proportionality coefficient, quickly suppressing the spread of the temperature gradient by increasing the coolant flow rate, and at the same time limiting the maximum flow rate increase to prevent thermal shock.

[0149] (2). When the deviation amount is within the tolerance range but there are local over-limit situations, give priority to adjusting the injection angle compensation value;

[0150] If the deviation is within the global tolerance range but there are local hot spots, based on the coordinates of the high-temperature area and the injection coverage model, dynamically calculate the nozzle angle compensation value, give priority to adjusting the injection angle of the corresponding area to enhance directional cooling, and at the same time maintain the flow stability of other areas.

[0151] (3). When the deviation amount continuously falls below the lower tolerance limit, reduce the flow rate proportionality coefficient and enable the waste heat recovery priority.

[0152] When the deviation continuously falls below the lower tolerance limit, gradually reduce the flow rate proportionality coefficient to reduce energy consumption, and increase the priority of the waste heat recovery subsystem, distributing the redundant coolant flow to the high-efficiency working range of the thermoelectric conversion device. The control strategy incorporates hysteresis anti-chattering logic to avoid command oscillations caused by instantaneous noise.

[0153] Dynamic parameter adjustment sub-module 243: Adjust the coolant flow rate proportionality coefficient and the injection angle compensation value according to the deviation amount until the temperature gradient error is less than the preset threshold;

[0154] According to the control parameter correction amount output by the fuzzy PID, the coolant flow rate proportionality coefficient and the injection angle compensation value are adjusted in real time. The flow rate adjustment adopts a progressive approximation strategy: based on the current flow rate, it is superimposed step by step according to the proportionality coefficient increment, and a fixed period is delayed after each adjustment to wait for the temperature field response until the global gradient error converges to the preset threshold. The injection angle compensation value is converted into the stepping motor drive pulse number of each nozzle through a coordinate mapping table. The angle correction amount is strictly matched with the spatial position of the high-temperature area. Based on the jet coverage model preset by computational fluid dynamics, the flow field distribution under different deflection angles is predicted, and then the nozzle pointing angle is corrected through the actual cooling effect feedback of the infrared thermal imager. The jet angle prediction is expressed by the following formula:

[0155]

[0156] Among them, represents the predicted jet angle at time ; represents the current moment; represents the integral variable, which is used to cumulatively calculate relevant quantities during the integral operation from to time period; represents the time lag; represents the kernel function, which reflects the influence weight of different moments on the prediction result at the current moment ; represents the complementary error function; represents the temperature value at time ; represents the reference temperature, which is used as the reference benchmark value of the temperature; represents the thermal diffusion coefficient.

[0157] The adjustment process embeds a saturation protection mechanism: the single-time flow rate adjustment amplitude does not exceed the system safety limit, and the angle compensation range is limited by the mechanical structure limit. If the error is still not eliminated after the maximum number of adjustments, it is determined as an uncontrollable working condition, triggering system-level degradation operation and reporting a fault code, and synchronously recording the temperature field data for offline analysis.

[0158] Performance evaluation and adaptive optimization module 25: Based on the temperature decay rate, pressure stability coefficient, and energy recovery efficiency value during the cooling process, use a neural network scoring model to generate a performance index, and dynamically adjust the algorithm weight and hardware response parameters, including the following sub-modules:

[0159] Data Monitoring and Acquisition Sub-module 251: Real-time collect the temperature decay rate, pressure stability coefficient, and energy recovery efficiency value during the cooling process;

[0160] The system real-time collects key parameters during the cooling process through multi-source sensors: The temperature decay rate is calculated from the sequential data of high-precision temperature sensors. The calculation of the temperature decay rate uses the weighted second derivative of the temperature change amount within a sliding window:

[0161]

[0162] where, represents the second derivative of the temperature weighted by the sliding window at time ; represents the half-width of the sliding window, that is, the range of the number of data points considered forward and backward from the current time ; represents the summation index variable, which is used to traverse the terms corresponding to different time offsets within the sliding window, and its value range is from to , representing different time intervals relative to the current time ; represents the standard deviation, which is a parameter in the weight function and controls the decay rate of the exponential part in the weight function; sinc(*) is the sinc function, which is a sine basis function; represents a parameter used in the sinc function and affects the shape of the weight function; represents the temperature value at time ; represents the temperature value at time , that is, relative to the current time , the temperature at the time when the time offset is ; represents the temperature value at time , that is, relative to the current time , the temperature at the time when the time offset is ; represents the time interval, which represents the time difference between two adjacent sampling times;

