Intelligent control method and system for heat supply system

Abstract

The invention discloses an intelligent control method and system for a heat supply system, and the method comprises the steps: calculating a temperature control response deviation ratio of each control period according to the difference between actual room temperatures and the difference between predicted room temperatures of adjacent control periods, calculating a response deviation duration index according to the time domain characteristics of a temperature control response deviation ratio sequence, and calculating the temperature control response deviation of each control period; according to mean value characteristics and fluctuation characteristics of a temperature control response deviation ratio sequence and a first-order difference sequence thereof, calculating a thermal characteristic change mode identification index, and combining the two to calculate update quantities of a heat transfer coefficient and a thermal inertia time constant; the room temperature of the next control period is predicted according to the updated heat transfer coefficient and the updated thermal inertia time constant, and an objective function of a multi-objective convex optimization problem is constructed based on the predicted room temperature and the valve adjustment amplitude of all users in the unit; and solving to obtain an optimal valve opening instruction to regulate and control the valve. The precision and stability of thermodynamic balance control are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of heating system. More particularly, the present application relates to a heating system intelligent control method and system. BACKGROUND

[0002] In the central heating system, "uneven heating" is a long-standing technical problem, which is manifested as the room temperature of the near-end user being too high and the room temperature of the far-end user being insufficient, seriously affecting user comfort and energy utilization efficiency; the traditional solution mainly relies on static hydraulic balance adjustment, which realizes basic flow distribution through the installation of fixed flow valves or differential pressure control valves.

[0003] However, this method cannot cope with the dynamic changes of building thermal characteristics in the heating system, such as the instantaneous increase of heat transfer coefficient caused by user window opening behavior, the change of internal heat source caused by personnel activity and sunlight irradiation, and the degradation of building thermal insulation performance over time, etc.

[0004] In recent years, intelligent heating control systems based on edge computing have gradually emerged, which realize local collaborative optimization by deploying intelligent valves in user branch; among them, multi-objective convex optimization algorithm has become the core control means due to its high calculation efficiency, ability to optimize thermal balance and operation stability, etc.

[0005] However, this algorithm highly depends on accurate thermal and hydraulic models, especially building thermal characteristic parameters, i.e. thermal inertia time constant and heat transfer coefficient ; when the model parameters are mismatched with the actual system, the optimization result will deviate seriously from the true optimal solution.

[0006] The existing technology usually adopts fixed model parameters or simple correction based on historical data, which cannot capture the dynamic changes of building thermal characteristics in real time; therefore, there is an urgent need for an adaptive online parameter correction method to improve the adaptability and control accuracy of multi-objective convex optimization algorithm in dynamic environment. SUMMARY

[0007] To solve the above technical problems that the existing mold life prediction model is single, cannot comprehensively reflect the change rule of mold life under different conditions, and the adaptability and accuracy are limited, the present application provides solutions in the following aspects.

[0008] In a first aspect, the present invention provides an intelligent control method for a heating system, comprising: collecting real-time data from the user in each control cycle, including actual room temperature and valve opening; predicting the room temperature of the current control cycle based on the user's heat transfer coefficient, thermal inertia time constant, actual room temperature, and valve opening in the previous control cycle; using the ratio of the difference between the actual room temperature and the difference between the predicted room temperature in adjacent control cycles as the temperature control response deviation rate for each control cycle; calculating the response deviation persistence index of the current control cycle based on the time-domain characteristics of the temperature control response deviation rate sequence in the analysis window of the user in the current control cycle; when the response deviation persistence index is greater than a persistence threshold: based on the mean value of the temperature control response deviation rate sequence and its first-order difference sequence... Based on the value characteristics and fluctuation characteristics, the thermal characteristic change pattern recognition index for the current control cycle is calculated. Combined with the temperature control response deviation rate, the update amount of the heat transfer coefficient and thermal inertia time constant for the current control cycle is calculated, and the heat transfer coefficient and thermal inertia time constant are updated. The valve opening of the user in the current control cycle is taken as the object to be solved: based on the updated heat transfer coefficient, thermal inertia time constant, actual room temperature and valve opening of the user in the current control cycle, the room temperature of the next control cycle is predicted. Based on the predicted room temperature and valve adjustment range of all users in the unit in the next control cycle, the objective function of the multi-objective convex optimization problem is constructed. The objective function is solved to obtain the optimal valve opening command for each user in the unit, and the valve opening is adjusted.

