Dummy supporting plate cooling system based on vehicle evaluation

Abstract

The invention relates to the technical field of vehicle evaluation, and discloses a dummy supporting plate cooling system based on vehicle evaluation, which comprises a steering control module used for controlling a supporting plate to adjust a steering posture according to a preset route; the speed regulation and control module is used for regulating the movement speed of the supporting plate and cooperatively works with the steering control module; the temperature control module is used for monitoring and dynamically adjusting the temperature of the supporting plate in real time; the GPS positioning module is used for acquiring the position information of the supporting plate and delimiting a test area; the signal receiving module is used for receiving sensor data and an external control instruction; the motor driving module is used for driving the supporting plate to steer and move; and the inertial navigation module is used for monitoring the motion state in real time through an accelerometer and a gyroscope, and regularly calibrating the inertial navigation module and the GPS positioning module when no GPS signal exists. The invention aims to solve the problem that the temperature of the supporting plate is too high due to the heating of the supporting plate and the external high-temperature environment in the testing process, and ensure the accuracy of the testing result and the safe operation of the dummy system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle evaluation, in particular to a dummy pallet cooling system based on vehicle evaluation. BACKGROUND

[0002] With the continuous development of automatic driving technology, the safety of automatic driving vehicles has become one of the focuses. The automatic emergency braking (AEB) system as part of the active safety system aims to prevent or mitigate the occurrence of collision accidents. However, during the AEB test process, the temperature control of the dummy pallet is crucial. The dummy pallet is usually used to simulate the motion trajectory and state of the dummy during the collision process, and it needs to work stably under various environmental conditions. However, the test process is often affected by temperature, especially in high temperature environments, the temperature of the dummy pallet is easy to rise. The excessively high temperature not only has a bad effect on the performance of the pallet material, but also may cause the "overheating" phenomenon of the pallet itself, and then cause the pallet to stop working. The current mainstream pallet has the following problems: The mainstream dummy pallet in the market lacks cooling mechanism configuration, mainly relying on the high temperature resistance of the material. The dummy pallet is mainly made of special composite materials or plastic materials, which have certain high temperature resistance, but their performance will be significantly affected in high temperature environments, which may cause deviation of the test results.

[0003] Especially in the summer outdoor test scene, the insufficient high temperature resistance of the pallet material may also cause overheating problems of the electronic components inside the pallet. The pallet usually needs to be built-in with sensors, wireless communication modules and control systems and other high-precision devices, the normal work of these devices is crucial for the test. If the temperature is too high, these components may have abnormal functions, affecting the collection and processing of data, and finally causing the device to fail. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application proposes a dummy pallet cooling system based on vehicle evaluation, aiming to solve the problem of excessively high temperature of the pallet caused by its own heating and external high temperature environment during the test process, and ensure the accuracy of the test results and the safe operation of the dummy system.

[0005] A dummy pallet cooling system based on vehicle evaluation, comprising: a steering control module for controlling the pallet to adjust the steering posture according to the preset route; a speed control module for adjusting the movement speed of the pallet and working cooperatively with the steering control module; a temperature control module for real-time monitoring and dynamic adjustment of the pallet temperature; a GPS positioning module for obtaining the position information of the pallet and delimiting the test area; A signal receiving module for receiving sensor data and external control instructions; A motor driving module for driving the rotation and movement of the platform; An inertial navigation module for monitoring the motion state in real time through an accelerometer and a gyroscope, and periodically calibrating with the GPS positioning module when there is no GPS signal.

[0006] Further, it also includes: A spiral duct; A condensate circulation unit that starts to absorb heat when the temperature exceeds a dynamic threshold; A base fan for cooling the spiral duct; A deep reinforcement learning model embedded in the temperature control module, with its input connected to: A GPS positioning module and an inertial navigation module for obtaining real-time position and motion state data; A motor driving module for obtaining temperature data under motor operating load; A signal receiving module for obtaining environmental temperature data and historical temperature data; The model predicts the temperature trend of the platform within a set time based on input data, and outputs condensate flow valve opening degree instructions and fan speed instructions.

[0007] Further, the outer wall of the spiral duct is provided with multiple layers of fins, and the extension direction of the fins is parallel to the airflow direction of the base fan.

