Chip high-precision temperature control system

Abstract

The invention discloses a chip high-precision temperature control system which comprises a main cooling loop, an auxiliary peak clipping cooling loop and an auxiliary valley filling heating loop, the main cooling loop is communicated with a heat exchange pipeline of a heating chip, and the auxiliary peak clipping cooling loop and the auxiliary valley filling heating loop are connected with the main cooling loop through a first gear pump and a second gear pump respectively. The first gear pump and the second gear pump are in signal connection with the control panel, the main cooling loop is connected with the output end and the backflow end of the main cooling loop liquid storage tank, the input end of the auxiliary peak clipping cooling loop is connected with the peak clipping cooling liquid storage tank, the output end of the auxiliary peak clipping cooling loop is connected with the first gear pump, and the output end of the auxiliary peak clipping cooling loop is connected with the control panel. The input end of the auxiliary valley filling heating loop is connected with the valley filling heating liquid storage tank, the output end of the auxiliary valley filling heating loop is connected with the secondary gear pump, the peak clipping cooling liquid storage tank is connected with the refrigerating system, and the valley filling heating liquid storage tank is connected with the heating system.

Description

Technical Field

[0001] The present invention relates to the technical field of chip production and manufacturing, and in particular to a high-precision chip temperature control system. Background Art

[0002] In the field of electronic products, the normal operation of chips has strict temperature requirements. At present, chip heat dissipation technology is one of the key links to ensure the stable operation of chips, and it has undergone a certain development process. From the early simple heat dissipation methods, it has gradually evolved into the widely used air cooling and liquid cooling technologies today. These technologies can meet the heat dissipation needs of chips under normal working conditions to a certain extent. Their main goal is to control the chip temperature within the specified range.

[0003] At the patent level, there are many patent technologies related to chip heat dissipation. They have innovated the heat dissipation structure, heat dissipation materials, heat dissipation principles, etc. from different angles, and have improved the heat dissipation efficiency and stability to a certain extent. However, these patent technologies mainly focus on optimizing the heat dissipation effect under stable power conditions.

[0004] At present, chip heat dissipation faces key problems. On the one hand, the adjustable range and adjustment speed of the heat dissipation capacity of the current heat dissipation method are relatively low, which makes it impossible to adjust the heat dissipation intensity in a timely and effective manner when the chip power changes. On the other hand, the heat dissipation device has a relatively large thermal capacity. This characteristic makes it difficult to cope with sudden changes in chip power consumption due to the existence of thermal inertia. When the chip power suddenly increases or decreases, the chip temperature will change sharply, which will have an adverse effect on the chip performance. The existing heat dissipation technology lacks an effective response to this. Although there are improvements and explorations, most of them only use fan or pump speed regulation. On the one hand, the adjustment range is limited, and on the other hand, there are obvious deficiencies in temperature control accuracy, which cannot meet the needs of high-precision temperature control of chips. Therefore, a new type of chip temperature control device is urgently needed to solve this problem. Summary of the invention

[0005] The purpose of the present invention is to overcome the defects of the prior art and provide a high-precision chip temperature control system to solve the problems raised in the above background technology.

[0006] A high-precision temperature control system for a chip comprises a control board and a heating chip, wherein the control board is connected to a heating core signal, and the heating chip feeds back a chip power consumption value and a chip temperature value to the control board. The system is characterized in that it also comprises a main cooling circuit, an auxiliary peak shaving cooling circuit, and an auxiliary valley filling heating circuit, wherein the main cooling circuit is connected to a heat exchange pipeline of the heating chip, and two branches, an auxiliary peak shaving cooling circuit and an auxiliary valley filling heating circuit, are arranged between the main cooling circuit and the heat exchange pipeline of the heating chip, wherein the auxiliary peak shaving cooling circuit and the auxiliary valley filling heating circuit are connected to the main cooling circuit through a first gear pump and a second gear pump, respectively, the first gear pump and the second gear pump are connected to the control board signal, respectively, the main cooling circuit is connected to the output end and the reflux end of the main cooling circuit liquid storage tank, the input end of the auxiliary peak shaving cooling circuit is connected to the peak shaving cooling liquid storage tank, the output end of the auxiliary peak shaving cooling circuit is connected to the first gear pump, the input end of the auxiliary valley filling heating circuit is connected to the valley filling heating liquid storage tank, the output end of the auxiliary valley filling heating circuit is connected to the second gear pump, the peak shaving cooling liquid storage tank is connected to a refrigeration system, and the valley filling heating liquid storage tank is connected to a heating system.

[0007] Furthermore, a heat exchanger is provided on the main cooling circuit at the reflux end close to the main cooling circuit liquid storage tank, and a fan is provided on the heat exchanger, and the fan is connected to the control board signal.

[0008] Furthermore, a main circuit pump is provided on the output end of the main cooling circuit close to the main cooling circuit liquid storage tank, and the main circuit pump is connected to the control board signal.

[0009] Furthermore, a solenoid valve is provided in the main cooling circuit near the main circuit pump end.

[0010] Further, the control panel includes a real-time data acquisition and processing module, a historical data analysis and prediction module, a PID controller module, a feedforward compensation module, a comprehensive control decision module, a physical parameter compensation module, a flow comprehensive adjustment module, a flow fine-tuning module, and a final flow adjustment instruction generation module;

[0011] The real-time data acquisition and processing module is connected to the heating chip, and its output is connected to the historical data analysis and prediction module and the PID controller module. The real-time data acquisition and processing module inputs the real-time power consumption P(t) and temperature T(t) data from the heating chip, and outputs the pre-processed power consumption and temperature data stream;

[0012] The historical data analysis and prediction module is connected to the real-time data acquisition and processing module and outputs to the feedforward compensation module. The historical data analysis and prediction module inputs the data received from the real-time data acquisition and processing module and outputs the power consumption prediction value.