[0009] Preferably, predicting the room temperature for the current control cycle based on the user's heat transfer coefficient, thermal inertia time constant, actual room temperature, and valve opening in the previous control cycle includes: In the formula, For users In the Predicted room temperature for each control cycle; For users In the The actual room temperature for each control cycle; To control the duration of the cycle; For users In the The thermal inertia time constant of each control cycle; To set the temperature; For users In the Heat transfer coefficient for each control cycle; For users In the The flow rate of each control cycle, and , For users In the Valve opening degree per control cycle This is the proportionality coefficient.

[0010] Preferably, the method for obtaining the analysis window of the current control cycle is as follows: [The text abruptly ends here, so the translation stops.] The control cycle to the first The control cycle constitutes the first control cycle. Analysis window for each control cycle, To analyze the length of the window; the user In the The temperature control response deviation rate of all temperature control responses in the analysis window of each control cycle constitutes the user's... In the Temperature control response deviation rate sequence for each control cycle.

[0011] Preferably, the step of calculating the response deviation persistence index for the current control cycle based on the time-domain characteristics of the temperature control response deviation rate sequence in the user's analysis window of the current control cycle includes: In the formula, For users In the The duration index of response deviation for each control cycle; The length of the analysis window; For users exist Temperature control response deviation rate per control cycle; This is an indicator function; it returns 1 when the condition is met and 0 otherwise. This is the deviation threshold; This indicates taking the absolute value.

[0012] Preferably, the step of calculating the thermal characteristic change pattern recognition index for the current control cycle based on the mean and fluctuation characteristics of the temperature control response deviation rate sequence and its first-order difference sequence includes: In the formula, For users In the Thermal characteristic change pattern recognition index for each control cycle; For users In the Standard deviation of the temperature control response deviation rate sequence for each control cycle; For users In the The mean of the temperature control response deviation rate sequence for each control cycle; For users In the The standard deviation of the first difference sequence of the temperature control response deviation rate sequence for each control cycle; This indicates taking the absolute value.

[0013] Preferably, the step of calculating the update amount of the heat transfer coefficient and thermal inertia time constant for the current control cycle by combining the temperature control response deviation rate includes: ; In the formula, For users In the Heat transfer coefficient per control cycle Update volume; For users In the Thermal inertia time constant for each control cycle Update volume; For users In the The mean of the temperature control response deviation rate sequence for each control cycle; For users In the The thermal inertia time constant of each control cycle; For users In the Heat transfer coefficient for each control cycle; For users In the Thermal characteristic change pattern recognition index for each control cycle; For users In the Pattern recognition weights for each control cycle.

[0014] Preferably, the user In the The formula for calculating the pattern recognition weight for each control cycle is: In the formula, For users In the Pattern recognition weights for each control cycle; For users In the Thermal characteristic change pattern recognition index for each control cycle; The shape parameter of the sigmoid function; It is the center point of the sigmoid function.

[0015] Preferably, the updating of the heat transfer coefficient and thermal inertia time constant includes: ; In the formula, For users In the The updated heat transfer coefficient for each control cycle; For users In the The updated thermal inertia time constant for each control cycle; For users In the Heat transfer coefficient per control cycle Update volume; For users In the Thermal inertia time constant for each control cycle Update volume; This is the learning rate.

[0016] Preferably, the expression for the objective function is: ;in, , For users In the Valve opening degree per control cycle The number of all users within the unit; In the first The maximum temperature difference within the unit during each control cycle is equal to the temperature difference between all users within the unit during the first control cycle. The difference between the maximum and minimum predicted room temperature for each control cycle; For users In the The valve adjustment range for each control cycle, and ; This indicates taking the absolute value.

[0017] Secondly, the present invention provides an intelligent control system for a heating system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned intelligent control method for a heating system is implemented.

[0018] By adopting the above technical solution, a computer program is generated from the above-mentioned intelligent control method for a heating system and stored in a memory so that it can be loaded and executed by a processor. Terminal devices can then be made based on the memory and the processor for convenient use.