[0163] The pressure stability coefficient is quantified by the ratio of the standard deviation to the mean of the hub air pressure, and the pressure fluctuation intensity is dynamically evaluated by combining the range analysis within the time window; The energy recovery efficiency value is calculated in real-time by the ratio of the output electric energy of the thermoelectric conversion device to the input power of the brake disc waste heat. The input power is calculated by the product of the temperature difference and the heat flow rate. After all parameters are aligned with time stamps by the hardware synchronization unit, they are encapsulated into a structured data set, and outlier filtering and data normalization processing are performed before transmission to ensure that the data input to the downstream module meets the model input specifications.

[0164] Performance evaluation and model calculation sub-module 252: Input the above parameters into the pre-trained neural network scoring model to output the comprehensive performance index;

[0165] The pre-trained lightweight neural network model is deployed in the embedded inference engine. The input layer receives the normalized temperature decay rate, pressure stability coefficient, and energy recovery efficiency values. The hidden layer extracts non-linear features through the activation function, and the output layer generates a comprehensive performance index in the range of 0-1, characterizing the global performance level of the current cooling system. During the model training stage, historical operating condition data and expert scoring labels are used for supervised learning, and an online incremental learning mechanism is introduced: regularly fine-tune the model weights with new data at the edge side to adapt to long-term drift problems such as material aging and environmental changes. The inference process embeds dynamic confidence detection. If the input parameter combination exceeds the training data distribution range, the expert rule base is triggered to assist in scoring to avoid optimization failure caused by model misjudgment.

[0166] Parameter optimization and adaptive adjustment sub-module 253: Adjust the proportional-integral-differential weights of the fuzzy PID algorithm according to the performance index, and optimize the response speed of the piezoelectric valve and the parameters of the thermoelectric conversion efficiency;

[0167] Based on the high or low performance index, the gradient descent method is used to online adjust the weight distribution of the proportional, integral, and differential terms of the fuzzy PID controller: when the performance is low, the proportional coefficient is preferentially increased to accelerate the system response; when the performance is high, the integral term is strengthened to suppress the steady-state error. At the same time, the weight overshoot is restricted by the constraint conditions to avoid control oscillation caused by parameter mutation. The optimization of the piezoelectric valve response speed is achieved by dynamically reconstructing the voltage-opening non-linear mapping relationship of the drive circuit. Switch to the low-gain mode in the high-performance interval to reduce power consumption, and enable the high-gain mode in the low-performance period to improve the flow regulation accuracy. At the same time, a dead zone compensation algorithm is embedded to eliminate the influence of mechanical hysteresis. The optimization of the thermoelectric conversion efficiency is achieved by dynamically adjusting the load impedance matching point of the TEG array through pulse width modulation, making it always approach the maximum power transfer state, and coordinating the coolant flow control to maintain the temperature difference in the high-efficiency working range of the thermoelectric material. The parameter adjustment process adopts a rolling time domain optimization strategy. The optimization effect is re-evaluated based on the latest performance index at fixed intervals. If the performance does not improve after continuous adjustments for multiple times, it will automatically roll back to the historical stable parameter set and trigger system-level fault diagnosis. All parameter adjustment instructions are sent through the security interface, strictly limited within the hardware physical limit, and the key operation records the log with time stamps, supporting offline traceability and policy iteration verification.

[0168] Corresponding to the above embodiments, an embodiment of the present invention provides a computer storage medium, including: at least one memory and at least one processor;

[0169] The memory is used to store one or more program instructions;

[0170] A processor for running one or more program instructions to execute an aircraft wheel cooling method;

[0171] Corresponding to the above embodiments, an embodiment of the present invention provides a computer-readable storage medium. The computer storage medium contains one or more program instructions for being executed by a processor to perform an aircraft wheel cooling method.

[0172] An embodiment of the present invention discloses a computer-readable storage medium. Computer program instructions are stored in the computer-readable storage medium. When the computer program instructions run on a computer, the computer is caused to execute the above-mentioned aircraft wheel cooling method.

