Thermodynamic system regulation and control system and method based on real-time entropy production feedback and electronic equipment

Abstract

The invention provides a thermodynamic system regulation and control system and method based on real-time entropy production feedback and electronic equipment. Comprising a multi-source sensing and entropy production calculation terminal, a dynamic regulation and control quantity generation module and an execution module, and the real-time entropy production rate of each target subsystem is determined through multi-dimensional physical field data, collected in real time, of each target subsystem in the thermodynamic system; according to the preset target entropy yield of each target subsystem and the real-time entropy yield of each target subsystem, the entropy production change rate and the accumulated entropy production deviation of each target subsystem are determined through a differential link and an integral link in an adaptive gain PID control method; and obtaining the real-time regulation and control quantity of each target subsystem, dynamically adjusting the actuator parameters of each target subsystem according to the real-time regulation and control quantity of each target subsystem, and driving the real-time entropy yield of each target subsystem to return to the target entropy yield of each target subsystem.

Description

Technical Field

[0001] This application relates to the field of automatic control technology for thermal systems, and in particular to a thermal system control system, method, and electronic device based on real-time entropy production feedback. Background Technology

[0002] A thermal system is an energy conversion and transfer system formed by connecting thermal equipment such as boilers, steam turbines, and heat exchangers through auxiliary equipment such as pipes and valves. Its core function is to convert thermal energy into mechanical energy or electrical energy through fuel combustion or external heat source input, and to maintain the stability of the thermodynamic parameters of the working fluid (such as water vapor, air, etc.) during the energy transfer process to ensure the efficient and stable operation of the system.

[0003] To reduce energy loss and ensure stable operation of the system under varying conditions, existing technologies typically employ optimized control strategies (such as model predictive control, fuzzy PID algorithms, and multi-objective optimization algorithms). By introducing real-time data-driven dynamic adjustment mechanisms and constructing multi-parameter collaborative optimization frameworks, the efficiency of thermal systems can be improved and irreversible losses reduced.

[0004] However, traditional PID control relies primarily on temperature and pressure feedback signals, resulting in significant response lag and difficulty in effectively suppressing transient entropy generation. Especially under scenarios of sudden load changes, system overshoot increases significantly, impacting system stability and energy efficiency. The fixed threshold method performs poorly in handling varying operating conditions, leading to… The sharp increase in losses triggers abrupt changes in local entropy production, further reducing system efficiency. Furthermore, fouling deteriorates heat transfer performance, increasing system energy consumption. The mismatch between rigid control mode and dynamic thermodynamic processes causes additional energy consumption at varying load rates. These factors, including losses, collectively limit the potential for improving the energy efficiency of thermal systems. Summary of the Invention

[0005] This application provides a thermodynamic system control system, method, and electronic device based on real-time entropy production feedback, in order to solve the problem that the existing thermodynamic system control methods have limited room for improving system energy efficiency.

[0006] In a first aspect, embodiments of this application provide a thermodynamic system control system based on real-time entropy production feedback, comprising:

[0007] A multi-source sensing and entropy production calculation terminal is used to determine the real-time entropy production rate of each target subsystem based on the multi-dimensional physical field data of each target subsystem in the real-time collected thermal system.

[0008] The dynamic control quantity generation module is used to determine the entropy production change rate and cumulative entropy production deviation of each target subsystem based on the preset target entropy production rate and the real-time entropy production rate of each target subsystem, through the derivative and integral links in the adaptive gain PID control method, and to obtain the real-time control quantity of each target subsystem based on the entropy production change rate, entropy production change rate gain, cumulative entropy production deviation, and cumulative entropy production deviation gain of each target subsystem.

[0009] The execution module is used to dynamically adjust the actuator parameters of each target subsystem according to the real-time control amount of each target subsystem, and drive the real-time entropy production rate of each target subsystem to return to the target entropy production rate of each target subsystem.

[0010] Furthermore, the dynamic control quantity generation module includes at least an adaptive gain unit, which is used for:

[0011] The real-time load rate of each target subsystem is determined based on the maximum rated power and real-time output power of each target subsystem.

[0012] If the rate of change of entropy production of each target subsystem is greater than the entropy production rate threshold, determine the load level of the real-time load rate of each target subsystem.

[0013] Based on the gain generation strategy corresponding to the load level, generate multiple gain combinations of entropy production change rate and cumulative entropy production deviation;

[0014] Calculate the reward value for each gain combination based on the preset reward function;

[0015] The gain combination corresponding to the maximum reward value is determined as the entropy production change rate gain and cumulative entropy production deviation gain of each target subsystem;

[0016] The real-time control quantity of each target subsystem is obtained based on the entropy production change rate and entropy production change rate gain of each target subsystem, as well as the cumulative entropy production deviation and cumulative entropy production deviation gain of each target subsystem.

[0017] Furthermore, based on the weighting generation strategy corresponding to the load level, multiple weighted combinations of entropy production change rate and cumulative entropy production deviation are generated, including:

[0018] When the load level is low, multiple first weight combinations are generated based on the principle that the cumulative entropy production deviation weight is greater than the entropy production change rate weight.

[0019] When the load level is high, multiple second weight combinations are generated based on the principle that the cumulative entropy production deviation weight is less than the entropy production change rate weight.

[0020] Furthermore, the adaptive gain unit is also used for:

[0021] The real-time load change rate of each target subsystem is determined based on the real-time load rate of each target subsystem.

[0022] If the real-time load change rate of the first target subsystem exceeds the preset threshold load change rate, the reward value of the first target subsystem is generated by the preset reward function based on the real-time entropy production rate, entropy production change rate, and real-time load rate of the first target subsystem.

[0023] Based on the range of the reward value of the first target subsystem, determine the correction direction and adjustment range of the entropy production change rate gain and the cumulative entropy production deviation gain of the first target subsystem;

[0024] Based on the correction direction and the adjustment magnitude, adjust the entropy production change rate gain and cumulative entropy production deviation gain of the first target subsystem to maximize the reward value of the first target subsystem.

[0025] The entropy production rate gain and cumulative entropy production deviation gain corresponding to the maximum reward value of the first target subsystem are determined as the entropy production rate gain and cumulative entropy production deviation gain of the first target subsystem.

[0026] Furthermore, the dynamic control quantity generation module further includes at least a target entropy yield determination unit, which is used for:

[0027] Based on the multi-dimensional physical field data of each target subsystem over a preset historical time period, determine the [data / data] of each target subsystem at each historical moment. Efficiency and entropy production rate;

[0028] The efficiency exceeding the preset value in each of the target subsystems Efficiency is determined to be maximum Efficiency set, based on the maximum Each in the efficiency set The entropy yield corresponding to efficiency generates a candidate set of entropy yields.