[0019] The beneficial effects of this invention are as follows:

[0020] This invention accurately identifies the type and degree of model mismatch by constructing a temperature control response deviation rate sequence and analyzing its time-domain characteristics, avoiding the continuous misjudgment problem caused by model mismatch in traditional methods. By introducing a thermal characteristic change pattern recognition index, it distinguishes between abrupt changes in heat transfer coefficient and changes in thermal inertia, enabling targeted parameter updates and ensuring that the system maintains good control performance even in scenarios with frequent changes in user behavior. In summary, this invention effectively solves the model mismatch problem caused by dynamic changes in building thermal characteristics in heating systems using multi-objective convex optimization algorithms, significantly improving the accuracy and stability of thermal balance control, and providing key technical support for the efficient operation of smart heating systems. Attached Figure Description

[0021] The above and other objects, features, and advantages of the present invention will become readily apparent from the following detailed description of exemplary embodiments, accompanied by the accompanying drawings. In the drawings, several embodiments of the invention are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein:

[0022] Figure 1 This is a flowchart illustrating an intelligent control method for a heating system according to the present invention;

[0023] Figure 2 This is a flowchart illustrating step S2 in the present invention;

[0024] Figure 3 This is a flowchart illustrating step S3 in the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0027] This invention discloses an intelligent control method for a heating system, referring to... Figure 1 This includes steps S1 to S4:

[0028] S1: Collect real-time data from the user during each control cycle, including actual room temperature and valve opening.

[0029] Edge computing is a distributed computing approach that moves data processing and analysis from a centralized cloud to edge devices or nodes closer to the data source. In this way, edge computing can significantly reduce data transmission latency, reduce bandwidth consumption, and improve the real-time performance and reliability of the system.

[0030] Key components of edge computing include edge devices and edge nodes. Edge devices are located at the source of data generation. In the heating system scenario, edge devices include temperature sensors that collect indoor temperature. Edge nodes serve as an intermediate layer between edge devices and the cloud, responsible for data aggregation, preprocessing, and preliminary analysis. In the heating system scenario, each unit has an edge node, which is used to aggregate data from all users within the unit and optimize the valve opening for all users within the unit.

[0031] During each control cycle, real-time data from the user is collected, including room temperature and valve opening. The room temperature data comes from the user's temperature sensor. The valve opening value ranges from [0,1], representing a continuous value from 0 to 1, indicating the valve is fully closed to fully open. Initially, in the first control cycle, the valve is fully open or half-open, so the valve opening is 1 or 0.5. The valve opening in subsequent control cycles is obtained by constructing and solving the objective function of a multi-objective convex optimization problem, with the valve opening as the object to be solved.

[0032] In heating systems, the choice of control cycle needs to balance response speed and system stability. Therefore, 5 minutes is chosen as the basic control cycle because the thermal inertia of the heating system determines that the room temperature changes relatively slowly. 5 minutes can capture meaningful temperature control response without causing system oscillation due to too frequent adjustments.

[0033] S2: Based on the user's heat transfer coefficient, thermal inertia time constant, actual room temperature, and valve opening in the previous control cycle, predict the user's room temperature in the current control cycle. Based on the difference between the actual room temperature and the difference between the predicted room temperature in adjacent control cycles, calculate the temperature control response deviation rate for each control cycle. Based on the time-domain characteristics of the temperature control response deviation rate sequence, calculate the response deviation persistence index.

[0034] It should be noted that in heating systems, the primary manifestation of model mismatch is the deviation between the actual temperature control response and the predicted response. By analyzing the characteristics of this deviation, the type and degree of model parameter mismatch can be deduced. The key to this step is to transform the raw temperature control data into meaningful feature indicators, providing a basis for subsequent model updates.

[0035] The flowchart for step S2 is shown below. Figure 2 The process includes steps S201 to S203, specifically as follows:

[0036] S201: The user's heat transfer coefficient, thermal inertia time constant, actual room temperature, and valve opening in the previous control cycle are used to predict the user's room temperature in the current control cycle.

[0037] Specifically, based on the user's heat transfer coefficient, thermal inertia time constant, actual room temperature, and valve opening in the previous control cycle, the room temperature for the current control cycle is predicted. The expression for the room temperature prediction model is as follows:

[0038] ;

[0039] In the formula, For users In the Predicted room temperature for each control cycle, in °C; For users In the The actual room temperature for each control cycle; To control the duration of the cycle, it is fixed at 5 minutes; For users In the The thermal inertia time constant for each control cycle, in seconds; To set the temperature, the range is typically [18, 22], in °C. For users In the The heat transfer coefficient for each control cycle is expressed in W / m²·K (watts per square meter per degree). For users In the The flow rate for each control cycle is expressed in m³ / h (cubic meters per hour). , For users In the Valve opening degree per control cycle This is the proportionality coefficient.