[0013] The PID controller module interacts with the real-time data acquisition and processing module to obtain current temperature information and input the target temperature set point T setpoint and current temperature T(t), outputs the flow regulation Q based on error calculation PID and pass the result to the comprehensive control decision module;

[0014] The feedforward compensation module receives the forecast data from the historical data analysis and forecast module and inputs the future power consumption change trend The result is sent to the comprehensive control decision module, and the feedforward adjustment value Q is output. feedforward ;

[0015] The comprehensive control decision module receives the adjustment quantity Q from the PID controller module PID And the feedforward adjustment quantity Q of the information of the feedforward compensation module feedforward , the output is connected to the physical parameter compensation module, and the output comprehensive flow regulation instruction Q final ;

[0016] The physical parameter compensation module receives the instruction of the comprehensive control decision module and inputs the comprehensive flow adjustment instruction Q final , fine-tuned based on actual physical conditions, and then passed to the flow comprehensive regulation module, which outputs the final flow regulation instruction after physical parameter adjustment;

[0017] The flow comprehensive regulation module receives instructions from the physical parameter compensation module, inputs the final flow regulation instruction, directly controls the operation of the two gear pumps, and outputs a driving signal to the first gear pump and the second gear pump;

[0018] The flow fine-tuning module is inserted after the physical parameter compensation module or directly connected before the flow comprehensive adjustment module, and is used to make fine adjustments to the flow when necessary;

[0019] The final flow regulation instruction generation module serves as an information aggregation point, receives status reports from all modules, and issues a final flow regulation instruction.

[0020] Furthermore, the feedforward compensation module outputs a feedforward adjustment value Q feedforward The calculation formula is:

[0021]

[0022] where k f is the feedforward gain coefficient, is the predicted future power consumption value, and P(t) is the current actual power consumption value.

[0023] Further, the PID controller module outputs a flow adjustment value Q based on the error calculation PID The calculation formula is:

[0024]

[0025] Where: e(t) = T setpoint -T(t) is the error signal, K p ,K i ,K d are the proportional, integral and differential coefficients respectively,

[0026] Q PID (t) is the flow adjustment amount calculated by the PID controller, which represents the coolant or heating fluid flow that needs to be adjusted at time t to correct the deviation between the current temperature and the target temperature.

[0027] Furthermore, the comprehensive control decision module outputs a comprehensive flow regulation instruction Q final The calculation formula is:

[0028] Q final (t) = α·Q PID (t)+β·Q feedforward

[0029] Among them, α is the weight factor of PID regulation, and β is the weight factor of feedforward compensation.

[0030] Beneficial effects of the present invention:

[0031] 1. High-precision temperature control

[0032] Through real-time monitoring of the chip temperature sensor and analysis of the chip power output, it is possible to accurately respond to sudden changes in chip power consumption. Whether the power consumption suddenly increases or decreases, the additional cooling and heating liquid circuits can respond quickly, releasing coolant of the corresponding temperature and flow rate to keep the chip temperature constant, effectively avoiding the problem of chip performance degradation caused by rapid temperature fluctuations and achieving high-precision temperature control.

[0033] 2. Improve chip performance stability

[0034] Because the chip temperature is precisely controlled, the chip can always operate at its best performance. It will not cause performance problems such as calculation errors and reduced processing speed due to overheating or overcooling caused by the inadequacy of traditional cooling methods in dealing with sudden changes in power consumption, greatly improving the stability and reliability of chip performance.

[0035] 3. Optimize the adaptability of the cooling system

[0036] Different from traditional heat dissipation technology, the present invention adds a special liquid circuit on the basis of the conventional cooling liquid circuit, so that the heat dissipation system can better adapt to the complex and changeable working conditions of the chip. It can not only effectively dissipate heat when the power consumption is stable, but also can cope with sudden changes in power consumption, thus enhancing the adaptability of the entire heat dissipation system to different working conditions.

[0037] 4. Extend the service life of the chip

[0038] A stable temperature environment can reduce material fatigue and damage to the chip caused by temperature stress. It avoids the accelerated aging of the chip at high temperatures and possible changes in physical properties at low temperatures, thereby extending the service life of the chip and reducing the risk of chip failure due to temperature problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a connection block diagram of the system of the present invention;

[0040] Figure 2 This is a control logic diagram of the control panel of the present invention.

[0041] In the figure: 1. Main cooling circuit storage tank, 2. Main cooling circuit, 3. Main circuit pump, 4. Valley filling heating storage tank, 5. Peak shaving cooling storage tank, 6. Heating system, 7. Refrigeration system, 8. Auxiliary peak shaving cooling circuit, 9. Auxiliary valley filling heating circuit, 10. First gear pump, 11. Second gear pump, 12. Heat generating chip, 13. Control board, 14. Fan, 15. Heat exchanger. DETAILED DESCRIPTION

[0042] The present invention proposes a high-precision temperature control system for chips. The chip's precise temperature control system is based on the premise that the chip supports the output of specific parameters. For example, some Intel CPU chips can output real-time power consumption values ​​and real-time chip temperatures. These two key parameters are input into the temperature control system of the present invention as control signals. This input method based on the chip's own real-time parameters lays a solid foundation for achieving precise temperature control. In the entire temperature control process, the chip's temperature sensor plays an indispensable role. It continuously monitors the chip temperature in real time and feeds back this accurate temperature information to the temperature control device, thereby providing a precise basis for the device's regulation and control actions, and ensuring the timeliness and accuracy of temperature control.

[0043] In terms of the choice of heat dissipation method, the present invention adopts liquid cooling technology and has the function of supporting pump and fan speed regulation. This liquid cooling method provides a strong guarantee for efficient heat dissipation, while the pump and fan speed regulation function further enhances the flexibility and adaptability of the heat dissipation system.

