An air conditioning system dynamic load demand prediction method, device and electronic equipment

Abstract

The application provides a kind of air conditioning system dynamic load demand prediction method, device and electronic equipment, it is related to technical field.The application obtains the monitoring data such as building thermal inertia data, circulating water data, internal disturbance data and external disturbance data in historical period in air conditioning system, carries out circulating water temperature identification, calculates the lag factor of the air conditioning system, realizes the calculation of the water temperature and cumulative flow required by air conditioning system when user room temperature changes from first room temperature to second room temperature within a set time, comprehensively considers the heat transfer performance and response delay factor when user end temperature changes.Thereby, the dynamic load prediction model trained based on influence characteristics and lag factor is used to predict the dynamic load demand of air conditioning system, comprehensively considers the heat transfer performance and response delay factor when air conditioning system end temperature changes, improves the accuracy of air conditioning system dynamic load demand prediction, and improves the control precision of air conditioning system end.

Description

Technical Field

[0001] This invention relates to the field of air conditioning technology, and in particular to a method, apparatus and electronic equipment for predicting dynamic load demand of air conditioning systems. Background Technology

[0002] In energy-saving optimization control of air conditioning systems, load forecasting is used to characterize the energy demand of end-users. Using this forecast as the control target, energy stations can achieve proactive regulation and control, thus demonstrating superior terminal control results, stability, and energy-saving effects compared to traditional station-level local automation. Therefore, the accuracy of load forecasting based on the supply-demand matching concept becomes a key factor affecting the final effectiveness of this technology.

[0003] In current mainstream technical solutions, the quantification of terminal load is typically represented by the cooling (heating) capacity supplied by the data center. Specifically, the load can be calculated based on historical operating data of the data center and the load calculation formula (Q = cmΔt). By combining relevant influencing factors, the load can be predicted and used as the input for optimized control of the data center.

[0004] However, this load forecasting method ignores the heat exchange performance and response delay factors at the user end. For example, if the heat exchange performance at the user end is normal, the air conditioning system operates according to the calculated load, and the user end temperature is within the normal range. When the heat exchange performance at the user end is good, the air conditioner only needs a small amount of cooling capacity to achieve user end temperature control. If the air conditioning system operates according to the calculated load, the user end temperature will be lower than the normal range, resulting in reduced control accuracy of the air conditioning system.

[0005] For example, due to the long piping of the air conditioning system, the response delay at the user end has a significant impact. If the response time at the user end is one hour, while the control interval of the air conditioning system is half an hour, it may result in the user end not responding after the air conditioning system adjusts, and the air conditioning system entering the next adjustment process, affecting the control accuracy of the air conditioning system terminal. Summary of the Invention

[0006] This invention provides a method, apparatus, and electronic device for predicting dynamic load demand in air conditioning systems, which can improve the accuracy of predicting dynamic load demand in air conditioning systems and improve the control precision of air conditioning system terminals.

[0007] In a first aspect, the present invention provides a method for predicting dynamic load demand of an air conditioning system. The method includes: acquiring monitoring data from the air conditioning system over a historical period, including building thermal inertia data, circulating water data, internal disturbance data, and external disturbance data; identifying the circulating water temperature based on the monitoring data from the air conditioning system over a historical period, and calculating the lag factor of the air conditioning system; the lag factor being the required water supply temperature and cumulative flow rate of the air conditioning system when the user's room temperature changes from a first room temperature to a second room temperature within a set time period; performing influence factor analysis based on the monitoring data to obtain the influence characteristics of the air conditioning system; the influence characteristics being features in the monitoring data that affect the lag factor; training a machine model based on the influence characteristics and the lag factor to obtain a dynamic load prediction model; the dynamic load prediction model taking the lag factor as output and the influence characteristics as input; and predicting the dynamic load demand of the air conditioning system based on the dynamic load prediction model.

[0008] In one possible implementation, the dynamic load demand of the air conditioning system is predicted based on a dynamic load prediction model, including: acquiring real-time monitoring data of the air conditioning system and the target room temperature; generating n-dimensional input features based on the real-time monitoring data and the target room temperature; inputting the n-dimensional input features into the dynamic load prediction model to obtain the water supply temperature and lag factor; and determining the water supply flow rate and control duration of the air conditioning system based on the water supply temperature and lag factor.