[0040] The initial value of the thermal inertia time constant is set according to the building type. For residential buildings, the initial value of the thermal inertia time constant ranges from [18000, 36000] in seconds. For office buildings, the initial value of the thermal inertia time constant ranges from [9000, 18000] in seconds. The initial value of the heat transfer coefficient is set according to the building insulation level. When the building insulation level is "good", the heat transfer coefficient ranges from [0.8, 1.2]. When the building insulation level is "average", the heat transfer coefficient ranges from [1.2, 1.8]. When the building insulation level is "poor", the heat transfer coefficient ranges from [1.8, 2.5].

[0041] Among them, the proportionality coefficient The ratio is set based on building type and user area; for residential buildings, the ratio is [not specified]. Equals 0.0045 multiplied by the user area; for office buildings, this is the scaling factor. It equals 0.0055 multiplied by the user area.

[0042] This formula originates from the discretization of the first-order differential equation of building thermal dynamics. It describes the relationship between the rate of change of room temperature and the current room temperature deviation and valve opening input, and is the core of the heating system model. Through this formula, the system can predict the room temperature change trend in the next control cycle under the current valve opening, providing a basis for optimization decisions.

[0043] S202: Calculate the temperature control response deviation rate of the user in each control cycle based on the difference between the actual room temperature and the predicted room temperature in adjacent control cycles.

[0044] It should be noted that by constructing the temperature control response deviation rate as a basic feature, the difference between the actual thermal characteristics and the model assumptions can be directly reflected.

[0045] Specifically, based on the difference between the actual room temperature and the predicted room temperature in adjacent control cycles, the user's temperature control response deviation rate in each control cycle is calculated. This rate represents the ratio of the actual temperature rise to the predicted temperature rise in each control cycle. The specific calculation formula is as follows:

[0046] ;

[0047] In the formula, For users In the Temperature control response deviation rate per control cycle; For users In the The actual room temperature for each control cycle; For users In the The actual room temperature for each control cycle; For users exist Predicted room temperature at any given time; For users In the Predicted room temperature for each control cycle.

[0048] It should be noted that, in order to avoid measurement noise interference, when When the value is less than the accuracy range of the temperature sensor, it is considered that there is no significant temperature control change. Setting it to 1 ensures that only meaningful temperature control responses are analyzed. For example, when the temperature sensor's accuracy range is [-0.5, 0.5], the accuracy range of the temperature sensor is 1.

[0049] This formula quantifies the accuracy of the model's predictions by comparing the ratio of the actual temperature rise to the predicted temperature rise; ideally, the two should be equal. =1; when the model mismatch occurs. It will deviate from 1; This provides a basic indicator for subsequent analysis, and its value directly reflects the degree of model mismatch: when A value less than 1 indicates that the actual temperature rise is less than predicted, possibly due to a lower heat transfer coefficient. Increase, for example, by opening a window; when A value >1 indicates that the actual temperature rise is greater than predicted, possibly due to additional heat sources, such as sunlight.

[0050] S203: Calculate the user's response deviation persistence index in the current control cycle based on the time-domain characteristics of the temperature control response deviation rate sequence in the analysis window of the current control cycle.

[0051] It should be noted that by introducing time-domain characteristic analysis of the temperature control response deviation rate and constructing a response deviation persistence index, different types of mismatches can be distinguished, namely transient noise or systematic mismatch, thus avoiding erroneous updates.

[0052] Specifically, the first The control cycle to the first The control cycle constitutes the first control cycle. Analysis window for each control cycle, The length of the analysis window is fixed at 12; the user... In the The temperature control response deviation rate of all temperature control responses in the analysis window of each control cycle constitutes the user's... In the Temperature control response deviation rate sequence for each control cycle.

[0053] Furthermore, user statistics In the The proportion of temperature control response deviation rate deviating from the value of 1 in the analysis window of each control cycle is used to extract the time-domain features of the temperature control response deviation rate and construct a user... In the The response deviation duration index for each control cycle is used to distinguish between transient noise and systematic mismatch.