[0008] Further, it also includes a dummy guide tube, which is provided with a condensing section at the connection with the platform, and a heat exhaust fan at the outlet of the condensing section.

[0009] Further, the working mechanism of the condensate circulation unit is: When the temperature exceeds the dynamic threshold, the condensate absorbs heat and vaporizes into gas, which flows to the condensing section under the pressure difference of the spiral duct; In the condensing section, the gas liquefies and releases heat, which is then discharged to the atmosphere through the dummy guide tube and the heat exhaust fan.

[0010] Further, the dynamic threshold in the temperature control module is dynamically set based on the material properties of the platform, the test accuracy requirements, and the environmental temperature.

[0011] Further, the deep reinforcement learning model is a time series convolutional network-long short-term memory network fusion model, which includes: A time series convolutional network layer that extracts long-time scale temperature dependence features through dilated causal convolution; A long short-term memory network layer for capturing short-term temperature fluctuation features; The linkage control unit outputs continuous control instructions of the condensate flow valve opening degree and the fan target rotating speed according to the prediction result.

[0012] Further, the reward function of the time sequence convolution network-long short term memory network fusion model is: + gamma Wherein, And Is a weight coefficient, Dynamically increases as the temperature approaches the threshold value; Dynamically increases the energy saving weight when the temperature is safe, and gamma is the initial initial factor.

[0013] Further, it further comprises an intelligent backflow control unit; For real-time monitoring of condensate flow and temperature, when the condensate cannot backflow due to high temperature, the backflow speed is automatically adjusted to maintain the heat exchange efficiency.

[0014] The application with the above technical scheme has the following advantages: The application adds a temperature control module to the traditional test support plate, sets an optimized design of the condenser pipe, and dynamically adjusts the size of the high temperature threshold value by embedding a deep reinforcement learning model, thereby solving the problem of high temperature of the dummy support plate in vehicle testing. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the specific embodiments of the present application, the drawings needed in the specific embodiments will be briefly introduced below. In all the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0016] Fig. 1 It is a structure diagram of a dummy support plate cooling system based on vehicle evaluation of the present application; Fig. 2 It is a system structure diagram of a dummy support plate cooling system based on vehicle evaluation of the present application; Fig. 3 It is a flow chart of a dummy support plate cooling system based on vehicle evaluation of the present application.

[0017] Reference signs: Spiral pipe 1, base fan 2, fin 3, dummy guide pipe 4. DETAILED DESCRIPTION

[0018] The embodiments of the technical scheme of the present application will be described in detail below with reference to the drawings. The following embodiments are only used to more clearly illustrate the technical scheme of the present application, and therefore only serve as examples, and cannot limit the protection scope of the present application.

[0019] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in the present application should be the general meanings understood by the skilled in the art to which the present application belongs. The terms "first", "second", etc. in the specification and claims of the embodiments of the present disclosure and the above-described drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances in order to implement the embodiments of the present disclosure described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. Unless otherwise specified, the term "a plurality of" means two or more. In the embodiments of the present disclosure, the character " / " represents a "or" relationship between the preceding and following objects. For example, A / B represents: A or B. The term "and / or" is a description of the association between objects, which means that there can be three relationships. For example, A and / or B, which means: A or B, or, A and B, the three relationships. The term "corresponding" can refer to an association or binding relationship, A corresponding to B means that there is an association or binding relationship between A and B.

[0020] As shown in Figs. 1-3 A dummy board cooling system based on vehicle evaluation, comprising: A steering control module for controlling the adjustment of the steering posture of the board according to the preset route; A speed control module for adjusting the movement speed of the board and working in conjunction with the steering control module; A temperature control module for real-time monitoring and dynamic adjustment of the temperature of the board; A GPS positioning module for obtaining the position information of the board and delineating the test area; A signal receiving module for receiving sensor data and external control instructions; A motor drive module for driving the steering and movement of the board; An inertial navigation module for real-time monitoring of the motion state through an accelerometer and a gyroscope, and periodic calibration with the GPS positioning module when there is no GPS signal.