[0044] On this basis, the device has a unique liquid circuit design. In addition to setting up a conventional cooling liquid circuit for basic heat dissipation functions, a heating liquid circuit and a refrigeration liquid circuit with peak-cutting and valley-filling functions are additionally added. These two special liquid circuits are the key innovations of the present invention. They are specifically used to deal with the complex and thorny problem of sudden changes in power consumption of the chip during operation. In order to achieve faster and more effective temperature control, the heating liquid circuit and the refrigeration liquid circuit are arranged at the entrance area of ​​the chip cooling liquid circuit as much as possible. Such a layout can effectively reduce the heat capacity and reduce the delay in the heat transfer process, so that it can respond more quickly to the chip temperature changes and achieve the goal of fast and constant chip temperature. Moreover, the heating liquid circuit and the refrigeration liquid circuit pipelines are driven by a first gear pump 10 and a second gear pump 11 with a flow accurate metering function. This gear pump can accurately control the flow of the coolant, and provides reliable hardware support for the subsequent accurate adjustment of the cooling or heating power according to the chip power consumption changes.

[0045] In a specific embodiment, Figure 1 As shown, a high-precision temperature control system for a chip includes a control board 13 and a heating chip 12. The control board 13 is connected to the heating chip 12 by signal. The heating chip 12 feeds back the chip power consumption value and the chip temperature value to the control board 13. The system also includes a main cooling circuit 2, an auxiliary peak-shaving cooling circuit 8, and an auxiliary valley-filling heating circuit 9. The main cooling circuit 2 is connected to the heat exchange pipeline of the heating chip 12. Two branches, namely, an auxiliary peak-shaving cooling circuit 8 and an auxiliary valley-filling heating circuit 9, are arranged between the main cooling circuit 2 and the heat exchange pipeline of the heating chip 12. The auxiliary peak-shaving cooling circuit 8 and the auxiliary valley-filling heating circuit 9 are connected to the main cooling circuit 2 through a first gear pump 10 and a second gear pump 11, respectively. The first gear pump 10 and the second gear pump 11 are connected to the control board 13 by signal, respectively. The main cooling circuit 2 is connected to the output end and the reflux end of the main cooling circuit liquid storage tank 1. A heat exchanger 15 is arranged on the main cooling circuit 2 near the reflux end of the main cooling circuit liquid storage tank 1. A fan 14 is arranged on the heat exchanger 15, and the fan 14 is connected to the control board 13 by signal. The main cooling circuit 2 is provided with a main circuit pump 3 at the output end near the main cooling circuit liquid storage tank 1, and the main circuit pump 3 is connected to the control board 13 by signal. The main cooling circuit 2 is provided with a solenoid valve near the end of the main circuit pump 3. The input end of the auxiliary peak shaving cooling circuit 8 is connected to the peak shaving cooling liquid storage tank 5, the output end of the auxiliary peak shaving cooling circuit 8 is connected to the first gear pump 10, the input end of the auxiliary valley filling heating circuit 9 is connected to the valley filling heating liquid storage tank 4, the output end of the auxiliary valley filling heating circuit 9 is connected to the second gear pump 11, the peak shaving cooling liquid storage tank 5 is connected to the refrigeration system 7, and the valley filling heating liquid storage tank 4 is connected to the heating system 6.

[0046] The conventional cooling liquid circuit of this chip plays a basic and vital role in the entire heat dissipation system. When the chip is in a constant power consumption working state, the conventional cooling liquid circuit effectively removes the heat generated by the chip by continuously and stably circulating the coolant. By reasonably adjusting the speed of the main circuit pump 3 and the fan 14, the temperature of the chip can be accurately controlled within the appropriate operating temperature range. For example, the chip temperature can be stabilized within the high-precision set value range of 70℃±0.2℃, ensuring the stable performance and reliable operation of the chip under stable power consumption.

[0047] When the power consumption of the chip suddenly increases, the additional peak-shaving refrigeration liquid circuit begins to play its key role. The operating mechanism of the refrigeration liquid circuit is very sophisticated. It can accurately release coolant with a lower temperature through the first gear pump 10 equipped with it according to the specific value of the increase in chip power consumption. In this process, the flow rate of the released coolant is strictly and accurately calculated to ensure that the cooling power generated by this part of the coolant during evaporation or heat exchange on the chip surface is exactly equal to the increase in chip power consumption. Through this precise matching method, the additional heat generated by the sudden increase in power consumption can be effectively offset, thereby ensuring that the chip temperature always remains constant, avoiding the adverse effects of the sudden temperature change caused by the sudden increase in power consumption on the chip performance.

[0048] When the chip power consumption suddenly decreases, the additional valley-filling heating liquid circuit starts to work. It uses the second gear pump 11 to accurately release the coolant with a higher temperature according to the specific situation of the chip power consumption reduction. The released heating liquid flow rate is also carefully calculated to ensure that the heating power generated by these heating liquids during heat exchange with the chip is exactly equal to the reduction value of the chip power consumption. In this way, even if the chip power consumption suddenly decreases, the chip temperature will not be affected and can still maintain stability, thereby further ensuring that the chip performance is not disturbed by temperature fluctuations and ensuring that the chip can operate stably and reliably under complex and changeable power consumption conditions.

[0049] The specific implementation is through the coordination between the modules configured on the control board 13, specifically:

[0050] like Figure 2 As shown, the control panel 13 includes a real-time data acquisition and processing module, a historical data analysis and prediction module, a PID controller module, a feedforward compensation module, a comprehensive control decision module, a physical parameter compensation module, a flow comprehensive adjustment module, a flow fine-tuning module, and a final flow adjustment instruction generation module;

[0051] The real-time data acquisition and processing module is connected to the heating chip 12, and its output is connected to the historical data analysis and prediction module and the PID controller module. The real-time data acquisition and processing module inputs the real-time power consumption P(t) and temperature T(t) data from the heating chip 12, and outputs the pre-processed power consumption and temperature data stream;

[0052] The historical data analysis and prediction module is connected to the real-time data acquisition and processing module and outputs to the feedforward compensation module. The historical data analysis and prediction module inputs the data received from the real-time data acquisition and processing module and outputs the power consumption prediction value.