[0054] user In the The formula for calculating the response deviation persistence index for each control cycle is:

[0055] ;

[0056] In the formula, For users In the The duration index of response deviation for each control cycle; The length of the analysis window; For users exist Temperature control response deviation rate per control cycle; This is an indicator function; it returns 1 when the condition is met and 0 otherwise. This is the deviation threshold; This indicates taking the absolute value.

[0057] This formula statistically analyzes the temperature control response deviation rate within the analysis window. The proportion of significant deviations from 1 is used to distinguish between transient noise and systematic mismatch: when the deviation persists, it indicates that there is indeed a model mismatch, rather than random fluctuations; therefore, the obtained response deviation persistence index... It quantifies the persistence of the bias; the larger the value, the more severe and persistent the model mismatch.

[0058] Among them, for the deviation threshold Based on the typical temperature control response characteristics of heating systems, the following settings are made: Under normal operating conditions, the fluctuation of the temperature control response deviation rate caused by random noise is usually less than ±15%, while the mismatch of the actual model usually causes a deviation of more than ±25%. Therefore, the deviation threshold is set as follows: A value of 0.2 indicates that a temperature control response deviation rate exceeding 20% ​​from 1 is considered a significant deviation.

[0059] S3: When the user's response deviation persistence index in the current control cycle is greater than the persistence threshold, calculate the thermal characteristic change pattern recognition index based on the mean and fluctuation characteristics of the temperature control response deviation rate sequence and its first-order difference sequence, and calculate the update amount of the heat transfer coefficient and thermal inertia time constant in combination with the temperature control response deviation rate.

[0060] In heating systems, the heat transfer coefficient Sudden changes (such as opening a window) and thermal inertia time constant The impact mechanisms of changes (such as insulation degradation) on the system are different and require targeted updates; therefore, this step analyzes the statistical characteristics of temperature control response deviation to distinguish different types of mismatch and calculates reasonable parameter update amounts accordingly.

[0061] Set the persistence threshold to 0.7 for users. In the Response deviation persistence index for each control cycle :

[0062] (1) When At that time, the heat transfer coefficient was not considered. and thermal inertia time constant Update.

[0063] (2) When When this occurs, it is determined to be a model mismatch, triggering the parameter update mechanism and adjusting the heat transfer coefficient. and thermal inertia time constant Perform the update; avoid unnecessary frequent updates by setting reasonable trigger conditions; refer to the flowchart of the update process in step S3. Figure 3 This includes steps S301 to S303, specifically:

[0064] S301: Calculate the thermal characteristic change pattern recognition index of the user in the current control cycle based on the mean and fluctuation characteristics of the user's temperature control response deviation rate sequence and its first-order difference sequence in the current control cycle.

[0065] Specifically, for users In the The temperature control response deviation rate sequence and its first-order difference sequence for each control cycle are analyzed. Based on the mean and fluctuation characteristics of the temperature control response deviation rate sequence and its first-order difference sequence, a user... In the The thermal characteristic change pattern recognition index for each control cycle.

[0066] user In the The formula for calculating the thermal characteristic change pattern recognition index for each control cycle is:

[0067] ;

[0068] In the formula, For users In the Thermal characteristic change pattern recognition index for each control cycle; For users In the The standard deviation of the temperature control response deviation rate sequence for each control cycle represents the overall degree of fluctuation. For users In the The mean of the temperature control response deviation rate sequence for each control cycle; For users In the The standard deviation of the first difference sequence of the temperature control response deviation rate sequence for each control cycle; This indicates taking the absolute value.

[0069] in, This indicates the degree to which the mean temperature control response deviation rate deviates from 1 at all times in the analysis window. It represents the degree of deviation from the standardized mean, reflecting the systematic nature of the deviation; This indicates the smoothness of fluctuations, reflecting the degree of drastic change in deviation.