[0029] Based on the premise of satisfying the device safety threshold and / or system stability boundary of each target subsystem, the minimum entropy production rate in the candidate set of entropy production rates is determined as the target entropy production rate of each target subsystem.

[0030] Furthermore, the execution module includes at least a hierarchical adjustment unit, which is used for:

[0031] Based on the pre-established correspondence between each target subsystem and the real-time control quantity, the actual control quantity of each target subsystem is determined;

[0032] Based on the actual control amount of each target subsystem and the weight parameters of each actuator in each target subsystem, the actual control amount of each actuator in each target subsystem is determined.

[0033] Adjust the parameters of each actuator in each target subsystem according to the actual control amount of each actuator in each target subsystem.

[0034] Furthermore, the execution module further includes at least a differential channel unit, the differential channel unit being used for:

[0035] In the event of a sudden crosswind causing eddies in the target subsystem, the system is triggered to detect fluid pressure fluctuations, velocity changes, and eddy morphology data of the target subsystem within a preset time.

[0036] Based on the fluid pressure fluctuations, velocity abrupt changes, and vortex morphology data of the target subsystem, determine whether the vortex leads to an increase in entropy;

[0037] When the eddy current causes an increase in entropy, the rotational speed of the corresponding actuator in the target subsystem is increased to suppress the sudden increase in entropy.

[0038] Furthermore, the execution module further includes at least an integration channel unit, the integration channel unit being used for:

[0039] If the target subsystem accumulates ash and scale slowly for more than a preset time threshold, a cleaning command is triggered to keep the real-time entropy production rate within a preset fluctuation range.

[0040] Secondly, embodiments of this application also propose a method for regulating a thermodynamic system based on real-time entropy production feedback, comprising:

[0041] Based on the multi-dimensional physical field data of each target subsystem in the real-time thermal system, the real-time entropy production rate of each target subsystem is determined.

[0042] Based on the preset target entropy production rate and the real-time entropy production rate of each target subsystem, the entropy production change rate and cumulative entropy production deviation of each target subsystem are determined through the derivative and integral links in the adaptive gain PID control method. Based on the entropy production change rate, entropy production change rate gain, cumulative entropy production deviation and cumulative entropy production deviation gain of each target subsystem, the real-time control quantity of each target subsystem is obtained.

[0043] Based on the real-time control parameters of each target subsystem, the actuator parameters of each target subsystem are dynamically adjusted to drive the real-time entropy production rate of each target subsystem back to the target entropy production rate of each target subsystem.

[0044] Thirdly, embodiments of this application also propose an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-mentioned thermodynamic system control method based on real-time entropy production feedback.

[0045] Compared with the prior art, this application has the following advantages:

[0046] This application's embodiments construct a closed-loop "sensing-computation-control" system, using entropy production rate as the real-time feedback object to directly reflect the irreversibility of the energy conversion process, fundamentally solving the response lag problem caused by the reliance on temperature and pressure feedback in traditional PID control. Its dynamic control module employs an adaptive gain PID control strategy, optimizing gain parameters through load rate grading and reward functions to dynamically adapt to varying operating conditions, effectively overcoming the limitations of the fixed threshold method under varying load scenarios. The system addresses the efficiency degradation caused by a surge in losses. Simultaneously, a tiered adjustment mechanism, a differential channel (for rapid response to eddy current entropy increases caused by sudden crosswinds), and an integral channel (for continuous adjustment of slowly changing processes such as dust accumulation and fouling) cover both short-term disturbances and long-term degradation scenarios, resolving the additional issues caused by the mismatch between traditional rigid control and dynamic thermodynamic processes. The system addresses energy loss issues. It achieves precise tracking and dynamic suppression of the irreversibility of energy conversion in the thermal system, significantly reducing entropy production and irreversible energy loss under typical scenarios such as variable load, sudden eddy current disturbances, and long-term ash accumulation. The adaptive gain mechanism effectively suppresses overshoot and ensures system operational stability through dynamic optimization of control parameters. The continuous adjustment of ash and fouling by the integral channel further maintains heat transfer performance, comprehensively reducing system energy consumption. This achieves a dual improvement in both energy efficiency and operational stability of the thermal system. Attached Figure Description

[0047] Figure 1 This paper shows a schematic diagram of the structure of a thermodynamic system control system based on real-time entropy production feedback provided in an embodiment of this application.

[0048] Figure 2 A flowchart of a thermodynamic system control method based on real-time entropy production feedback provided in an embodiment of this application is shown;

[0049] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0050] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0051] Figure 1 A schematic diagram of the structure of a thermodynamic system control system based on real-time entropy production feedback provided in an embodiment of this application is shown. (Refer to...) Figure 1 This application provides a thermodynamic system control system based on real-time entropy production feedback, comprising:

[0052] The multi-source sensing and entropy production calculation terminal 10 is used to determine the real-time entropy production rate of each target subsystem based on the multi-dimensional physical field data of each target subsystem in the real-time collected thermal system.

[0053] Among them, multi-dimensional physical field data refers to the spatial distribution data covering multiple physical field dimensions such as temperature, pressure, flow rate, and concentration, which are collected in real time by the deployed sensor network during the operation of each subsystem in the target system. The real-time entropy production rate of each target subsystem refers to the amount of entropy generated per unit time by the subsystem at the current moment due to internal irreversible processes (such as heat conduction, viscous friction, diffusion, chemical reaction, and other dissipation effects). It is a key thermodynamic indicator for measuring the energy dissipation rate and irreversibility of the subsystem.

[0054] Specifically, the multi-dimensional physical field data includes: temperature field data, used to characterize the real-time temperature values ​​of each spatial point within the subsystem, reflecting the energy distribution state; pressure field data, used to characterize the pressure values ​​at each point, reflecting the mechanical environment of the fluid; velocity field data, used to characterize the fluid velocity and direction at each point, describing the transport characteristics of mass and momentum; and concentration field data, used for entropy production analysis of chemical reactions or diffusion processes. This application does not impose specific limitations on the embodiments.

[0055] Real-time entropy production rate is calculated based on the entropy balance equation of the second law of thermodynamics in existing technologies. The entropy production rate cloud map of each subsystem maps discrete entropy production rate density values ​​to the entire spatial domain of the subsystem, forming a continuous entropy production rate distribution field, which is then visualized using color gradients or contour lines. This is used to directly identify regions with significantly higher entropy production rates than the average level (such as areas in heat exchangers where low flow rates lead to thickened thermal boundary layers, or areas near pump impellers where viscous friction is intense), thus locating high-entropy production areas. The color intensity reflects the relative magnitude of entropy production rates in different regions, aiding in the identification of major dissipation processes. The continuous distribution of the entropy production rate cloud map allows for analysis of the variation patterns of entropy production rate along the flow direction and temperature gradient direction (e.g., entropy production rate gradually increases along the pipe axis).