[0009] In one possible implementation, based on monitoring data from the air conditioning system over a historical period, circulating water temperature is identified, and the hysteresis factor of the air conditioning system is calculated. This includes: cleaning the monitoring data to obtain cleaned data; performing pattern recognition based on the cleaned data to determine the operating data of the air conditioning system under stable operating conditions; and performing circulating water temperature identification and calculating the hysteresis factor based on the operating data of the air conditioning system under stable operating conditions.

[0010] In one possible implementation, based on the operating data of the air conditioning system under stable operating conditions, circulating water temperature is identified, and a lag factor is calculated. This includes: identifying multiple abrupt change nodes in the supply water temperature over a historical period based on the stable operating data; identifying multiple abrupt change nodes in the return water temperature over a historical period based on the stable operating data; matching the multiple abrupt change nodes in the supply water temperature and the multiple abrupt change nodes in the return water temperature to obtain multiple pairs of supply and return water temperature abrupt change nodes; calculating the time interval between each pair of supply and return water temperature abrupt change nodes; determining the cumulative flow of each pair of supply and return water temperature abrupt change nodes based on the instantaneous flow rate of each pair of supply and return water temperature abrupt change nodes and the time interval between each pair of supply and return water temperature abrupt change nodes; querying the stable operating data to determine the user's first room temperature corresponding to the supply water abrupt change node and the user's second room temperature corresponding to the return water abrupt change node in each pair of supply and return water temperature abrupt change nodes; and determining the lag factor based on the user's first room temperature and second room temperature, as well as the cumulative flow rate and supply water temperature of each pair of supply and return water temperature abrupt change nodes.

[0011] In one possible implementation, an impact factor analysis is performed based on monitoring data from the air conditioning system to obtain the impact characteristics of the air conditioning system. This includes: segmenting the monitoring data in the air conditioning system based on multiple pairs of supply and return water temperature abrupt change nodes to obtain monitoring data for multiple time periods within multiple response intervals; performing an impact factor analysis based on the monitoring data for multiple time periods within multiple response intervals to obtain the impact characteristics of the air conditioning system; the impact characteristics include building thermal inertia characteristics, circulating water characteristics, internal disturbance characteristics, and external disturbance characteristics; building thermal inertia characteristics include continuous cooling duration, outdoor temperature during cooling, and indoor temperature during cooling; circulating water characteristics include instantaneous circulating water flow rate, cumulative circulating water flow rate, supply water temperature, and return water temperature; internal disturbance characteristics include room temperature data and indoor occupant activity intensity; external disturbance characteristics include outdoor temperature, outdoor wind speed, and irradiance.

[0012] In one possible implementation, a dynamic load prediction model is obtained by training a machine model based on the impact features and lag factors, including: generating training samples with the impact features as input features and the lag factors as output features; training the machine model based on the training samples to obtain an initial model; removing the input features from the training samples with replacement, performing feature ablation, retraining the machine model, optimizing the initial model, and obtaining the dynamic load prediction model.

[0013] In one possible implementation, the input features in the training samples are removed with replacement, feature ablation is performed, the machine model is retrained, and the initial model is optimized to obtain a dynamic load prediction model. This includes: Step 1: Calculate the determination coefficient of the initial model and initialize it as the global optimum; Step 2: Initialize the number of features removed from the input features to 1; Step 3: Based on the number of features removed, remove features that have been removed from the input features to obtain multiple optimized input features; Step 4: Based on each optimized input feature and a lag factor, train the machine model to obtain multiple optimized models; Step 5: Calculate the determination coefficient of each optimized model, and the maximum determination coefficient among the multiple optimized models; Step 6: If the maximum determination coefficient among the multiple optimized models is greater than the determination coefficient of the global optimum, then the optimized model corresponding to the maximum determination coefficient is determined as the global optimum; if the maximum determination coefficient among the multiple optimized models is less than or equal to the determination coefficient of the global optimum, then it remains unchanged; Step 7: Increase the number of features removed by 1, repeat steps 3 to 7 until the number of features removed is greater than the number of features in the input features, exit the iteration process, and execute step 8; Step 8: Determine the global optimum as the dynamic load prediction model.