[0070] This formula combines the mean and fluctuation characteristics of the deviation to construct a single index that can distinguish different types of mismatch, namely the thermal characteristic change pattern recognition index: heat transfer coefficient. Sudden changes typically cause a step change in the temperature control response deviation rate over time, manifested as a significant deviation of the mean from 1 but with relatively small fluctuations; thermal inertia time constant Changes typically cause the temperature control response deviation rate to drift slowly over time, exhibiting a mean close to 1 but with significant fluctuations; therefore, the thermal characteristic change pattern recognition index... A larger value indicates that the average temperature control response deviation rate deviates significantly from 1 and fluctuates less, which is judged to be the heat transfer coefficient. mutation; The smaller the value, the closer the average temperature control response deviation rate is to 1, but the greater the fluctuation, which is judged to be the thermal inertia time constant. Changes; the construction of this index achieves a linear mapping from temperature control response deviation characteristics to model mismatch types.

[0071] It should be noted that by introducing a thermal characteristic change pattern recognition index, a linear mapping from deviation characteristics to mismatch types is achieved.

[0072] S302: Based on the user's temperature control response deviation rate and thermal characteristic change pattern recognition index in the current control cycle, calculate the update amount of the user's heat transfer coefficient and thermal inertia time constant in the current control cycle.

[0073] It should be noted that the user's thermal characteristic change pattern recognition index during the control cycle It can be used to distinguish heat transfer coefficients and thermal inertia time constant The relative change contribution is thus determined by the S-shaped function. Therefore, converting the thermal characteristic change pattern recognition index into pattern recognition weights ensures that the update focus matches the mismatch type.

[0074] Specifically, users In the The formula for calculating the pattern recognition weight for each control cycle is:

[0075] ;

[0076] In the formula, For users In the Pattern recognition weights for each control cycle; For users In the Thermal characteristic change pattern recognition index for each control cycle; The shape parameter of the sigmoid function is fixed at 5. The center point of the sigmoid function is fixed at 1.

[0077] Among them, the sigmoid function is Smooth transition nearby, ensuring when When it approaches 1, The process smoothly transitions from 0 to 1, avoiding abrupt changes during the update process. By using an S-shaped function to convert the thermal characteristic change pattern recognition index into pattern recognition weights, the smoothness of the update process can be ensured, preventing system oscillations.

[0078] Furthermore, based on users In the The mean and pattern recognition weights of the temperature control response deviation rate sequence for each control cycle are used to calculate the user's... In the The update amount of heat transfer coefficient and thermal inertia time constant for each control cycle.

[0079] user In the Heat transfer coefficient per control cycle and thermal inertia time constant The formula for calculating the update amount is:

[0080] ;

[0081] ;

[0082] In the formula, For users In the Heat transfer coefficient per control cycle Update volume; For users In the Thermal inertia time constant for each control cycle Update volume; For users In the The mean of the temperature control response deviation rate sequence for each control cycle; For users In the The thermal inertia time constant of each control cycle; For users In the Heat transfer coefficient for each control cycle; For users In the Thermal characteristic change pattern recognition index for each control cycle; For users In the Pattern recognition weights for each control cycle.

[0083] Due to the temperature control response deviation rate A value less than 1 indicates that the actual temperature rise is less than predicted, possibly due to a lower heat transfer coefficient. This is caused by enlargement, therefore, through This represents the proportion of actual heat transfer capacity relative to model assumptions, and is then used to calculate user... heat transfer coefficient Update volume ,and The smaller the value, the more users... heat transfer coefficient Update volume The larger the deviation rate; due to the temperature control response deviation rate A value greater than 1 indicates that the actual temperature rise is greater than predicted, possibly due to an additional heat source. Therefore, by... This represents the proportion of actual thermal inertia relative to the model assumptions, and is then used to calculate user... thermal inertia time constant Update volume ,and The smaller the value, the more users... thermal inertia time constant Update volume The smaller.

[0084] Among them, when When it is large, Approximately equal to 1, primarily updating the heat transfer coefficient. ;when When smaller, Approximately equal to 0, primarily updating the thermal inertia time constant. ;when When the value approaches 1, the heat transfer coefficient is updated simultaneously. and thermal inertia time constant And the proportion is from Sure.

[0085] S303: Based on the user's updated heat transfer coefficient and thermal inertia time constant in the current control cycle, update the user's heat transfer coefficient and thermal inertia time constant in the previous control cycle, and obtain the updated heat transfer coefficient and thermal inertia time constant for the current control cycle.