[0056] By collecting multi-dimensional physical field data of each target subsystem in the thermal system in real time, the real-time entropy production rate of each target subsystem can be determined, which can accurately quantify the irreversibility of energy dissipation of each target subsystem and provide core indicators that directly reflect the operational health status of each target subsystem for dynamic regulation.

[0057] The dynamic control quantity generation module 20 is used to determine the entropy production change rate and cumulative entropy production deviation of each target subsystem based on the preset target entropy production rate and the real-time entropy production rate of each target subsystem, through the derivative and integral links in the adaptive gain PID control method. Based on the entropy production change rate, entropy production change rate gain, cumulative entropy production deviation, and cumulative entropy production deviation gain of each target subsystem, the real-time control quantity of each target subsystem is obtained.

[0058] Among them, the adaptive gain PID control method refers to the PID control strategy in which the gain parameter is dynamically adjusted according to the real-time state of the target subsystem. The preset target entropy production rate refers to the benchmark value of the entropy production rate set for each target subsystem, which represents the optimal irreversibility of each target subsystem. The entropy production change rate refers to the rate of change of the real-time entropy production rate of the target subsystem over time. The cumulative entropy production deviation refers to the cumulative error between the real-time entropy production rate and the target entropy production rate over a period of time.

[0059] In some embodiments, the dynamic control quantity generation module 20 includes at least an adaptive gain unit, which is used for:

[0060] The real-time load rate of each target subsystem is determined based on its maximum rated power and real-time output power.

[0061] When the rate of change of entropy production of each target subsystem is greater than the entropy production rate threshold, determine the load level of the real-time load rate of each target subsystem;

[0062] Based on the gain generation strategy corresponding to the load level, generate multiple gain combinations of entropy production change rate and cumulative entropy production deviation;

[0063] Calculate the reward value for each gain combination based on the preset reward function;

[0064] The gain combination corresponding to the maximum reward value is determined as the entropy production change rate gain and cumulative entropy production deviation gain for each target subsystem;

[0065] Based on the entropy production change rate and entropy production change rate gain of each target subsystem, as well as the cumulative entropy production deviation and cumulative entropy production deviation gain of each target subsystem, the real-time control quantity of each target subsystem is obtained.

[0066] The real-time load factor is calculated as the ratio of the target subsystem's real-time output power to its maximum rated power, used to quantify the load level under the current operating conditions. Load levels are defined according to preset load factor ranges; for example, a low load level corresponds to a load factor below 40%, and a high load level corresponds to a load factor above 80%. The gain generation strategy employs differentiated weighting principles for different load levels; for example, the weight of cumulative entropy production deviation is increased at low load levels, while the weight of entropy production change rate is increased at high load levels. The reward function uses entropy production rate deviation, the magnitude of control quantity changes, and system stability indicators as input parameters, and calculates the reward value through weighted averages.

[0067] Specifically, when the entropy production change rate of the target subsystem exceeds a preset threshold, the adaptive gain unit first calculates its real-time load rate and determines its load level. If it is at a low load level, it generates multiple combinations primarily based on the cumulative entropy production deviation gain; if it is at a high load level, it generates multiple combinations primarily based on the entropy production change rate gain. The reward value of each gain combination is calculated through real-time simulation of the control effect, and the combination with the largest reward value is selected as the current optimal gain parameter. For example, under low load conditions, the cumulative entropy production deviation gain is set to 0.8, and the entropy production change rate gain is set to 0.2; under high load conditions, the cumulative entropy production deviation gain is adjusted to 0.3, and the entropy production change rate gain is adjusted to 0.7. By dynamically adjusting the gain parameters, the control quantity can quickly respond to sudden entropy increases while avoiding the accumulation of steady-state errors, achieving the optimal control effect under different load conditions.

[0068] For example, the adaptive gain unit first determines the real-time load rate of each target subsystem based on its maximum rated power and real-time output power. For instance, for a steam turbine with a rated power of 100MW, a real-time load rate of 75% is achieved when the current real-time output power is 75MW. When the entropy production change rate of the target subsystem exceeds a preset entropy production rate threshold, the load level is determined based on the real-time load rate. Specifically, the load rate can be divided into three levels: low load (0-30%), medium load (30%-70%), and high load (70%-100%). Therefore, based on different load levels, corresponding gain generation strategies are adopted to generate multiple gain combinations with entropy production change rates and cumulative entropy production deviations. For example, at the low load level, gain combinations such as (0.5, 2.0), (0.8, 1.5), and (1.0, 1.2) can be generated, where the first value is the entropy production change rate gain, and the second value is the cumulative entropy production deviation gain. Next, a reward value for each gain combination is calculated using a preset reward function. The reward function can consider factors such as entropy production suppression effect and response speed. The gain combination that obtains the maximum reward value is determined as the entropy production change rate gain and the cumulative entropy production deviation gain of the current target subsystem. Finally, based on the entropy production change rate and corresponding gain of the target subsystem, as well as the cumulative entropy production deviation and corresponding gain, the real-time control quantity of the target subsystem is calculated. For example, a linear combination method can be used to calculate the real-time control quantity of the target subsystem.

[0069] The embodiments of this application dynamically adjust the control gain according to the real-time system load, maintaining good control performance under different operating conditions. Simultaneously, a reward mechanism is introduced to optimize gain selection, further improving the entropy production suppression effect. Furthermore, by considering both the entropy production change rate and the cumulative entropy production deviation, it can both quickly respond to system fluctuations and eliminate long-term deviations, thereby achieving more precise and stable entropy production control.

[0070] In some embodiments, based on a weighting strategy corresponding to the load level, multiple weighted combinations of entropy production change rate and cumulative entropy production deviation are generated, including:

[0071] When the load level is low, multiple first weight combinations are generated based on the principle that the cumulative entropy production deviation weight is greater than the entropy production change rate weight.

[0072] When the load level is high, multiple second weight combinations are generated based on the principle that the cumulative entropy production deviation weight is less than the entropy production change rate weight.

[0073] The load level classification is based on a comparison between the real-time load rate and a preset threshold. For example, a load rate below 40% is defined as a low load level, and above 80% is defined as a high load level. In the first weight combination, the cumulative entropy production deviation weight ranges from 0.6 to 0.8, and the entropy production change rate weight ranges from 0.2 to 0.4. In the second weight combination, the cumulative entropy production deviation weight is adjusted to 0.3 to 0.5, and the entropy production change rate weight is increased to 0.5 to 0.7. The weight combinations are generated through permutation and combination, with the sum of the weight coefficients of each combination remaining at 1. The generated weight combinations are input into a reward function for effect evaluation. The reward function includes the absolute value of the entropy production rate deviation, the fluctuation range of the control quantity, and equipment safety constraints.