[0021] In some embodiments, further comprising: A spiral pipe 1; A condensate circulation unit for starting condensate flow heat absorption when the temperature exceeds the dynamic threshold; A base fan 2 for cooling the spiral pipe 1; A deep reinforcement learning model embedded in the temperature control module, the input end of which is connected to: The GPS positioning module and the inertial navigation module for obtaining real-time position and motion state data; The motor drive module for obtaining temperature data under the running load of the motor; The signal receiving module is configured to obtain ambient temperature data and historical temperature data. The model predicts the temperature change trend of the support plate within a set time based on the input data, and outputs the condensate flow valve opening degree instruction and the fan speed instruction.

[0022] Specifically, this embodiment includes two test environments, specifically, summer high-temperature test and long-time frequent test.

[0023] The core module of the system is the temperature control module, which is responsible for real-time monitoring and adjusting the temperature of the support plate to ensure that the support plate does not overheat, thereby affecting the test results.

[0024] In actual work, the working principle of the temperature control module is to monitor the temperature of the support plate through the sensor in the temperature control module, and start the corresponding adjustment mechanism according to the temperature change. The increase of the temperature of the support plate is mainly caused by two aspects of heat: one is the heat of the system itself, and the other is the high temperature of the external environment.

[0025] Especially in summer or high-temperature test environment, the temperature of the external environment will significantly affect the temperature of the support plate. Among them, the temperature control module uses the embedded deep reinforcement learning model to intelligently dynamically adjust the temperature according to the historical data and real-time environmental data. The training process includes multiple data sources.

[0026] For example, the real-time position and inertial navigation data provided by the GPS positioning module, the temperature change record of the device during the historical test, the temperature data under the load of the motor driving module, and the environmental temperature data under the current test condition. These data will be input into the model, and the model will further optimize the temperature adjustment strategy by learning the relationship between the historical temperature change model, the running state of the device and the external environmental factors.

[0027] The deep reinforcement learning module can predict the temperature change trend of the support plate within 1-3 minutes after the device is started based on the currently obtained data. These prediction results will be fed back to the display screen of the temperature control module in real time for the operator to refer to.

[0028] When the temperature prediction value approaches or exceeds the set safety high temperature threshold, the system will automatically start the condensate circulation and low-speed operation of the heat exhaust fan in advance to prevent the rapid rise of temperature and avoid the possible system fluctuations and failures. This early warning mechanism can effectively regulate the internal temperature of the device and avoid the impact of temperature changes on system stability. When the load is high, the temperature control module can increase the cooling intensity in time according to the real-time load data of the motor drive module, and increase the condensate circulation speed or adjust the speed of the heat exhaust fan to cope with the additional heat brought by the load. On the contrary, when the ambient temperature is low, the system will automatically sense the external temperature change and accordingly reduce the cooling intensity to avoid excessive cooling, which not only helps to keep the device within the ideal working range, but also effectively saves energy. The system will intelligently adjust the cooling strategy according to the changes in load and ambient temperature, ensuring the performance of the device while maximizing the energy use efficiency.

[0029] In some embodiments, the outer wall of the spiral pipe 1 is provided with multiple layers of fins 3, and the extension direction of the fins 3 is parallel to the airflow direction of the base fan 2.

[0030] In some embodiments, a dummy guide pipe 4 is further included, and the connection between the dummy guide pipe 4 and the supporting plate is provided with a condensing section, and the outlet of the condensing section is provided with a heat exhaust fan.

[0031] In some embodiments, the working mechanism of the condensate circulation unit is as follows: When the temperature exceeds the dynamic threshold, the condensate absorbs heat and vaporizes into gas, which flows to the condensing section under the pressure difference of the spiral pipe 1; In the condensing section, the gas is liquefied and releases heat, and the heat is discharged to the atmosphere through the dummy guide pipe 4 and the heat exhaust fan.

[0032] In some embodiments, the dynamic threshold in the temperature control module is dynamically set based on the material characteristics of the supporting plate, the test accuracy requirements, and the ambient temperature.

[0033] In some embodiments, an intelligent backflow control unit is further included; which is used to monitor the condensate flow and temperature in real time, and automatically adjusts the backflow speed to maintain the heat exchange efficiency when the condensate cannot backflow due to high temperature.