[0014] In one possible implementation, the dynamic load demand of the air conditioning system is predicted based on a dynamic load forecasting model. This process includes: dividing each feature into intervals based on a set interval in the training samples to obtain multiple value intervals for each feature; traversing each training sample to determine the value interval corresponding to each training sample; determining the feature coverage of the training samples based on the value interval corresponding to each training sample and the multiple value intervals for each feature; and determining the evaluation result of the dynamic load forecasting model based on the feature coverage of the training samples. The evaluation result includes passing and failing the evaluation.

[0015] Secondly, embodiments of the present invention provide a dynamic load demand prediction device for an air conditioning system. The device includes: a communication module for acquiring monitoring data from the air conditioning system over a historical period, including building thermal inertia data, circulating water data, internal disturbance data, and external disturbance data; a processing module for identifying the circulating water temperature based on the monitoring data from the air conditioning system over a historical period and calculating the lag factor of the air conditioning system; the lag factor being the required water supply temperature and cumulative flow rate of the air conditioning system when the user's room temperature changes from a first room temperature to a second room temperature within a set time period; performing influence factor analysis based on the monitoring data to obtain the influence characteristics of the air conditioning system; the influence characteristics being features in the monitoring data that affect the lag factor; training a machine model based on the influence characteristics and the lag factor to obtain a dynamic load prediction model; the dynamic load prediction model outputting the lag factor and inputting the influence characteristics; and predicting the dynamic load demand of the air conditioning system based on the dynamic load prediction model.

[0016] In one possible implementation, the communication module is specifically used to acquire real-time monitoring data and target room temperature of the air conditioning system; the processing module is specifically used to generate n-dimensional input features based on the real-time monitoring data and target room temperature; input the n-dimensional input features into the dynamic load prediction model to obtain the water supply temperature and lag factor; and determine the water supply flow rate and control duration of the air conditioning system based on the water supply temperature and lag factor.

[0017] In one possible implementation, the processing module is specifically used to clean the monitoring data to obtain cleaned data; based on the cleaned data, to perform pattern recognition to determine the operating data of the air conditioning system under stable operating conditions; and based on the operating data of the air conditioning system under stable operating conditions, to identify the circulating water temperature and calculate the hysteresis factor.

[0018] In one possible implementation, the processing module is specifically used to: identify multiple abrupt changes in supply water temperature over a historical period based on stable operating data; identify multiple abrupt changes in return water temperature over a historical period based on stable operating data; match the multiple abrupt changes in supply water temperature and return water temperature to obtain multiple pairs of abrupt changes in supply and return water temperature; calculate the time interval between each pair of abrupt changes in supply and return water temperature; determine the cumulative flow of each pair of abrupt changes in supply and return water temperature based on the instantaneous flow rate of each pair of abrupt changes in supply and return water temperature and the time interval between each pair of abrupt changes in supply and return water temperature; query the stable operating data to determine the user's first room temperature corresponding to the supply water abrupt change in each pair of abrupt changes in supply water temperature and the user's second room temperature corresponding to the return water abrupt change in supply water temperature; and determine the lag factor based on the user's first room temperature and second room temperature, as well as the cumulative flow rate and supply water temperature of each pair of abrupt changes in supply and return water temperature.

[0019] In one possible implementation, the processing module is specifically used to segment the monitoring data of the air conditioning system based on multiple pairs of supply and return water temperature abrupt change nodes, obtaining monitoring data for multiple time periods in multiple response intervals; based on the monitoring data for multiple time periods in multiple response intervals, it performs influencing factor analysis to obtain the influence characteristics of the air conditioning system; the influence characteristics include building thermal inertia characteristics, circulating water characteristics, internal disturbance characteristics, and external disturbance characteristics; the building thermal inertia characteristics include continuous cooling duration, outdoor temperature during cooling, and indoor temperature during cooling; the circulating water characteristics include instantaneous circulating water flow rate, cumulative circulating water flow rate, supply water temperature, and return water temperature; the internal disturbance characteristics include room temperature data and indoor occupant activity intensity; the external disturbance characteristics include outdoor temperature, outdoor wind speed, and irradiance.