[0074] Specifically, when the real-time load rate is at a low load level, the system prioritizes combinations with higher cumulative entropy production deviation weights. For example, under conditions of ash and scale buildup in the air-cooled condenser, the integral channel weight is increased to 0.7, prompting timely triggering of cleaning commands and controlling entropy production rate fluctuations within ±2%. When the load rate jumps to a high load level, the entropy production change rate weight is increased to 0.6, enabling the differential channel to detect entropy production rate anomalies within 0.2 seconds when sudden crosswinds cause eddies, and compressing the entropy increase from 42% to 8% by increasing the fan speed. The weight combination generation process is achieved through an exhaustive method, generating at least 5 candidate combinations for each load level. After evaluation, the combination with the highest reward value is selected for real-time control.

[0075] In some embodiments, the adaptive gain unit is further configured to:

[0076] Determine the real-time load change rate of each target subsystem based on its real-time load rate.

[0077] If the real-time load change rate of the first target subsystem exceeds the preset threshold load change rate, the reward value of the first target subsystem is generated by a preset reward function based on the real-time entropy production rate, entropy production change rate and real-time load rate of the first target subsystem.

[0078] Based on the range of the reward value of the first target subsystem, determine the correction direction and adjustment range of the entropy production change rate gain and the cumulative entropy production deviation gain of the first target subsystem;

[0079] Based on the correction direction and adjustment magnitude, adjust the entropy production change rate gain and cumulative entropy production deviation gain of the first target subsystem to maximize the reward value of the first target subsystem.

[0080] The entropy production rate gain and cumulative entropy production deviation gain corresponding to the maximum reward value of the first target subsystem are determined as the entropy production rate gain and cumulative entropy production deviation gain of the first target subsystem.

[0081] The real-time load change rate is calculated using sampling data from a power sensor, with a sampling interval of 200 milliseconds. The threshold load change rate is set at 15% of the device's rated power. When the real-time load change rate exceeds this threshold, a reward value calculation module based on the current entropy production rate, entropy production change rate, and load rate is triggered. The reward value range is divided into three levels, corresponding to negative correction of the gain parameter, maintaining the current value, and positive correction, respectively. The adjustment range is determined based on a lookup table trained from historical data, with a correction step size set to a variable range of 0.1-0.3.

[0082] Specifically, when a wind turbine experiences a load change rate of 18% due to turbulence, the system triggers a gain adjustment process. The real-time entropy production rate is 42 W / K, the entropy production change rate is +15 W / (K·s), and the load factor is 82%. The reward function calculates R = 3.2, falling within the range [0, 5). According to preset rules, the entropy production change rate gain needs to be corrected upwards by 0.2, and the cumulative entropy production deviation gain needs to be corrected downwards by 0.1. After adjustment, the reward value is recalculated and increased to 4.8, and the correction continues until the R value reaches above 5.0. Finally, the gain parameters are adjusted from the initial Kp = 1.2, Ki = 0.8 to Kp = 1.5, Ki = 0.6, reducing the response time of the control output by 40% and lowering the entropy production rate fluctuation range to within ±2 W / K.

[0083] For example, the adaptive gain unit determines the real-time load change rate based on the real-time load rate of each target subsystem. For instance, for a target subsystem, by collecting its output power data, the ratio of power change per unit time to rated power is calculated to obtain the real-time load change rate. When the real-time load change rate of the first target subsystem exceeds a preset threshold load change rate, a gain adjustment mechanism is triggered. Specifically, the threshold load change rate is set to 5% / minute. If the real-time load change rate is detected to reach 6% / minute, subsequent steps are initiated. Based on the real-time entropy production rate, entropy production change rate, and real-time load rate of the first target subsystem, a reward value is generated using a preset reward function. The calculated reward value is compared with a pre-divided reward interval to determine the correction direction and adjustment magnitude of the entropy production change rate gain and the cumulative entropy production deviation gain. For example, the reward value can be divided into three intervals: high, medium, and low, corresponding to increasing, maintaining, and decreasing the gain, respectively. Based on the determined correction direction and adjustment magnitude, the entropy production change rate gain and the cumulative entropy production deviation gain of the first target subsystem are adjusted. The adjustment step size can be set to ±10% of the current gain value. Repeat the above gain adjustment process until the reward value reaches its maximum. The entropy production change rate gain and the cumulative entropy production deviation gain corresponding to the maximum reward value are then used as the final determined values.

[0084] This application's embodiments employ adaptive adjustment of the entropy production feedback control gain. When the system load changes rapidly, the control parameters can be adjusted promptly, maintaining control sensitivity and stability. This improves the control performance of the thermal system under dynamic operating conditions, reduces entropy production fluctuations, and minimizes irreversible system losses.

[0085] In some embodiments, the dynamic control quantity generation module 20 further includes at least a target entropy yield determination unit, the target entropy yield determination unit being used for:

[0086] Based on the multi-dimensional physical field data of each target subsystem over a preset historical time period, determine the target subsystem at each historical moment. Efficiency and entropy production rate;

[0087] The efficiency exceeding the preset value in each target subsystem Efficiency is determined to be maximum Efficiency set, based on the maximum Each in the efficiency set The entropy yield corresponding to efficiency generates a candidate set of entropy yields.

[0088] Based on the premise of satisfying the equipment safety threshold and / or system stability boundary of each target subsystem, the minimum entropy production rate in the candidate set of entropy production rates is determined as the target entropy production rate of each target subsystem.

[0089] The preset historical time period selects a typical operating cycle of the system; for example, for thermal power units, it selects 720 hours of operating data including different load rates and ambient temperatures. The preset efficiency is set to historical. The 85th percentile of the efficiency distribution ensures the selection of high-efficiency operating points. Equipment safety thresholds include the pressure vessel's pressure limit and the vibration threshold of rotating parts. The system stability boundary is calculated using the Lyapunov exponent method, which will not be elaborated further in this application's embodiments.

[0090] Specifically, the multi-dimensional physical field data is cleaned and then input into... An efficiency calculation model, based on the first and second laws of thermodynamics. Maximum. The efficiency set was selected using a sliding time window algorithm, with a window width of 1 hour and a step size of 10 minutes. The candidate set of entropy productivity was established using an association analysis algorithm. The mapping relationship between efficiency and entropy productivity is established using the Pareto front method to extract the optimal solution set. Determining the minimum entropy productivity requires constraint verification; when a candidate value exceeds a safety threshold, a suboptimal value is automatically selected and an early warning is triggered. For example, a steam turbine at 90% load has historically reached its maximum... The efficiency corresponds to an entropy production rate of 1.2 kW / K, but the current cooling water temperature exceeds the standard. The system automatically adjusts the target value to 1.3 kW / K to ensure that the condenser vacuum is maintained within a safe range.