[0173] In an embodiment of the present invention, the processor may be an integrated circuit chip with signal processing capabilities. The processor may be a general-purpose processor, a digital signal processor (DSP for short), an application specific integrated circuit (ASIC for short), a field programmable gate array (FPGA for short), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0174] It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor. The steps of the method disclosed in combination with the embodiments of the present invention may be directly embodied as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module may be located in a random access memory, flash memory, read-only memory, programmable read-only memory, or electrically erasable programmable memory, register, a mature storage medium in the art. The processor reads the information in the storage medium and combines its hardware to complete the steps of the above method.

[0175] The storage medium may be a memory, for example, it may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories.

[0176] Among them, the non-volatile memory may be a read-only memory (ROM for short), a programmable read-only memory (PROM for short), an erasable programmable read-only memory (EPROM for short), an electrically erasable programmable read-only memory (EEPROM for short), or a flash memory.

[0177] The volatile memory may be a Random Access Memory (RAM) which serves as an external cache. By way of example but not limitation, many forms of RAM are available, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DRRAM).

[0178] The storage media described in the embodiments of the present invention are intended to include, but not be limited to, these and any other suitable types of memory.

[0179] Those skilled in the art should be aware that, in one or more of the above examples, the functions described in the present invention can be implemented by a combination of hardware and software. When applying software, the corresponding functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium. The computer-readable medium includes computer storage media and communication media, where the communication media includes any medium that facilitates the transfer of a computer program from one place to another. The storage media can be any available medium accessible by a general-purpose or special-purpose computer.

[0180] The specific embodiments described above further elaborate on the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, and improvements made on the basis of the technical solutions of the present invention shall be included within the protection scope of the present invention.

Claims

1. An aircraft wheel cooling method, characterized in that, Including: S1. The temperature and pressure sensor collects the temperature of the brake disc and the wheel hub air pressure in real time, and the embedded processing module calculates the temperature change rate and the pressure difference to generate a set of initial thermodynamic parameters. S2. Based on the set of initial parameters, solve the thermal stress distribution value of the brake disc, and combine with the material threshold database to generate a cooling instruction set for the coolant flow rate and injection angle. S3. According to the cooling instruction set, adjust the opening value of the piezoelectric valve in the annular pipeline, adjust the injection area according to the flow rate-surface area heat exchange efficiency, and at the same time recover the waste heat through the thermoelectric conversion device and optimize the system power consumption balance value. S4. The infrared thermal imager obtains the temperature field distribution, compares it with the initial parameter tolerance range, and corrects the injection angle and flow rate instructions through the fuzzy PID algorithm to eliminate the temperature gradient error. Specifically, it includes: obtaining the brake disc temperature field distribution data through the infrared thermal imager, and extracting the actual temperature gradient curve; comparing the actual temperature gradient with the initial parameter tolerance range, calculating the deviation amount and inputting it into the fuzzy PID controller; adjusting the coolant flow rate proportionality coefficient and the injection angle compensation value according to the deviation amount until the temperature gradient error is less than the preset threshold. Among them, comparing the actual temperature gradient with the initial parameter tolerance range, calculating the deviation amount and inputting it into the fuzzy PID controller includes the following sub-steps: when the temperature gradient deviation amount is greater than the upper tolerance limit, preferentially increase the flow rate proportionality coefficient; when the deviation amount is within the tolerance range but there are local over-limit conditions, preferentially adjust the injection angle compensation value; when the deviation amount continuously falls below the lower tolerance limit, reduce the flow rate proportionality coefficient and enable the waste heat recovery priority. S5. Based on the temperature decay rate, pressure stability coefficient and energy recovery efficiency value during the cooling process, use the neural network scoring model to generate an efficiency index, and dynamically adjust the algorithm weight and hardware response parameters.

2. The aircraft wheel cooling method according to claim 1, wherein The temperature and pressure sensor collects the temperature of the brake disc and the wheel hub air pressure in real time, and the embedded processing module calculates the temperature change rate and the pressure difference to generate a set of initial thermodynamic parameters, including the following sub-steps: Perform filtering and denoising processing on the temperature and air pressure data collected by the temperature and pressure sensor, and extract the effective signal segment. Calculate the real-time change rate of the brake disc temperature as the temperature change rate according to the effective signal segment, and calculate the difference between the wheel hub air pressure and the standard air pressure value as the pressure difference. Associate and store the temperature change rate, pressure difference and time stamp to generate a set of initial thermodynamic parameters including temperature, air pressure, temperature change rate and pressure difference.