[0020] In one possible implementation, the processing module is specifically used to generate training samples with impact features as input features and lag factors as output features; to train a machine model based on the training samples to obtain an initial model; to remove the input features from the training samples with replacement, to perform feature ablation, to retrain the machine model, to optimize the initial model, and to obtain a dynamic load prediction model.

[0021] In one possible implementation, the processing module specifically performs the following steps: Step 1: Calculate the determination coefficient of the initial model and initialize the initial model as the global optimum; Step 2: Initialize the feature removal count in the input features to 1; Step 3: Based on the feature removal count, remove features from the input features to obtain multiple optimized input features; Step 4: Based on each optimized input feature and lag factor, train a machine model to obtain multiple optimized models; Step 5: Calculate the determination coefficient of each optimized model, and the maximum determination coefficient among the multiple optimized models; Step 6: If the maximum determination coefficient among the multiple optimized models is greater than the determination coefficient of the global optimum, then determine the optimized model corresponding to the maximum determination coefficient as the global optimum; if the maximum determination coefficient among the multiple optimized models is less than or equal to the determination coefficient of the global optimum, then keep it unchanged; Step 7: Increase the feature removal count by 1, repeat steps 3 to 7 until the feature removal count is greater than the number of features in the input features, exit the iteration process, and execute step 8; Step 8: Determine the global optimum as the dynamic load prediction model.

[0022] In one possible implementation, the processing module is further configured to divide each feature into intervals based on a set interval in the training samples, thereby obtaining multiple value intervals for each feature; traverse each training sample to determine the value interval corresponding to each training sample; determine the feature coverage of the training samples based on the value interval corresponding to each training sample and the multiple value intervals for each feature; and determine the evaluation result of the dynamic load prediction model based on the feature coverage of the training samples, wherein the evaluation result includes evaluation passed and evaluation failed.

[0023] Thirdly, embodiments of the present invention provide an electronic device including a memory and a processor. The memory stores a computer program, and the processor is configured to call and run the computer program stored in the memory to perform the steps of the method as described in the first aspect and any possible implementation thereof.

[0024] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, characterized in that, when the computer program is executed by a processor, it implements the steps of the method as described in the first aspect and any possible implementation thereof.

[0025] This invention provides a method, apparatus, and electronic device for predicting dynamic load demand in an air conditioning system. By acquiring historical monitoring data such as building thermal inertia data, circulating water data, internal disturbance data, and external disturbance data of the air conditioning system, this invention identifies the circulating water temperature and calculates the lag factor of the air conditioning system. This enables the calculation of the required supply water temperature and cumulative flow rate of the air conditioning system when the user's room temperature changes from a first room temperature to a second room temperature within a set time period, comprehensively considering the heat exchange performance and response delay factors when the user's temperature changes. Subsequently, the influence characteristics of the lag factor in the monitoring data are used to train a machine model, resulting in a dynamic load prediction model. Therefore, based on the dynamic load prediction model, the dynamic load demand of the air conditioning system is predicted, comprehensively considering the heat exchange performance and response delay factors when the temperature changes at the air conditioning system's terminals, thus improving the accuracy of dynamic load demand prediction and the control precision of the air conditioning system's terminals. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a flowchart illustrating a dynamic load demand forecasting method for an air conditioning system provided in an embodiment of the present invention;

[0028] Figure 2 This is a schematic diagram of a sudden temperature change node in the supply and return water provided in an embodiment of the present invention;

[0029] Figure 3 This is a schematic diagram illustrating the influence characteristics of an air conditioning system provided in an embodiment of the present invention;

[0030] Figure 4 This is a schematic diagram of the structure of a dynamic load demand prediction device for an air conditioning system provided in an embodiment of the present invention;

[0031] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0032] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0033] In the description of this invention, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The term "and / or" in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Furthermore, "at least one" and "more than one" refer to two or more. The terms "first," "second," etc., do not limit the quantity or order of execution, and "first," "second," etc., do not necessarily imply differences.

[0034] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner to facilitate understanding.