[0091] For example, the target entropy productivity determination unit first acquires multi-dimensional physical field data such as temperature, pressure, and flow rate for each target subsystem over the past 24 hours. Then, based on this historical data, it calculates the entropy productivity of each target subsystem at each historical moment. Efficiency and entropy production rate. Next, each target subsystem... Data points with an efficiency greater than 90% are determined as the maximum. Efficiency set. Based on this maximum Each in the efficiency set The entropy yield corresponding to efficiency generates a candidate set of entropy yields. Finally, under the premise of meeting the upper limit of safe temperature and the lower limit of safe pressure of the target subsystem equipment, the minimum entropy yield in the candidate set of entropy yields is determined as the target entropy yield of the first target subsystem. For example, for a steam turbine subsystem, by analyzing its 24-hour historical data, it is found that... There are 50 operating points with an efficiency greater than 90%. The entropy productivity corresponding to these 50 operating points is formed into a candidate set. Under the premise of ensuring that the turbine blade temperature does not exceed 550℃ and the steam pressure is not lower than 16MPa, the minimum entropy productivity of 2.5kJ / (kg·K) is selected from the candidate set as the target entropy productivity of the steam turbine.

[0092] This application's embodiments automatically determine the optimal target entropy production rate for each subsystem based on historical operating data, avoiding the irrationality that may result from manually setting target values. Simultaneously, by considering equipment safety thresholds and system stability boundaries, the feasibility and safety of the target entropy production rate are ensured. This method enables the thermal system to minimize irreversible losses and improve overall system efficiency while ensuring safe and stable operation.

[0093] The execution module 30 is used to dynamically adjust the actuator parameters of each target subsystem according to the real-time control amount of each target subsystem, and drive the real-time entropy production rate of each target subsystem to regress to the target entropy production rate of each target subsystem.

[0094] Specifically, the execution module 30 refers to the device that converts the control quantity into physical execution action. It is implemented by using a frequency converter to adjust the fan speed and an electric valve opening regulator to directly adjust the actuator parameters of the subsystem.

[0095] For example, the multi-source sensing and entropy production calculation terminal 10 collects multi-dimensional physical field data of each target subsystem in the thermal system and calculates the entropy production rate of each target subsystem in real time. The dynamic control quantity generation module 20 receives the preset target entropy production rate and real-time entropy production rate of each target subsystem, and uses the derivative and integral components in the adaptive gain PID control method to calculate the entropy production change rate and cumulative entropy production deviation, respectively. Then, based on the entropy production change rate, entropy production change rate gain, cumulative entropy production deviation, and cumulative entropy production deviation gain, it generates the real-time control quantity for each target subsystem. The execution module 30 dynamically adjusts the actuator parameters of each target subsystem according to the real-time control quantity, driving the real-time entropy production rate to return to the target entropy production rate.

[0096] The multi-source sensing and entropy production calculation terminal 10 collects physical field data such as temperature, pressure, and flow rate through multiple sensors and calculates the entropy production rate using thermodynamic formulas. In the dynamic control quantity generation module 20, the differential element calculates the rate of change of entropy production to quickly respond to transient changes, and the integral element calculates the cumulative deviation to eliminate steady-state errors. The adaptive gain mechanism dynamically adjusts the gain parameters according to the system operating conditions to improve control accuracy. The execution module 30 achieves closed-loop control of the entropy production rate by adjusting actuator parameters such as valve opening and pump speed.

[0097] In this embodiment, a dual-feedback mechanism—using a differential channel to suppress transient entropy increase and an integral channel to eliminate steady-state deviation—combined with adaptive gain adjustment, achieves dynamic optimal control under varying operating conditions, overcoming the minute-level delay bottleneck of traditional indirect control modes. This forms a closed-loop feedback control centered on entropy production rate, enabling real-time monitoring and dynamic suppression of irreversible losses in the thermal system. The differential-integral dual-channel structure can simultaneously handle sudden disturbances and gradually changing processes, while the adaptive gain mechanism adapts to varying operating conditions, thereby achieving optimal control of the thermal system. Dynamic optimization of efficiency.

[0098] This application's embodiments construct a closed-loop "sensing-computation-control" system, using entropy production rate as the real-time feedback object to directly reflect the irreversibility of the energy conversion process, fundamentally solving the response lag problem caused by the reliance on temperature and pressure feedback in traditional PID control. Its dynamic control module employs an adaptive gain PID control strategy, optimizing gain parameters through load rate grading and reward functions to dynamically adapt to varying operating conditions, effectively overcoming the limitations of the fixed threshold method under varying load scenarios. The system addresses the efficiency degradation caused by a surge in losses. Simultaneously, a tiered adjustment mechanism, a differential channel (for rapid response to eddy current entropy increases caused by sudden crosswinds), and an integral channel (for continuous adjustment of slowly changing processes such as dust accumulation and fouling) cover both short-term disturbances and long-term degradation scenarios, resolving the additional issues caused by the mismatch between traditional rigid control and dynamic thermodynamic processes. The system addresses energy loss issues. It achieves precise tracking and dynamic suppression of the irreversibility of energy conversion in the thermal system, significantly reducing entropy production and irreversible energy loss under typical scenarios such as variable load, sudden eddy current disturbances, and long-term ash accumulation. The adaptive gain mechanism effectively suppresses overshoot and ensures system operational stability through dynamic optimization of control parameters. The continuous adjustment of ash and fouling by the integral channel further maintains heat transfer performance, comprehensively reducing system energy consumption. This achieves a dual improvement in both energy efficiency and operational stability of the thermal system.

[0099] In some embodiments, the execution module 30 includes at least a hierarchical adjustment unit, which is used for:

[0100] Based on the pre-established correspondence between each target subsystem and the real-time control quantity, the actual control quantity of each target subsystem is determined;

[0101] Based on the actual control amount of each target subsystem and the weight parameters of each actuator in each target subsystem, determine the actual control amount of each actuator in each target subsystem;

[0102] Adjust the parameters of each actuator in each target subsystem according to the actual control amount of each actuator in each target subsystem.

[0103] In each target subsystem, the parameters of each actuator are determined based on the real-time control quantity and the weight parameters of each actuator in the first target subsystem. The pre-established correspondence is obtained through training with historical operating data and includes mapping rules between different entropy yield deviation ranges and control quantity amplitudes. The weight parameters are determined based on the thermodynamic influence coefficient of the actuator in the target subsystem, which is obtained by calculating the partial derivative of the power change of each actuator with respect to the entropy yield of the subsystem. Actuator parameter adjustment adopts an incremental control method, with each adjustment not exceeding 5% of the actuator's rated parameters.