[0053] The historical data analysis and prediction module is one of the key components in the chip's high-precision temperature control system. It is responsible for predicting future power consumption trends based on the stored historical power consumption and temperature data using time series analysis or machine learning algorithms. This module is designed to predict possible power consumption fluctuations in the chip in advance, thereby providing the necessary information for the feedforward compensation module, enabling the system to respond before power consumption changes occur, and improving the accuracy and timeliness of temperature control.

[0054] The working principle of the historical data analysis and prediction module is:

[0055] The module receives pre-processed power consumption and temperature data from the real-time data acquisition and processing module and stores it in the internal database for subsequent analysis. The received data is further cleaned to remove outliers, fill in missing data points, and convert the data into a form suitable for analysis. Through time domain and frequency domain analysis, statistical feature calculation and other methods, key features that can reflect the system behavior pattern are extracted, such as periodic changes, trend components, seasonal factors, etc. According to the application scenario and technical requirements, select a suitable time series prediction model or machine learning algorithm (such as ARIMA, LSTM neural network, etc.). Use historical data to train the selected model and optimize the model parameters to obtain the best prediction performance. Use the trained model to predict the power consumption changes in the future period and generate power consumption prediction values. The prediction time range can be set according to actual needs, which can be a short time window (such as a few minutes to a few hours) or a longer time window (such as a few days or longer). The predicted power consumption change trend is sent to the feedforward compensation module as the basis for adjusting the flow of coolant or heating liquid. At the same time, a closed-loop control system is established to regularly evaluate the accuracy of the prediction results, and the model is updated or retrained according to the actual operation situation to ensure the continued effectiveness of the prediction effect.

[0056] For example, in a specific embodiment, the system needs to predict the power consumption change trend within the next 10 minutes. The historical data analysis and prediction module performs the following steps:

[0057] 1. Receive data: The historical data analysis and prediction module receives the latest power consumption and temperature data from the real-time data acquisition and processing module.

[0058] 2. Data processing: Perform preprocessing steps such as cleaning, filling missing values, and smoothing noise on the newly received data.

[0059] 3. Feature extraction: Calculate the time series characteristics of the current data, such as mean, variance, autocorrelation coefficient, etc., while taking into account past periodicity and trend information.

[0060] 4. Model application: Apply a pre-trained time series prediction model (such as ARIMA or LSTM) to predict the changing trend of power consumption in the next 10 minutes based on current features and historical data.

[0061] 5. Output prediction results: Output the predicted power consumption trend It is sent to the feedforward compensation module so that it can prepare corresponding flow regulation measures in advance.

[0062] The historical data analysis and prediction module can provide reliable prediction information for the feedforward compensation module, providing strong support for the efficient operation of the entire chip high-precision temperature control system.

[0063] The PID controller module interacts with the real-time data acquisition and processing module to obtain current temperature information and input the target temperature set point T setpoint and current temperature T(t), outputs the flow regulation Q based on error calculation PID The PID controller module is one of the core control components in the chip's high-precision temperature control system. It is responsible for calculating the required flow adjustment amount Q based on the error between the current temperature and the target temperature. PID The module maintains the chip temperature near the set point by precisely adjusting the flow of cooling or heating liquid to achieve stable and efficient temperature control.

[0064] The working principle of the PID controller module is as follows: the PID controller module obtains the real-time measurement value of the current temperature T(t) from the real-time data acquisition and processing module, and reads the target temperature set point T setpoint , calculate the current temperature T(t) and the target temperature T setpoint The error between e(t) = T setpoint -T(t). Apply PID control algorithm to calculate the flow adjustment amount Q that needs to be adjusted PID (t), the formula is as follows:

[0065]

[0066] Where: e(t) = Tsetpoint -T(t) is the error signal, K p ,K i ,K d are the proportional, integral and differential coefficients respectively,

[0067] Q PID (t) is the flow adjustment amount calculated by the PID controller, which indicates the flow of coolant or heating liquid that needs to be adjusted at time t to correct the deviation between the current temperature and the target temperature. The three parameters K of the PID controller are optimized through experimental debugging and automatic parameter adjustment algorithm. p ,K i ,K d , in order to achieve the best control performance. The calculated flow adjustment amount Q PID (t) Sent to the comprehensive control decision module as part of the basis for formulating the final flow regulation instructions.

[0068] For example, in a specific embodiment, the system needs to maintain the chip temperature within the range of 70°C ± 0.2°C, and the PID controller module performs the following steps:

[0069] 1. Receive temperature information: The PID controller module receives the latest temperature data T(t) from the real-time data acquisition and processing module, and reads the target temperature T set by the user. setpoint =70℃.

[0070] 2. Error calculation: Calculate the error between the current temperature and the target temperature e(t) = 70°C - T(t).

[0071] 3. PID control algorithm application: Apply the PID control algorithm to calculate the flow adjustment amount Q that needs to be adjusted based on the error e(t) PID (t).

[0072] 4. Parameter tuning: If the control effect is not ideal, you can adjust the PID parameter K p ,K i ,K d , to obtain better control performance.

[0073] 5. Output adjustment amount: The calculated flow adjustment amount Q PID (t) is sent to the integrated control decision module so that it can combine the information of feedforward compensation to formulate the final flow regulation instruction.

[0074] The feedforward compensation module receives the forecast data from the historical data analysis and forecast module and inputs the future power consumption change trend The result is sent to the comprehensive control decision module, and the feedforward adjustment value Q is output. feedforward ;

[0075] The feedforward compensation module is an important component of the chip high-precision temperature control system. It is designed to calculate and output the required flow adjustment amount Q in advance based on the future power consumption change trend provided by the historical data analysis and prediction module. feedforward The module enables faster and more precise temperature control by pre-adjusting the flow of coolant or heating fluid to offset expected power consumption fluctuations:

[0076] Feedforward compensation module workflow:

[0077] The feedforward compensation module receives the future power consumption change trend from the historical data analysis and prediction module The actual current power consumption P(t) provided by the real-time data acquisition and processing module. These data are the basis for the calculation of the feedforward compensation module. If the forecast shows that the power consumption will increase in the future (for example, due to increased load or running higher performance applications), the feedforward compensation module needs to prepare additional cooling capacity. Conversely, if the forecast shows that the power consumption will decrease, it is necessary to reduce the coolant flow or introduce heating fluid to maintain temperature stability.