[0086] Specifically, based on the update amount of the user's heat transfer coefficient and thermal inertia time constant in the current control cycle, the user's heat transfer coefficient and thermal inertia time constant in the previous control cycle are updated to obtain the user's updated heat transfer coefficient and thermal inertia time constant in the current control cycle.

[0087] user In the The formulas for calculating the updated heat transfer coefficient and thermal inertia time constant for each control cycle are as follows:

[0088] ;

[0089] ;

[0090] In the formula, For users In the The updated heat transfer coefficient for each control cycle; For users In the The updated thermal inertia time constant for each control cycle; For users In the Heat transfer coefficient per control cycle Update volume; For users In the Thermal inertia time constant for each control cycle Update volume; This is the learning rate.

[0091] When the learning rate is too small, the update is too slow; when the learning rate is too large, it is easy to cause system oscillation. Therefore, the learning rate is set to 0.3.

[0092] It should be noted that, through the learning rate By controlling the proportion of updates, parameters can be updated to avoid system instability caused by a large-scale update at once, thus achieving a balance between update speed and system stability.

[0093] S4: Take the valve opening of the current control cycle as the object to be solved: predict the room temperature of the next control cycle based on the updated heat transfer coefficient and thermal inertia time constant, construct the objective function of the multi-objective convex optimization problem based on the predicted room temperature and valve adjustment range of all users in the unit, solve the optimal valve opening command for each user in the unit, and adjust the valve opening.

[0094] Specifically, the valve opening degree of the user in the current control cycle is taken as the object to be solved. Based on the updated heat transfer coefficient, thermal inertia time constant, actual room temperature and valve opening degree of the user in the current control cycle, the room temperature of the user in the next control cycle is predicted based on the room temperature prediction model in step S201.

[0095] Furthermore, taking the valve opening degree of the user in the current control cycle as the object to be solved, and based on the predicted room temperature and valve adjustment range of all users in the unit in the next control cycle, an objective function for a multi-objective convex optimization problem is constructed, which can balance thermal balance and stability. Here, the valve adjustment range is equal to the difference between the valve opening degree of the user in the current control cycle and the previous control cycle. The objective function is solved using the interior point method to obtain the optimal valve opening degree sequence, which includes the optimal valve opening degree command for each user in the unit.

[0096] The objective function is expressed as follows:

[0097] ;

[0098] in, , For users In the Valve opening degree per control cycle The number of all users within the unit; In the first The maximum temperature difference within the unit during each control cycle is equal to the temperature difference between all users within the unit during the first control cycle. The difference between the maximum and minimum predicted room temperature for each control cycle; For users In the The valve adjustment range for each control cycle, and ; This indicates that all users within the unit are in the [number]th [period]. The sum of valve adjustment amplitudes for each control cycle; This indicates taking the absolute value.

[0099] Finally, based on the optimal valve opening command for each user within the unit contained in the optimal valve opening sequence, the valve opening of each user within the unit is adjusted.

[0100] This invention also discloses an intelligent control system for a heating system, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement an intelligent control method for a heating system according to the present invention.

[0101] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

Claims

1. A method for intelligent control of a heating system, characterized in that, include: During each control cycle, real-time data from the user is collected, including actual room temperature and valve opening. Based on the user's heat transfer coefficient, thermal inertia time constant, actual room temperature, and valve opening in the previous control cycle, the room temperature of the current control cycle is predicted; the ratio of the difference between the actual room temperature and the difference between the predicted room temperature in adjacent control cycles is used as the temperature control response deviation rate for each control cycle; based on the time-domain characteristics of the temperature control response deviation rate sequence in the analysis window of the current control cycle, the response deviation persistence index of the current control cycle is calculated. When the response deviation persistence index is greater than the persistence threshold: Based on the mean and fluctuation characteristics of the temperature control response deviation rate sequence and its first-order difference sequence, calculate the thermal characteristic change pattern recognition index of the current control cycle. Combined with the temperature control response deviation rate, calculate the update amount of the heat transfer coefficient and thermal inertia time constant of the current control cycle, and update the heat transfer coefficient and thermal inertia time constant. The valve opening of the user in the current control cycle is taken as the object to be solved: based on the updated heat transfer coefficient, thermal inertia time constant, actual room temperature and valve opening of the user in the current control cycle, the room temperature of the next control cycle is predicted; based on the predicted room temperature and valve adjustment range of all users in the unit in the next control cycle, the objective function of the multi-objective convex optimization problem is constructed; the objective function is solved to obtain the optimal valve opening command for each user in the unit, and the valve opening is adjusted.