[0104] For example, in a scenario involving the group control of air-cooled condenser fans, when an entropy production rate deviation of +8% is detected, the tiered adjustment unit first queries the corresponding relationship table to determine the total control amount as a 12% increase in fan speed. Subsequently, based on the flow field weight parameters of each fan's region (0.35 for region A, 0.45 for region B, and 0.20 for region C), it calculates that fans in region A need to increase their speed by 4.2%, those in region B by 5.4%, and those in region C by 2.4%. Finally, a speed adjustment command is sent to each fan controller via the CAN bus, enabling each fan to complete synchronized speed adjustment within 0.5 seconds. Compared to a uniform speed control scheme, this tiered control method increases the entropy production rate recovery speed by 40% and avoids secondary eddies caused by local over-adjustment.

[0105] In adjusting the parameters of each actuator in each target subsystem, the hierarchical adjustment unit first determines the actual control quantity of each target subsystem based on the pre-established correspondence between each target subsystem and the real-time control quantity. For example, for a heat exchanger subsystem, a mapping relationship between the real-time control quantity and parameters such as flow rate and temperature is pre-established. Further, the hierarchical adjustment unit determines the actual control quantity of each actuator based on the actual control quantity of each target subsystem and the weight parameters of each actuator. Specifically, assuming a subsystem has three actuators with weights of 0.5, 0.3, and 0.2, and an actual control quantity of 100 units, the actual control quantities of the three actuators are 50, 30, and 20 units, respectively. Finally, the hierarchical adjustment unit adjusts the corresponding actuator parameters based on the actual control quantity of each actuator. For example, for a valve actuator, the opening degree is adjusted; for a pump actuator, the speed is adjusted, etc. This achieves hierarchical and refined control of the entire thermal system.

[0106] Through the above technical solutions, the embodiments of this application achieve refined and hierarchical control of the thermal system. By considering the characteristics of different subsystems and actuators, a one-size-fits-all control approach is avoided, improving the overall operating efficiency of the system. Simultaneously, the pre-established correspondences and weight parameters ensure the rationality and reliability of the control, reducing system fluctuations caused by improper control. Furthermore, the hierarchical adjustment method also improves the system's response speed, enabling it to react more quickly to changes in entropy production, thereby more effectively suppressing irreversible losses in the system.

[0107] In some embodiments, the execution module 30 further includes at least a differential channel unit, which is used for:

[0108] In the event of a sudden crosswind causing eddies in the target subsystem, the system is triggered to detect fluid pressure fluctuations, velocity changes, and eddy morphology data of the target subsystem within a preset time.

[0109] Based on the fluid pressure fluctuations, velocity changes, and vortex morphology data of the target subsystem, determine whether the vortex leads to an increase in entropy;

[0110] When entropy increases due to eddy currents, the rotational speed of the corresponding actuator in the target subsystem is increased to suppress sudden entropy increases.

[0111] The trigger condition for detection is a vortex event caused by a sudden crosswind. Detection parameters include fluid pressure fluctuations, sudden changes in flow velocity, and vortex morphology data. The process of determining entropy increase is achieved by real-time analysis of the correlation between pressure fluctuation amplitude, flow velocity change rate, and vortex morphology. The operation of increasing the actuator speed is dynamically adjusted according to the degree of entropy increase; for example, the pressure fluctuation amplitude is mapped to a speed increment through a proportional relationship.

[0112] For example, when crosswinds induce eddies, multiphysics sensors collect fluid pressure, flow velocity, and eddy morphology data within 0.2 seconds. The edge computing unit calculates the entropy increase under the current eddy state using a pre-defined fluid dynamics model. If a pressure fluctuation exceeds a threshold and the direction of the sudden velocity change aligns with the eddy rotation direction, it is considered a valid entropy increase event. At this point, the controller sends a speed increase command to the fan actuator, with a command response delay of less than 10 milliseconds. Implementation examples show that this solution can increase the fan speed to 115% of the design value within 3 seconds, reducing the entropy increase caused by sudden crosswinds from 42% to 8%, effectively maintaining system stability. efficiency.

[0113] When a sudden crosswind causes eddies in the target subsystem, the differential channel unit triggers the detection of fluid pressure fluctuations, velocity changes, and eddy morphology data within a preset time. Specifically, the differential channel unit detects fluid pressure fluctuations using a pressure sensor, velocity changes using a flow meter, and eddy morphology data using an eddy detector. The preset time can be set to 10 seconds. Further, based on the detected fluid pressure fluctuations, velocity changes, and eddy morphology data, the differential channel unit determines whether the eddies cause entropy increase. For example, if the fluid pressure fluctuation exceeds 5%, the velocity change exceeds 10%, and the eddy morphology data shows an eddy intensity of level 3, it is determined that the eddies cause entropy increase. Therefore, if entropy increase is determined to be caused by eddies, the differential channel unit increases the rotational speed of the corresponding actuator in the target subsystem to suppress the sudden entropy increase. Specifically, the actuator speed can be increased by 20% for 30 seconds, and then gradually restored to normal speed.

[0114] This application embodiment rapidly detects and responds to entropy increase caused by sudden eddies. By monitoring changes in fluid parameters in real time, it promptly captures the moment eddies form and quickly takes measures to increase the actuator speed to suppress entropy increase. This effectively reduces energy loss caused by eddies and improves the overall efficiency of the system. Simultaneously, because the actuator speed is only increased briefly when necessary, energy waste caused by prolonged high-speed operation is avoided, achieving a balance between precise control and energy efficiency optimization.

[0115] In some embodiments, the execution module 30 further includes at least an integration channel unit, the integration channel unit being used for:

[0116] If the target subsystem accumulates dust and scale slowly for more than a preset time threshold, a cleaning command is triggered to keep the real-time entropy production rate within a preset fluctuation range.

[0117] The integration channel unit comprises a time threshold determination module and an instruction triggering module. The time threshold determination module continuously monitors the duration of the cumulative entropy production deviation. When the deviation exceeds a preset time threshold, it sends an activation signal to the instruction triggering module. The instruction triggering module has built-in cleaning program logic that generates a corresponding cleaning instruction sequence based on the activation signal. The preset time threshold ranges from 48 to 120 hours, with the specific value dynamically adjusted based on historical dust concentration data of the target subsystem. The preset fluctuation range is set to ±2% to ±5% of the target entropy production rate.

[0118] For example, the integral channel unit identifies slowly changing entropy production trends by integrating the accumulated entropy production deviation. When ash and scale buildup causes the entropy production rate to rise slowly at a rate of 0.05%-0.2% / hour, the cumulative deviation of the integral channel reaches the trigger threshold within a 72-hour continuous monitoring period. At this point, a cleaning command is activated, controlling the high-pressure water gun to perform a pulsed flushing operation. During the flushing process, real-time entropy production rate monitoring data is fed back to the integral channel. When the entropy production rate is detected to fall back to the preset fluctuation range, the cleaning command automatically terminates. This process allows the system to intervene in the early stages of ash and scale buildup, avoiding irreversible efficiency losses caused by the continuous accumulation of deviations.