[0078] Set the feedforward gain coefficient k f , set the feedforward gain coefficient k f is a critical step that determines the system's response strength to the predicted power consumption changes. f The following aspects need to be considered:

[0079] System response characteristics: Understand the dynamic response characteristics of the system, including the time delay required from adjusting the flow to actually affecting the chip temperature. This can be determined experimentally and is usually known during the design phase.

[0080] Experimental data: The data obtained through experiments can be used to evaluate different k f For example, various power consumption changes are simulated in a laboratory environment, and the corresponding temperature response curves are recorded to optimize k f .

[0081] Stability considerations: Ensure k f The selection should not cause the system to overreact, that is, avoid temperature fluctuations beyond the allowable range due to excessive compensation. At the same time, sufficient response speed should be ensured to offset the temperature impact caused by power consumption changes in a timely manner.

[0082] Safety: Considering the safety and reliability of the system in extreme cases, k f It should be within a reasonable range, neither overly conservative nor overly radical.

[0083] Then according to the set k fAnd the received power consumption change trend, apply the formula to calculate the feedforward adjustment amount Q feedforward :

[0084]

[0085] in:

[0086] where k f is the feedforward gain coefficient, is the predicted future power consumption value, and P(t) is the current actual power consumption value.

[0087] The calculated feedforward adjustment quantity Q feedforward Sent to the comprehensive control decision module as part of the basis for formulating the final flow regulation instructions.

[0088] For example, in a specific embodiment, the system needs to respond in advance to the power consumption change trend within the next 10 minutes, and the feedforward compensation module performs the following steps:

[0089] 1. Receive prediction data: The feedforward compensation module receives the latest power consumption prediction value And the current actual power consumption (P(t).

[0090] 2. Power consumption trend analysis: Analyze the power consumption trend in the next 10 minutes to determine whether the power consumption will increase or decrease, and estimate the impact of these changes on temperature.

[0091] 3. Set the feedforward gain coefficient k f :Based on the system response characteristics and previous experimental data, set an appropriate k f .

[0092] 4. Calculation of feedforward adjustment amount: Apply the formula to calculate the feedforward adjustment amount Q feedforward (t) to prepare corresponding flow regulation measures.

[0093] 5. Output regulation: The calculated feedforward regulation Q feedforward (t) is sent to the comprehensive control decision module so that it can combine the information of the PID controller to formulate the final flow regulation instructions.

[0094] The comprehensive control decision module receives the adjustment quantity Q from the PID controller module PID And the feedforward adjustment quantity Q of the information of the feedforward compensation module feedforward , the output is connected to the physical parameter compensation module, and the output comprehensive flow regulation instruction Q final The integrated control decision module is one of the core components in the chip high-precision temperature control system. It is responsible for integrating the adjustment amount from the PID controller module and the information from the feedforward compensation module (i.e., the feedforward adjustment amount), and calculating the final flow control instruction Q based on this information.final (t). This module combines the advantages of real-time feedback control and predictive control to achieve more accurate and timely adjustment of the flow of cooling or heating liquid, thereby ensuring the stability and efficiency of chip temperature.

[0095] The working principle of the comprehensive control decision module is:

[0096] The integrated control decision module receives the flow adjustment value Q calculated based on the current temperature error from the PID controller module. PID (t), and at the same time receives the feedforward adjustment value Q calculated based on the future power consumption change trend prediction from the feedforward compensation module feedforward (t). According to the response characteristics of the system and experimental data, the weight factors α and β are set. α represents the weight of the PID adjustment amount, and β represents the weight of the feedforward compensation amount. These two factors determine the relative importance of the two adjustment methods in the final flow adjustment instruction.

[0097] Apply the formula to calculate the comprehensive flow regulation index Q final (t), the formula is as follows:

[0098] Q final (t) = α·Q PID (t)+β·Q feedforward

[0099] Among them, α is the weight factor of PID regulation, and β is the weight factor of feedforward compensation.

[0100] The calculated comprehensive flow regulation instruction Q final (t) is sent to the physical parameter compensation module as the basis for the next fine-tuning.

[0101] For example, in a specific embodiment, the system needs to maintain the chip temperature within the range of 70°C ± 0.2°C and to cope with the power consumption change trend within the next 10 minutes. The comprehensive control decision module performs the following steps:

[0102] 1. Receive adjustment information: The comprehensive control decision module receives the latest flow adjustment value Q from the PID controller module PID (t), and the feedforward adjustment value Q provided by the feedforward compensation module feedforward (t).

[0103] 2. Weight factor setting: Based on the system response characteristics and previous experimental data, set the weight factors α and β.

[0104] 3. Calculation of comprehensive flow regulation instruction: Apply the formula to calculate the comprehensive flow regulation instruction Q final (t) to prepare corresponding flow regulation measures.

[0105] 4. Output adjustment amount: The calculated comprehensive flow adjustment instruction Q final (t) is sent to the physical parameter compensation module so that it can formulate the final flow regulation instructions based on the actual situation.

[0106] The integrated control decision module can provide reliable flow regulation information for the physical parameter compensation module, providing strong support for the efficient operation of the entire chip high-precision temperature control system. In addition, it works in conjunction with the PID controller module and the feedforward compensation module to form a fast-response, high-precision temperature control system.