[0034] Specifically, the operation mode of the temperature control module is that it first controls the flow of condensate by monitoring the temperature in the spiral pipe 1. When the temperature is higher than the set high temperature threshold, the condensate will start to flow in the spiral pipe 1 and absorb heat through the condensation process. During this process, the condensation gas will be driven by the pressure difference in the spiral pipe 1, carrying a large amount of heat to the condensing section, completing the heat exchange process.

[0035] In addition, the base fan 2 at the bottom of the pallet is activated to assist the spiral duct 1 in dissipating heat. When the condensate is heated to the vaporization temperature, the condensed gas will carry heat and enter the condensation section. In the condensation section, the condensed gas will re-liquefy and release heat. To improve condensation efficiency, the condenser is located at the connection between the pallet and the dummy conduit 4 and is equipped with a heat exhaust fan. The heat exhaust fan discharges the heat released during the condensation process through the dummy conduit 4 into the dummy body, and ultimately discharges the heat into the atmosphere, ensuring that the temperature of the pallet is effectively controlled.

[0036] To improve the device's heat exchange efficiency, the device incorporates multiple layers of circular fins 3 around the spiral duct 1 surrounding the base fan 2. These fins increase the surface area of ​​the spiral duct 1, thereby increasing the efficiency of heat transfer. The base fan 2 adjusts its speed via an intelligent control system to maximize heat dissipation and prevent overheating.

[0037] To ensure the temperature control module operates properly, the intelligent reflux control unit automatically adjusts the condensate flow rate when the condensate in spiral pipe 1 is unable to flow back into the system. This monitors the flow rate and temperature of the liquid in real time, ensuring that the condensate flows back at the optimal rate, maximizing heat exchange and preventing overheating or underheating.

[0038] Through the collaborative work of the above-mentioned design and control system, the dummy pallet cooling device for vehicle testing can effectively reduce the pallet temperature during testing, ensuring that the dummy pallet can still operate normally in high-temperature environments, thereby ensuring the accuracy and safety of vehicle testing.

[0039] In some embodiments, the deep reinforcement learning model is a temporal convolutional network-long short-term memory network fusion model, including: Temporal convolutional network layer, which extracts long-time scale temperature-dependent features through dilated causal convolution; Long short-term memory network layer, used to capture short-term temperature fluctuation characteristics; The linkage control unit outputs continuous control instructions for the condensate flow valve opening and the fan target speed based on the prediction results.

[0040] In some embodiments, the reward function of the temporal convolutional network-long short-term memory network fusion model is: +γ in, and is the weight coefficient, It increases dynamically as the temperature approaches the threshold; When the temperature is safe, the energy-saving weight is dynamically increased, and γ is the initial predetermined factor.

[0041] Specifically, the control is performed by an improved long short-term memory network model, a physical knowledge injection time series convolution network-long short-term memory network is proposed, which can further improve the performance of the model on time series data, especially the processing ability for long sequence. Under this design, the time series convolution network is used to capture long-time dependence, and the long short-term memory network is used to refine the short-time dependence of time series, which combines the advantages of the two and makes the model more accurate and better in control effect.

[0042] State space construction unit: convert multi-source heterogeneous data into state vector of deep reinforcement learning module, including real-time temperature data sequence (sampling value of past 60 seconds of pallet / key monitoring point, sampling rate 1 Hz), motion state feature (moving speed, acceleration, heading angle, pitch variance calculated by GPS / inertial navigation data), load feature (heat capacity coefficient, weight, mass center position encoding vector mapped by cargo type), environmental parameters (temperature and humidity, light intensity), equipment state (condensate flow, fan speed, battery power), historical temperature feature (similarity of historical curve of the same line and load with current temperature).

[0043] The time series convolution network-long short-term memory network model needs to input these multi-source data into the network for joint modeling. The specific process is as follows: Input data: including temperature, load, GPS / inertial navigation data. These input data are usually processed into multiple feature channels and are preliminarily time series feature extracted by time series convolution network.

[0044] Convolution layer: first, the input data is convolved by the time series convolution network layer (dilated causal convolution). The time series convolution network can capture the dependence relationship in a longer time scale through the dilated causal convolution structure. Multi-layer convolution can gradually extract high-level time series features For example, assuming that the historical temperature data is a multi-channel input, the time series convolution network layer extracts the historical temperature trend at each time point through different convolution kernels and dilation strategies.