3. A method for cooling an aircraft wheel according to claim 1, characterized in that, Based on the set of initial parameters, solve the thermal stress distribution value of the brake disc, and combine with the material threshold database to generate a cooling instruction set for the coolant flow rate and injection angle, including the following sub-steps: Based on the set of initial thermodynamic parameters, simulate the thermal stress distribution of the brake disc by the finite element analysis method, and output the stress peak value and gradient value of each region. According to the temperature resistance threshold and compressive strength threshold in the material threshold database, match the coolant demand level corresponding to the current stress distribution value. Combine the demand level with the heat exchange efficiency model to generate a cooling instruction set including coolant flow rate, injection angle and duration.

4. A method for cooling an aircraft wheel according to claim 1, characterized in that, According to the cooling instruction set, adjust the opening value of the piezoelectric valve in the annular pipeline, adjust the injection area according to the flow rate-surface area heat exchange efficiency, and at the same time recover the waste heat through the thermoelectric conversion device and optimize the system power consumption balance value, including the following sub-steps: Adjust the opening degree of the corresponding piezoelectric valve in the annular pipeline proportionally according to the flow value in the cooling instruction set to control the coolant flow rate; Based on the flow velocity-surface area heat exchange efficiency model, dynamically adjust the spraying area to cover the peak area of thermal stress; Convert the waste heat into electric energy through a thermoelectric conversion device, calculate the system power consumption balance value in real time, and optimize the energy distribution strategy.

5. A method for cooling an aircraft wheel according to claim 1, characterized in that, Based on the temperature decay rate, pressure stability coefficient, and energy recovery efficiency value during the cooling process, use a neural network scoring model to generate an efficiency index, and dynamically adjust the algorithm weights and hardware response parameters, including the following sub-steps: Collect the temperature decay rate, pressure stability coefficient, and energy recovery efficiency value during the cooling process in real time; Input the above parameters into the pre-trained neural network scoring model to output the comprehensive efficiency index; Adjust the proportional-integral-derivative weights of the fuzzy PID algorithm according to the efficiency index, and optimize the response speed of the piezoelectric valve and the thermoelectric conversion efficiency parameters.

6. An aircraft wheel cooling device, characterized in that, Including: Data acquisition and preprocessing module: The temperature and pressure sensors collect the brake disc temperature and hub air pressure in real time, and the embedded processing module calculates the temperature change rate and pressure difference to generate the initial thermodynamic parameter set; Thermal stress analysis and instruction generation module: Based on the initial parameter set, solve the thermal stress distribution value of the brake disc, and combine with the material threshold database to generate the cooling instruction set of the coolant flow rate and spraying angle; Cooling execution and waste heat recovery module: According to the cooling instruction set, adjust the opening degree value of the piezoelectric valve in the annular pipeline, adjust the spraying area according to the flow velocity-surface area heat exchange efficiency, and at the same time recover the waste heat through the thermoelectric conversion device and optimize the system power consumption balance value; Temperature field feedback and dynamic correction module: The infrared thermal imager obtains the temperature field distribution, compares it with the initial parameter tolerance range, and corrects the spraying angle and flow rate instruction through the fuzzy PID algorithm to eliminate the temperature gradient error; specifically including: obtaining the brake disc temperature field distribution data through the infrared thermal imager, and extracting the actual temperature gradient curve; comparing the actual temperature gradient with the initial parameter tolerance range, calculating the deviation amount and inputting it into the fuzzy PID controller; adjusting the coolant flow rate proportionality coefficient and spraying angle compensation value according to the deviation amount until the temperature gradient error is less than the preset threshold; Among them, comparing the actual temperature gradient with the initial parameter tolerance range, calculating the deviation amount and inputting it into the fuzzy PID controller includes: when the temperature gradient deviation amount is greater than the tolerance upper limit, preferentially increase the flow rate proportionality coefficient; when the deviation amount is within the tolerance range but there are local overlimits, preferentially adjust the spraying angle compensation value; when the deviation amount continues to be lower than the tolerance lower limit, reduce the flow rate proportionality coefficient and enable the waste heat recovery priority; Efficiency evaluation and adaptive optimization module: Based on the temperature decay rate, pressure stability coefficient, and energy recovery efficiency value during the cooling process, use a neural network scoring model to generate an efficiency index, and dynamically adjust the algorithm weights and hardware response parameters.

7. A computer storage medium, characterized in that, Including: At least one memory and at least one processor; The memory is used to store one or more program instructions; The processor is used to run one or more program instructions to execute an aircraft wheel cooling method according to any one of claims 1-5.

Images