[0107] The physical parameter compensation module receives the instruction of the comprehensive control decision module and inputs the comprehensive flow adjustment instruction Q final , fine-tuned based on actual physical conditions, and then passed to the flow comprehensive regulation module, which outputs the final flow regulation instruction after physical parameter adjustment;

[0108] The physical parameter compensation module receives the instruction of the comprehensive control decision module and inputs the comprehensive flow adjustment instruction Q final , fine-tuned based on actual physical conditions, and then passed to the flow comprehensive regulation module, which outputs the final flow regulation instruction after physical parameter adjustment. The specific details are as follows:

[0109] The physical parameter compensation module first receives the comprehensive flow adjustment instruction Q from the comprehensive control decision module final , parse the flow regulation information in the instruction and prepare to make further adjustments based on actual physical conditions. Actual physical conditions include:

[0110] Pressure drop compensation: The built-in pressure sensor is used to monitor the pressure change during the flow of the coolant or heating liquid in real time, and the speed of the first gear pump 10 and the second gear pump 11 is dynamically adjusted according to the monitored pressure drop to ensure that the required flow rate is maintained unchanged.

[0111] Pumping efficiency optimization: Use the pre-stored database of pumping efficiency curves to select the most appropriate pump speed set point. Based on the pumping efficiency under the current working conditions, the pump speed is optimized to minimize energy loss and maximize flow accuracy.

[0112] Flow path resistance assessment: Based on historical data and experiments, calculate the actual flow deviation caused by factors such as pipe diameter and number of elbows. Based on the calculation results, appropriate compensation is made to the flow to offset the impact of these factors.

[0113] Temperature dependency correction: Consider the effect of temperature on liquid viscosity. As the temperature increases or decreases, the liquid viscosity changes, which affects the flow rate. Based on real-time temperature data, the flow setpoint is automatically adjusted to ensure that the expected flow rate is achieved at different temperatures.

[0114] Compensation for system aging effects: Take into account possible aging phenomena that may occur over time, such as reduced pump performance or increased deposits on the inner wall of pipes. Regularly check and adjust system parameters to ensure long-term stable operation.

[0115] Combined with the monitoring and analysis of all the above physical conditions, the final flow regulation instruction after compensation is calculated. The final flow regulation instruction after physical parameter adjustment is sent to the flow comprehensive regulation module. It is directly connected to the flow comprehensive regulation module through the signal line.

[0116] The flow comprehensive regulation module receives instructions from the physical parameter compensation module, inputs the final flow regulation instruction, directly controls the operation of the two gear pumps, and outputs a driving signal to the first gear pump 10 and the second gear pump 11;

[0117] The flow comprehensive regulation module is one of the key execution components in the chip high-precision temperature control system, responsible for receiving the final flow regulation instruction Q from the physical parameter compensation module. final , and directly controls the operation of the first gear pump 10 and the second gear pump 11 according to the instruction. By accurately adjusting the speed of the two gear pumps, the flow comprehensive regulation module can ensure that the coolant or heating liquid flows through the heat-generating chip 12 at the required flow rate and pressure, thereby achieving efficient and stable control of the chip temperature. Input: Final flow regulation instruction Q final , calculated by the physical parameter compensation module, contains the flow regulation information after compensation for actual physical conditions.

[0118] Output: driving signal to the first gear pump 10 and the second gear pump 11, used to control the rotation speed of the two gear pumps to achieve accurate output of the required flow rate.

[0119] Working principle of comprehensive flow regulation module:

[0120] The flow comprehensive regulation module receives the final flow regulation instruction Q from the physical parameter compensation module final , which has taken into account the impact of actual physical conditions such as pressure drop, pumping efficiency, flow path resistance, temperature dependence and system aging effects.

[0121] Parse the received final flow adjustment instruction Q final (t), determine the target flow rate Q that each gear pump needs to achieve target,1 (t) or Q target,2 (t).

[0122] According to the target flow Q target,1 (t) or Q target,2(t), combined with the pre-stored pumping efficiency curve database, the flow demand is converted into the corresponding gear pump speed setting value N target,1 (t) or (N target,2 (t). This step takes into account the differences in pumping efficiency under different working conditions and optimizes the speed to minimize energy loss and maximize flow accuracy.

[0123] Send a drive signal to the first gear pump 10 or the second gear pump 11 to set their speeds to N respectively. target,1 (t) or N target,2 (t) ensures that the cooling liquid or heating liquid flows to the heat generating chip 12 at a predetermined flow rate and pressure.

[0124] The flow comprehensive regulation module can not only accurately execute the flow regulation instructions provided by the physical parameter compensation module, but also dynamically respond to changes in system status to ensure the efficient operation of the entire chip high-precision temperature control system. In addition, it works closely with the physical parameter compensation module to form a fast-response, high-precision temperature control system.

[0125] The flow fine-tuning module is inserted after the physical parameter compensation module or directly connected before the flow comprehensive adjustment module, and is used to make fine adjustments to the flow when necessary;

[0126] The flow fine-tuning module is one of the fine-tuning components in the chip's high-precision temperature control system. Its main function is to make fine adjustments to the flow of the coolant or heating liquid when necessary. This module is inserted after the physical parameter compensation module or directly connected before the flow comprehensive regulation module to ensure that the flow regulation instructions can be executed most accurately according to actual needs. By introducing a flow fine-tuning mechanism, the temperature control accuracy of the system can be further improved to meet the stringent requirements for chip temperature control.

[0127] Working principle of flow fine-tuning module:

[0128] The flow fine-tuning module first receives the final flow adjustment instruction Q from the physical parameter compensation module. final (t), analyze the target flow rate Q that each gear pump needs to achieve target,1 (t) and Q target,2 (t).

[0129] Implement a closed-loop control system and use the built-in flow sensor to monitor the actual flow rate Q in real time actual,1 (t) and Q actual,2 (t), and the target flow Q target,1 (t) and Q target,2 (t) to identify any possible flow deviation ΔQ(t).

[0130] According to the detected flow deviation ΔQ(t), the proportional integral derivative PID control algorithm or other appropriate adjustment algorithm is applied to calculate the required fine-tuning amount ΔQ adjust (t). This step is intended to minimize flow deviation and ensure that the coolant or heating fluid flows strictly according to the set value.