[0119] For example, in a thermal power plant boiler system, when the heating surfaces of the tail flue slowly accumulate ash and scale due to long-term operation, the integral channel unit continuously monitors the real-time entropy production rate trend of this subsystem. When the entropy production rate is detected to show a monotonically increasing trend for 72 consecutive hours and exceed the preset fluctuation threshold, the integral channel unit automatically generates a cleaning control command. This command, through the DCS system, links the steam soot blowing device to control the rotating nozzles to periodically purge the heating surface tube bundle at a steam pressure of 0.8 MPa. Each purging lasts for 120 seconds, with the interval adjusted to 15 minutes. During the purging process, the entropy production rate change curve is monitored in real time, and the cleaning process is terminated when the entropy production rate is detected to return to the center value of the preset fluctuation band.

[0120] Through the above technical solution, this application effectively solves the problem of continuous deviation in entropy production rate caused by slow scaling during long-term operation of thermal equipment. It accurately identifies gradual anomalies through the time accumulation effect of the integral channel, triggering targeted maintenance operations and avoiding energy efficiency losses caused by adjustment lag in traditional control methods. This ensures the system maintains thermodynamic equilibrium under slow scaling conditions, reducing the probability of unplanned shutdowns.

[0121] Secondly, Figure 2 A flowchart of a thermodynamic system control method based on real-time entropy production feedback according to an embodiment of this application is shown. (Refer to...) Figure 2 This application also proposes a method for regulating a thermodynamic system based on real-time entropy production feedback, including:

[0122] S10. Determine the real-time entropy production rate of each target subsystem based on the multi-dimensional physical field data of each target subsystem in the real-time collected thermal system.

[0123] S20. Based on the preset target entropy production rate and the real-time entropy production rate of each target subsystem, the entropy production change rate and cumulative entropy production deviation of each target subsystem are determined through the derivative and integral links in the adaptive gain PID control method. Based on the entropy production change rate, entropy production change rate gain, cumulative entropy production deviation and cumulative entropy production deviation gain of each target subsystem, the real-time control quantity of each target subsystem is obtained.

[0124] S30. Based on the real-time control amount of each target subsystem, dynamically adjust the actuator parameters of each target subsystem to drive the real-time entropy production rate of each target subsystem back to the target entropy production rate of each target subsystem.

[0125] In some embodiments, the real-time control parameters for each target subsystem are obtained according to the following steps:

[0126] The real-time load rate of each target subsystem is determined based on its maximum rated power and real-time output power.

[0127] When the rate of change of entropy production of each target subsystem is greater than the entropy production rate threshold, determine the load level of the real-time load rate of each target subsystem;

[0128] Based on the gain generation strategy corresponding to the load level, generate multiple gain combinations of entropy production change rate and cumulative entropy production deviation;

[0129] Calculate the reward value for each gain combination based on the preset reward function;

[0130] The gain combination corresponding to the maximum reward value is determined as the entropy production change rate gain and cumulative entropy production deviation gain for each target subsystem;

[0131] Based on the entropy production change rate and entropy production change rate gain of each target subsystem, as well as the cumulative entropy production deviation and cumulative entropy production deviation gain of each target subsystem, the real-time control quantity of each target subsystem is obtained.

[0132] In some embodiments, based on a weighting strategy corresponding to the load level, multiple weighted combinations of entropy production change rate and cumulative entropy production deviation are generated, including:

[0133] When the load level is low, multiple first weight combinations are generated based on the principle that the cumulative entropy production deviation weight is greater than the entropy production change rate weight.

[0134] When the load level is high, multiple second weight combinations are generated based on the principle that the cumulative entropy production deviation weight is less than the entropy production change rate weight.

[0135] In some embodiments, the process of determining the entropy production rate change gain and the cumulative entropy production deviation gain includes:

[0136] Determine the real-time load change rate of each target subsystem based on its real-time load rate.

[0137] If the real-time load change rate of the first target subsystem exceeds the preset threshold load change rate, the reward value of the first target subsystem is generated by a preset reward function based on the real-time entropy production rate, entropy production change rate and real-time load rate of the first target subsystem.

[0138] Based on the range of the reward value of the first target subsystem, determine the correction direction and adjustment range of the entropy production change rate gain and the cumulative entropy production deviation gain of the first target subsystem;

[0139] Based on the correction direction and adjustment magnitude, adjust the entropy production change rate gain and cumulative entropy production deviation gain of the first target subsystem to maximize the reward value of the first target subsystem.

[0140] The entropy production rate gain and cumulative entropy production deviation gain corresponding to the maximum reward value of the first target subsystem are determined as the entropy production rate gain and cumulative entropy production deviation gain of the first target subsystem.

[0141] In some embodiments, the process of determining the target entropy yield of each target subsystem includes:

[0142] Based on the multi-dimensional physical field data of each target subsystem over a preset historical time period, determine the target subsystem at each historical moment. Efficiency and entropy production rate;

[0143] The efficiency exceeding the preset value in each target subsystem Efficiency is determined to be maximum Efficiency set, based on the maximum Each in the efficiency set The entropy yield corresponding to efficiency generates a candidate set of entropy yields.

[0144] Based on the premise of satisfying the equipment safety threshold and / or system stability boundary of each target subsystem, the minimum entropy production rate in the candidate set of entropy production rates is determined as the target entropy production rate of each target subsystem.

[0145] In some embodiments, the adjustment process for each actuator parameter in each target subsystem includes:

[0146] Based on the pre-established correspondence between each target subsystem and the real-time control quantity, the actual control quantity of each target subsystem is determined;

[0147] Based on the actual control amount of each target subsystem and the weight parameters of each actuator in each target subsystem, determine the actual control amount of each actuator in each target subsystem;

[0148] Adjust the parameters of each actuator in each target subsystem according to the actual control amount of each actuator in each target subsystem.

[0149] In some embodiments, when a sudden crosswind causes eddies in the target subsystem, the system is triggered to detect fluid pressure fluctuations, velocity changes, and eddy morphology data of the target subsystem within a preset time.

[0150] Based on the fluid pressure fluctuations, velocity changes, and vortex morphology data of the target subsystem, determine whether the vortex leads to an increase in entropy;

[0151] When entropy increases due to eddy currents, the rotational speed of the corresponding actuator in the target subsystem is increased to suppress sudden entropy increases.

[0152] In some embodiments, if the target subsystem slowly accumulates ash and scale for more than a preset time threshold, a cleaning command is triggered to keep the real-time entropy production rate within a preset fluctuation range.

[0153] Thirdly, Figure 3 A schematic diagram of the structure of an electronic device according to an embodiment of this application is shown. (Refer to...) Figure 3 This application also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-mentioned thermodynamic system control method based on real-time entropy production feedback.

[0154] In this embodiment, the computer device includes a processor, memory, and network interface connected via a system bus.