[0045] 2) Time series convolution network-long short-term memory network module architecture The structure of the time series convolution network-long short-term memory network model can be built through the following steps: Time series convolution network layer: process the input time series data through multiple convolution layers, and use dilated convolution to obtain the dependence relationship in a long time window. The advantage of dilated convolution is that it can capture longer dependence relationships without increasing more parameters or layers, improving the time series modeling ability.

[0046] Convolution kernel: Assuming the use of a 3x3 convolution kernel, by setting different dilation factors (dilation factors of 2, 4, 8, etc. are set according to the needs of the system), different time scales can be extracted at each layer.

[0047] LSTM layer: After the time convolution network, a long short-term memory network layer is added to capture short-term dependencies. Long short-term memory network can dynamically adjust the gating mechanism according to the past state to decide whether to save the past memory, which is suitable for tasks where the temperature change has a greater impact on the current time.

[0048] Output layer: After joint processing by the time convolution network and the long short-term memory network, the model outputs the predicted temperature value and the corresponding cooling action strategy (such as condensate flow, fan speed adjustment).

[0049] Action space definition unit: Output continuous control instructions to the actuator, including: condensate flow valve opening adjustment amount (range 0-100%, where 0-30% is low, 30%-80% is medium, and 80%-100% is high), fan target speed adjustment amount (range 0-Max_RPM speed).

[0050] (3) Reward function calculation unit: dynamically quantize multi-objective optimization results, including temperature stability and energy consumption control, specific calculation as follows: 1) Temperature stability: design rewards through "temperature deviation threshold". For example, if the deviation between the system temperature and the target temperature is small (such as less than a certain set threshold), a higher reward is given; if the deviation is large, a negative reward is given.

[0051] Temperature stability reward function design: Where, is the current temperature, is the target temperature. If the deviation is small, the reward is large, and vice versa.

[0052] 2) Energy consumption control: design rewards according to the energy consumption of the device. For example, when the energy consumption is low, give a positive reward; when the energy consumption is high, give a negative reward.

[0053] Energy consumption control reward function design: Where, represents the total power consumption of the pallet device 3) Comprehensive reward function: the final reward can be the weighted sum of the two, considering both temperature stability and energy saving control.

[0054] Comprehensive reward function design: +γ wherein, and are weight coefficients, dynamically increases as the temperature approaches the threshold value; dynamically increases the energy saving weight when the temperature is safe, and γ is an initial factor.

[0055] Prediction result-control decision linkage unit Through the improved long short-term memory network model for control, a physical knowledge injected time sequence convolution network-long short-term memory network is proposed to predict the future 60s / 120s / 180s temperature value and confidence interval.

[0056] When the predicted temperature exceeds the threshold value, the reinforcement learning strategy network is triggered to calculate: 1) Condensate early start time, based on the temperature prediction result, the model will determine whether to adjust the temperature control measure in advance. When the predicted future temperature exceeds the set threshold value, the model can take measures in advance (such as adjusting the condensate flow or fan speed), and this advance action can be determined by calculating the difference between the predicted time point and the current time, so as to determine when to start adjusting the control strategy. Assuming that the model predicts that the temperature will exceed the set threshold value within 2 minutes, the timing of early start can be determined by the following logic : wherein, is the predicted temperature exceeding the threshold value time difference.

[0057] 2) Fan initial speed setting value (satisfying the predicted temperature rise x heat capacity < heat dissipation power x time constraint): once the start time is determined, the model can calculate the initial condensate flow or fan speed and other adjustment parameters according to the predicted temperature value and the expected temperature drop.

[0058] Subsequent closed-loop adjustment parameters (dynamically optimize actions according to real-time temperature feedback): through methods such as policy gradient or Q-learning, update the model parameters so that they can make better decisions in the new environment.

[0059] It should be noted that for each of the foregoing method embodiments, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the application is not limited by the described action sequence, because according to the application, certain steps can be performed in other order or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the actions and modules involved are not necessarily required by the application.