[0131] The original final flow regulation instruction Q final (t) and the calculated fine-tuning amount ΔQ adjust (t) Combine and generate a new flow regulation instruction Q adjusted (t):

[0132] Q adjusted (t) = Q final (t)+ΔQ adjust (t)

[0133] The updated flow control instruction (Q adjusted (t) is sent to the flow comprehensive regulation module to drive the first gear pump (10) and the second gear pump (11) to perform more precise operations.

[0134] For example, in a specific embodiment, the system needs to keep the chip temperature within the range of 70°C ± 0.2°C and to cope with the power consumption change trend in the next 10 minutes, the flow fine-tuning module performs the following steps:

[0135] 1. Receive and parse instructions: The flow fine-tuning module receives the final flow adjustment instruction Q from the physical parameter compensation module. final (t), parse out the target flow Q target,1 (t) and Q target,2 (t).

[0136] 2. Real-time monitoring of flow deviation: Implement a closed-loop control system to monitor the actual flow Q in real time actual,1 (t) and Q actual,2 (t), and the target flow Q target,1 (t) and Q target,2 (t) and identify the flow deviation ΔQ(t).

[0137] 3. Calculate the fine-tuning amount: Calculate the required fine-tuning amount ΔQ based on the detected flow deviation ΔQ(t) adjust (t).

[0138] 4. Update the flow control instruction: Combine the original final flow control instruction Q final (t) and the calculated fine-tuning amount ΔQ adjust (t), generate a new flow regulation instruction Q adjusted (t).

[0139] 5. Send the updated command: Send the updated flow adjustment command Q adjusted (t) is sent to the flow comprehensive regulation module to drive the first gear pump 10 and the second gear pump 11 to operate more accurately.

[0140] The flow fine-tuning module can not only accurately correct the flow adjustment instructions provided by the physical parameter compensation module, but also dynamically respond to changes in the system state to ensure the efficient operation of the entire chip high-precision temperature control system. In addition, it works closely with the physical parameter compensation module and the flow comprehensive adjustment module to form a fast-response, high-precision temperature control system.

[0141] The final flow control instruction generation module serves as an information aggregation point, receives status reports from all modules, and issues the final flow control instruction. The final flow control instruction generation module is the information aggregation point and decision-making core in the chip high-precision temperature control system. It is responsible for receiving status reports from the real-time data acquisition and processing module, historical data analysis and prediction module, PID controller module, feedforward compensation module, comprehensive control decision module, and physical parameter compensation module, and integrates the final flow control instruction Q based on this information. final The module ensures the coordinated operation of the entire temperature control system and realizes the precise control of the flow of the cooling liquid or the heating liquid to maintain the temperature of the heating chip 12 within the set range.

[0142] Input status reports from various control and monitoring modules, including but not limited to:

[0143] The real-time data acquisition and processing module provides the current power consumption (P(t) and temperature (T(t)).

[0144] The historical data analysis and prediction module provides future power consumption trend

[0145] The error-based flow regulation Q calculated by the PID controller module PID (t).

[0146] The feedforward adjustment value Q output by the feedforward compensation module feedforward (t).

[0147] The comprehensive flow regulation instruction Q calculated by the comprehensive control decision module final,compensated (t).

[0148] Flow regulation instruction Q after adjustment by physical parameter compensation module final,physical (t).

[0149] The final flow regulation instruction is output, which is used to directly control the operation of the first gear pump 10 and the second gear pump 11 after all module information is integrated.

[0150] The working principle of the final flow regulation instruction generation module:

[0151] The final flow regulation instruction generation module receives the latest status report from each relevant module to obtain key information about the current system status and expected changes. The received information is summarized and analyzed to evaluate the consistency and relationship between data from different sources, such as checking whether the real-time temperature meets the predicted value and confirming the coordination between modules. Based on the summarized information, decide how to adjust the flow of coolant or heating liquid. Determine the most appropriate flow regulation strategy taking into account the response characteristics, stability requirements and safety considerations of the system. Convert the above decisions into specific flow regulation instructions to ensure that they reflect not only the changes in real-time temperature and power consumption, but also take into account future power consumption trends and the impact of actual physical conditions. Send the final flow regulation instruction to the comprehensive flow regulation module, which directly controls the operation of the first gear pump 10 and the second gear pump 11.

[0152] The final flow control instruction generation module can not only accurately integrate information from different modules, but also dynamically respond to changes in system status to ensure the efficient operation of the entire chip high-precision temperature control system. In addition, as an information aggregation point, it ensures the coordination and consistency between the subsystems, together forming a fast-response, high-precision temperature control system.

[0153] The present invention can accurately respond to sudden changes in chip power consumption through real-time monitoring of the chip temperature sensor and analysis of the chip power consumption output. Whether the power consumption increases or decreases suddenly, the additional cooling liquid circuit and heating liquid circuit can respond quickly, release the coolant of the corresponding temperature and flow rate, keep the chip temperature constant, effectively avoid the problem of chip performance degradation caused by rapid temperature fluctuations, and achieve high-precision temperature control effect. Due to the precise control of the chip temperature, the chip can always work in the best performance state. Due to the shortcomings of traditional heat dissipation methods in dealing with sudden changes in power consumption, there will be no performance problems such as calculation errors and reduced processing speed caused by overheating or overcooling, which greatly improves the stability and reliability of chip performance. Different from traditional heat dissipation technology, the present invention adds a special liquid circuit on the basis of the conventional cooling liquid circuit, so that the heat dissipation system can better adapt to the complex and changeable working state of the chip. Not only can it effectively dissipate heat when the power consumption is stable, but it can also be at ease in the power consumption mutation scenario, enhancing the adaptability of the entire heat dissipation system to different working conditions. A stable temperature environment can reduce material fatigue and damage caused by temperature stress on the chip. It avoids the accelerated aging of the chip at high temperatures and the possible changes in physical properties at low temperatures, thereby extending the service life of the chip and reducing the risk of chip failure due to temperature problems.