2. The intelligent control method for a heating system according to claim 1, characterized in that, The method of predicting the room temperature for the current control cycle based on the user's heat transfer coefficient, thermal inertia time constant, actual room temperature, and valve opening in the previous control cycle includes: ； In the formula, For users In the Predicted room temperature for each control cycle; For users In the The actual room temperature for each control cycle; To control the duration of the cycle; For users In the The thermal inertia time constant of each control cycle; To set the temperature; For users In the Heat transfer coefficient for each control cycle; For users In the The flow rate of each control cycle, and , For users In the Valve opening degree per control cycle This is the proportionality coefficient.

3. The intelligent control method for a heating system according to claim 1, characterized in that, The method for obtaining the analysis window of the current control cycle is as follows: The first The control cycle to the first The control cycle constitutes the first control cycle. Analysis window for each control cycle, To analyze the length of the window; the user In the The temperature control response deviation rate of all temperature control responses in the analysis window of each control cycle constitutes the user's... In the Temperature control response deviation rate sequence for each control cycle.

4. The intelligent control method for a heating system according to claim 1, characterized in that, The step of calculating the response deviation persistence index for the current control cycle based on the time-domain characteristics of the temperature control response deviation rate sequence within the user's analysis window of the current control cycle includes: ； In the formula, For users In the The duration index of response deviation for each control cycle; The length of the analysis window; For users exist Temperature control response deviation rate per control cycle; This is an indicator function; it returns 1 when the condition is met and 0 otherwise. This is the deviation threshold; This indicates taking the absolute value.

5. The intelligent control method for a heating system according to claim 1, characterized in that, The calculation of the thermal characteristic change pattern recognition index for the current control cycle based on the mean and fluctuation characteristics of the temperature control response deviation rate sequence and its first-order difference sequence includes: ； In the formula, For users In the Thermal characteristic change pattern recognition index for each control cycle; For users In the Standard deviation of the temperature control response deviation rate sequence for each control cycle; For users In the The mean of the temperature control response deviation rate sequence for each control cycle; For users In the The standard deviation of the first difference sequence of the temperature control response deviation rate sequence for each control cycle; This indicates taking the absolute value.

6. The intelligent control method for a heating system according to claim 1, characterized in that, The calculation of the update amount of the heat transfer coefficient and thermal inertia time constant for the current control cycle, based on the temperature control response deviation rate, includes: ； ； In the formula, For users In the Heat transfer coefficient per control cycle Update volume; For users In the Thermal inertia time constant for each control cycle Update volume; For users In the The mean of the temperature control response deviation rate sequence for each control cycle; For users In the The thermal inertia time constant of each control cycle; For users In the Heat transfer coefficient for each control cycle; For users In the Thermal characteristic change pattern recognition index for each control cycle; For users In the Pattern recognition weights for each control cycle.

7. The intelligent control method for a heating system according to claim 6, characterized in that, The user In the The formula for calculating the pattern recognition weight for each control cycle is: ； In the formula, For users In the Pattern recognition weights for each control cycle; For users In the Thermal characteristic change pattern recognition index for each control cycle; The shape parameter of the sigmoid function; It is the center point of the sigmoid function.

8. The intelligent control method for a heating system according to claim 1, characterized in that, The updated heat transfer coefficient and thermal inertia time constant include: ； ； In the formula, For users In the The updated heat transfer coefficient for each control cycle; For users In the The updated thermal inertia time constant for each control cycle; For users In the Heat transfer coefficient per control cycle Update volume; For users In the Thermal inertia time constant for each control cycle Update volume; This is the learning rate.

9. The intelligent control method for a heating system according to claim 1, characterized in that, The expression for the objective function is: ； in, , For users In the Valve opening degree per control cycle The number of all users within the unit; In the first The maximum temperature difference within the unit during each control cycle is equal to the temperature difference between all users within the unit during the first control cycle. The difference between the maximum and minimum predicted room temperature for each control cycle; For users In the The valve adjustment range for each control cycle, and ; This indicates taking the absolute value.

10. An intelligent control system for a heating system, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement a heating system intelligent control method according to any one of claims 1-9.

Images