[0060] In the above embodiments, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0061] In several embodiments provided in the present application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units or components shown or discussed can be indirect coupling or communication connection through some interfaces, apparatuses or units, and can be electrical or other forms.

[0062] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0063] In addition, the function units in each embodiment of the application can be integrated in one processing unit, or each unit can be physically present alone, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of software program module.

[0064] The above describes the embodiments of the present application in detail, and the specific examples are applied to the principle and implementation mode of the present application. The above embodiment description is only used to help understand the method and its core idea of the present application; at the same time, for the general technical personnel in the art, according to the idea of the present application, the specific implementation mode and application.

Claims

1. A dummy pallet cooling system based on vehicle evaluation, characterized in that: include: Steering control module, used to control the pallet to adjust the steering posture according to the preset route; The speed control module is used to adjust the movement speed of the pallet and work in conjunction with the steering control module; Temperature control module, used for real-time monitoring and dynamic adjustment of pallet temperature; GPS positioning module, used to obtain the position information of the pallet and demarcate the test area; Signal receiving module, used to receive sensor data and external control instructions; Motor drive module, used to drive the steering and movement of the pallet; The inertial navigation module is used to monitor the motion status in real time through the accelerometer and gyroscope, and to regularly calibrate with the GPS positioning module when there is no GPS signal.

2. A dummy support plate cooling system based on vehicle evaluation according to claim 1, characterized in that: Also includes: spiral pipe; The condensate circulation unit starts the condensate flow to absorb heat when the temperature exceeds the dynamic threshold; The base fan is used to dissipate heat from the spiral pipe; A deep reinforcement learning model is embedded in the temperature control module, with its input connected to: GPS positioning module and inertial navigation module are used to obtain real-time position and motion status data; The motor drive module is used to obtain the temperature data of the motor under load; Signal receiving module, used to obtain ambient temperature data and historical temperature data; The model predicts the temperature change trend of the pallet within a set time based on input data, and outputs a condensate flow valve opening instruction and a fan speed instruction.

3. The dummy support plate cooling system based on vehicle evaluation according to claim 2 is characterized in that: The outer wall of the spiral pipe is provided with multiple layers of fins, and the extending direction of the fins is parallel to the airflow direction of the base fan.

4. The dummy support plate cooling system based on vehicle evaluation according to claim 2, characterized in that: It also includes a dummy duct, a condensation section is provided at the connection between the dummy duct and the supporting plate, and a heat exhaust fan is provided at the outlet of the condensation section.

5. The dummy support plate cooling system based on vehicle evaluation according to claim 4 is characterized in that: The working mechanism of the condensate circulation unit is: When the temperature exceeds the dynamic threshold, the condensate absorbs heat and vaporizes into gas, which flows to the condensation section driven by the pressure difference in the spiral pipe. In the condensation section, the gas liquefies and releases heat, which is then discharged to the atmosphere through the dummy's body via the dummy's ducts and heat exhaust fans.

6. The dummy support plate cooling system based on vehicle evaluation according to claim 1, characterized in that: The dynamic threshold in the temperature control module is dynamically set based on the material characteristics of the support plate, the test accuracy requirements and the ambient temperature.

7. The dummy support plate cooling system based on vehicle evaluation according to claim 2, characterized in that: The deep reinforcement learning model is a temporal convolutional network-long short-term memory network fusion model, including: Temporal convolutional network layer, which extracts long-time scale temperature-dependent features through dilated causal convolution; Long short-term memory network layer, used to capture short-term temperature fluctuation characteristics; The linkage control unit outputs continuous control instructions for the condensate flow valve opening and the fan target speed based on the prediction results.

8. The dummy support plate cooling system based on vehicle evaluation according to claim 7 is characterized in that: The reward function of the temporal convolutional network-long short-term memory network fusion model is: +g in, and is the weight coefficient, It increases dynamically as the temperature approaches the threshold; When the temperature is safe, the energy-saving weight is dynamically increased, and γ is the initial predetermined factor.

9. The dummy support plate cooling system based on vehicle evaluation according to claim 2, characterized in that: Also included is an intelligent reflux control unit; It is used to monitor the condensate flow and temperature in real time. When the condensate cannot reflux due to high temperature, it automatically adjusts the reflux speed to maintain heat exchange efficiency.

Images