Claims

1. A high-precision chip temperature control system, comprising a control board (13) and a heating chip (12), wherein the control board (13) is connected to the heating core (12) by signal, and the heating chip (12) feeds back the chip power consumption value and the chip temperature value to the control board (13), characterized in that: The invention also comprises a main cooling circuit (2), an auxiliary peak shaving cooling circuit (8), and an auxiliary valley filling heating circuit (9); the main cooling circuit (2) is connected to a heat exchange pipeline of a heating chip (12); two branches, namely an auxiliary peak shaving cooling circuit (8) and an auxiliary valley filling heating circuit (9), are arranged between the main cooling circuit (2) and the heat exchange pipeline of the heating chip (12); the auxiliary peak shaving cooling circuit (8) and the auxiliary valley filling heating circuit (9) are connected to the main cooling circuit (2) via a first gear pump (10) and a second gear pump (11), respectively; the first gear pump (10) and the second gear pump (11) are connected to a control circuit (12) and a control circuit (13); The signal connection of the board (13) is connected, the main cooling circuit (2) is connected to the output end and the return end of the main cooling circuit liquid storage tank (1), the input end of the auxiliary peak shaving cooling circuit (8) is connected to the peak shaving cooling liquid storage tank (5), the output end of the auxiliary peak shaving cooling circuit (8) is connected to the first gear pump (10), the input end of the auxiliary valley filling heating circuit (9) is connected to the valley filling heating liquid storage tank (4), the output end of the auxiliary valley filling heating circuit (9) is connected to the second gear pump (11), the peak shaving cooling liquid storage tank (5) is connected to the refrigeration system (7), and the valley filling heating liquid storage tank (4) is connected to the heating system (6).

2. A high-precision chip temperature control system according to claim 1, characterized in that: A heat exchanger (15) is provided on the main cooling circuit (2) at the return end close to the main cooling circuit liquid storage tank (1), and a fan (14) is provided on the heat exchanger (15). The fan (14) is connected to the control panel (13) by signal.

3. A high-precision chip temperature control system according to claim 1, characterized in that: A main circuit pump (3) is provided at the output end of the main cooling circuit (2) close to the main cooling circuit liquid storage tank (1), and the main circuit pump (3) is connected to the control panel (13) via a signal.

4. A high-precision chip temperature control system according to claim 1, characterized in that: The main cooling circuit (2) is provided with a solenoid valve near the end of the main circuit pump (3).

5. The high-precision chip temperature control system according to claim 1, characterized in that: The control panel (13) comprises a real-time data acquisition and processing module, a historical data analysis and prediction module, a PID controller module, a feedforward compensation module, a comprehensive control decision module, a physical parameter compensation module, a flow comprehensive adjustment module, a flow fine-tuning module, and a final flow adjustment instruction generation module; The real-time data acquisition and processing module is connected to the heating chip (12), and its output is connected to the historical data analysis and prediction module and the PID controller module. The real-time data acquisition and processing module inputs the real-time power consumption P(t) and temperature T(t) data from the heating chip (12), and outputs the pre-processed power consumption and temperature data stream; The historical data analysis and prediction module is connected to the real-time data acquisition and processing module and outputs to the feedforward compensation module. The historical data analysis and prediction module inputs the data received from the real-time data acquisition and processing module and outputs the power consumption prediction value. The PID controller module interacts with the real-time data acquisition and processing module to obtain current temperature information and input the target temperature set point T setpoint and current temperature T(t), outputs the flow regulation Q based on error calculation PID , and pass the results to the comprehensive control decision module; The feedforward compensation module receives the forecast data from the historical data analysis and forecast module and inputs the future power consumption change trend The result is sent to the comprehensive control decision module to output the feedforward adjustment value Q feedforward ; The comprehensive control decision module receives the adjustment quantity Q from the PID controller module PID And the feedforward adjustment quantity Q of the information of the feedforward compensation module feedforward , the output is connected to the physical parameter compensation module, and the output comprehensive flow regulation instruction Q final ; The physical parameter compensation module receives the instruction of the comprehensive control decision module and inputs the comprehensive flow adjustment instruction Q final , fine-tuned based on actual physical conditions, and then passed to the flow comprehensive regulation module, which outputs the final flow regulation instruction after physical parameter adjustment; The flow comprehensive regulation module receives instructions from the physical parameter compensation module, inputs the final flow regulation instruction, directly controls the operation of the two gear pumps, and outputs a driving signal to the first gear pump (10) and the second gear pump (11); The flow fine-tuning module is inserted after the physical parameter compensation module or directly connected before the flow comprehensive adjustment module, and is used to make fine adjustments to the flow when necessary; The final flow regulation instruction generation module serves as an information aggregation point, receives status reports from all modules, and issues a final flow regulation instruction.

6. A high-precision chip temperature control system according to claim 5, characterized in that: The feedforward compensation module outputs the feedforward adjustment value Q feedforward The calculation formula is: where k f is the feedforward gain coefficient, is the predicted future power consumption value, and P(t) is the current actual power consumption value.

7. The high-precision chip temperature control system according to claim 5, characterized in that: The PID controller module outputs the flow regulation quantity Q based on the error calculation PID The calculation formula is: Where: e(t) = T setpoint -T(t) is the error signal, K p ,K i ,K d are the proportional, integral and differential coefficients respectively, Q PID (t) is the flow adjustment amount calculated by the PID controller, which represents the coolant or heating fluid flow that needs to be adjusted at time t to correct the deviation between the current temperature and the target temperature.

8. The high-precision chip temperature control system according to claim 5, characterized in that: The comprehensive control decision module outputs the comprehensive flow regulation instruction Q final The calculation formula is: Q final (t)=α·Q PID (t)+β·Q feedforward Among them, α is the weight factor of PID regulation, and β is the weight factor of feedforward compensation.

Images