[0155] The processor of this computer device provides computational and control capabilities. The memory of the computer device includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is used to store data samples. The network interface of the computer device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements any thermodynamic system control method based on real-time entropy production feedback.

[0156] Those skilled in the art will understand that Figure 3The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0157] As the method embodiments are basically similar to the system embodiments, the description is relatively simple, and relevant parts can be found in the description of the system embodiments.

[0158] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0159] The above provides a detailed description of the thermodynamic system control system, method, and electronic device based on real-time entropy production feedback provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A real-time entropy production feedback based thermal system regulation system, characterized in that, The application relates to a multi-source perception and entropy production calculation terminal for determining the real-time entropy production rate of each target subsystem in a thermal system according to multi-dimensional physical field data of the each target subsystem collected in real time. A dynamic control quantity generation module is configured to determine the entropy production change rate and cumulative entropy production deviation of the each target subsystem through the differential link and integral link in a PID control method with adaptive gain according to the preset target entropy production rate of the each target subsystem and the real-time entropy production rate of the each target subsystem, and obtain the real-time control quantity of the each target subsystem according to the entropy production change rate, entropy production change rate gain, cumulative entropy production deviation and cumulative entropy production deviation gain of the each target subsystem. An execution module is configured to dynamically adjust the actuator parameters of the each target subsystem according to the real-time control quantity of the each target subsystem, and drive the real-time entropy production rate of the each target subsystem to return to the target entropy production rate of the each target subsystem. The dynamic control quantity generation module comprises at least an adaptive gain unit, which is configured to:

2. The system of claim 1, wherein, determine the real-time load rate of the each target subsystem according to the maximum rated power and real-time output power of the each target subsystem; determine the load level of the real-time load rate of the each target subsystem in the case that the entropy production change rate of the each target subsystem is greater than an entropy production rate threshold value; generate a plurality of gain combinations of the entropy production change rate and cumulative entropy production deviation according to a gain generation strategy corresponding to the load level; calculate the reward value of each gain combination according to a preset reward function; determine the gain combination corresponding to the maximum reward value as the entropy production change rate gain and cumulative entropy production deviation gain of the each target subsystem; obtain the real-time control quantity of the each target subsystem according to the entropy production change rate and entropy production change rate gain of the each target subsystem, and the cumulative entropy production deviation and cumulative entropy production deviation gain of the each target subsystem. generate a plurality of weight combinations of the entropy production change rate and cumulative entropy production deviation according to a weight generation strategy corresponding to the load level, comprising:

3. The system of claim 2, wherein, generate a plurality of first weight combinations according to the principle that the cumulative entropy production deviation weight is greater than the entropy production change rate weight in the case that the load level is a low load level; generate a plurality of second weight combinations according to the principle that the cumulative entropy production deviation weight is less than the entropy production change rate weight in the case that the load level is a high load level. The adaptive gain unit is further configured to:

4. The system of claim 2, wherein, determine the real-time load change rate of the each target subsystem according to the real-time load rate of the each target subsystem; generate the reward value of the first target subsystem through the preset reward function according to the real-time entropy production rate, entropy production change rate and real-time load rate of the first target subsystem in the case that the real-time load change rate of the first target subsystem exceeds a preset threshold load change rate; determine the correction direction and adjustment amplitude of the entropy production change rate gain and cumulative entropy production deviation gain of the first target subsystem according to the interval of the reward value of the first target subsystem. ​ adjust, according to the correction direction and the adjustment amplitude, the entropy production rate change gain and the cumulative entropy production deviation gain of the first target subsystem to maximize a reward value of the first target subsystem; determine the entropy production rate change gain and the cumulative entropy production deviation gain corresponding to the maximum reward value of the first target subsystem as the entropy production rate change gain and the cumulative entropy production deviation gain of the first target subsystem.

5. The system of claim 1, wherein, The dynamic regulation amount generation module further comprises a target entropy production rate determination unit, which is configured to: determining, according to the multi-dimensional physical field data of each target subsystem in a preset historical time period, the multi-dimensional physical field data of each target subsystem at each historical time point efficiency and entropy production rate; determining an efficiency set in each of the target subsystems efficiency is determined as maximum efficiency set, according to the maximum each of the efficiency set entropy production rate corresponding to each of the efficiency set generates an entropy production rate candidate set; determine the minimum entropy production rate in the entropy production rate candidate set as the target entropy production rate of each target subsystem, under the premise that the equipment safety threshold and / or the system stability boundary of each target subsystem are satisfied.

6. The system of claim 1, wherein, The execution module comprises a hierarchical adjustment unit, which is configured to: determine the actual regulation amount of each target subsystem according to the corresponding relationship between each target subsystem and the real-time regulation amount established in advance; determine the actual regulation amount of each actuator in each target subsystem according to the actual regulation amount of each target subsystem and the weight parameter of each actuator in the target subsystem; adjust the parameter of each actuator in each target subsystem according to the actual regulation amount of each actuator in the target subsystem.

7. The system of claim 1, wherein, The execution module further comprises a differential channel unit, which is configured to: trigger the detection of the fluid pressure fluctuation, flow rate mutation and vortex pattern data of the target subsystem within a preset time in the case of vortex caused by sudden wind of the target subsystem; determine whether the vortex causes entropy increase according to the fluid pressure fluctuation, flow rate mutation and vortex pattern data of the target subsystem; in the case that the vortex causes entropy increase, increase the rotating speed of the corresponding actuator in the target subsystem to suppress the sudden entropy increase.

8. The system of claim 1, wherein, The execution module further comprises an integral channel unit, which is configured to: trigger a cleaning instruction to maintain the real-time entropy production rate within a preset fluctuation range in the case that the slow fouling of the target subsystem exceeds the preset time threshold.

9. A method for regulating a thermodynamic system based on real-time entropy production feedback, characterized in that, The method comprises: determining the real-time entropy production rate of each target subsystem in the heat system according to the multi-dimensional physical field data of each target subsystem collected in real time; determining the entropy production rate change and the cumulative entropy production deviation of each target subsystem through the differential link and the integral link in the PID control method with adaptive gain according to the preset target entropy production rate of each target subsystem and the real-time entropy production rate of each target subsystem, and obtaining the real-time regulation amount of each target subsystem according to the entropy production rate change, the entropy production rate change gain, the cumulative entropy production deviation and the cumulative entropy production deviation gain of each target subsystem; dynamically adjusting the actuator parameter of each target subsystem according to the real-time regulation amount of each target subsystem to drive the real-time entropy production rate of each target subsystem to return to the target entropy production rate of each target subsystem.

10. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to realize the heat system regulation method based on real-time entropy production feedback according to claim 9. The processor executes the computer program to realize the heat system regulation method based on real-time entropy production feedback according to claim 9.

Images