[0035] Furthermore, the terms "comprising" and "having," and any variations thereof, used in the description of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or modules is not limited to the steps or modules listed, but may optionally include other steps or modules not listed, or may optionally include other steps or modules inherent to such process, method, product, or device.

[0036] To make the objectives, technical solutions, and advantages of the present invention clearer, the following description will be provided in conjunction with the accompanying drawings and specific embodiments.

[0037] Currently, load forecasting for air conditioning systems is mostly based on historical operating data from the air conditioning room, and this data is used as the input for room control.

[0038] This load forecasting method, which uses the cooling (heating) capacity of the air conditioning room as the load demand for equipment optimization control, has shortcomings. For example, it cannot quantify the time of terminal temperature response during the control process. When the terminal room temperature fluctuates, the cooling (heating) supply of the energy station needs to be adjusted to reach the target temperature. If this demand is predicted as a load quantity, the time to reach the target room temperature cannot be quantitatively constrained. Furthermore, this load forecasting method uses the cooling (heating) capacity of the air conditioning room as the load demand, ignoring the heat exchange performance of the terminals. When predicting the required load at the terminals, using this load quantity as the control target during the chiller station optimization control phase, the neglect of the terminal heat exchange performance leads to situations where the heat exchange temperature difference exceeds the boundary under certain operating conditions, causing the system's operating conditions to continuously deteriorate, with the circulating water temperature continuously decreasing or increasing. Even if this situation is considered and corresponding constraints are applied, the requirement for load supply and demand matching is still disrupted, reducing the energy-saving effect of the control.

[0039] Furthermore, this load forecasting method neglects the response delay factor when constructing the load forecasting model. The main factors contributing to load response delay include the transmission and distribution delay of the cooling medium in the pipeline network and the heat transfer delay of the terminal thermal response. In most projects, the latter can be ignored; therefore, the load response delay mentioned in this paper refers to the transmission and distribution delay. The degree of load response delay varies across different types of businesses. In ordinary public buildings, this delay is typically 3-10 minutes, while the control interval is often half an hour or one hour, so its impact on the results is minimal. However, in regional energy systems, due to the long pipeline network, the delay can be as high as one hour, making this factor non-negligible. Conventional load forecasting technologies typically ignore this delay factor, resulting in significant problems with adaptability to different projects and thus affecting the accuracy of terminal temperature control.

[0040] To solve the above technical problems, such as Figure 1 As shown, this embodiment of the invention provides a method for dynamic load demand forecasting of an air conditioning system. The method includes steps S101-S105.

[0041] S101. Obtain monitoring data from the air conditioning system during historical periods.

[0042] In this embodiment of the application, the monitoring data includes building thermal inertia data, circulating water data, internal disturbance data, and external disturbance data.

[0043] In some embodiments, building thermal inertia data include continuous cooling duration, average outdoor temperature during the cooling period, and average indoor temperature during the cooling period.

[0044] It should be noted that building thermal inertia is a property of buildings in that they exhibit differences in the time it takes for heat to be absorbed and released according to the laws of heat. When the external ambient temperature changes rapidly, the temperature inside the building will only change slightly, and this change has a certain delay.

[0045] The thermal inertia of a building is mainly affected by the materials of its cladding structure. Materials with good thermal inertia are mainly heavy building materials, such as soil, brick, concrete, and stone. Materials with poor thermal inertia are mainly lightweight building materials, such as wood, foam, and plastics.

[0046] Building thermal inertia data refers to specific thermophysical building parameters that are built using data-driven models and do not require direct acquisition. Building thermal inertia data is characterized by relevant building operation data, which can characterize factors related to building thermal inertia. Examples include continuous cooling duration, average indoor temperature during cooling periods, and average outdoor temperature during cooling periods.

[0047] In some embodiments, the circulating water data includes instantaneous circulating water flow rate, cumulative circulating water flow rate, supply water temperature, and return water temperature.

[0048] In some embodiments, the disturbance data includes the current average room temperature and the current indoor occupant activity intensity.

[0049] In some embodiments, external disturbance data include outdoor temperature, outdoor wind speed, and irradiance.

[0050] S102. Based on monitoring data of the air conditioning system during historical periods, identify the circulating water temperature and calculate the hysteresis factor of the air conditioning system.

[0051] In this embodiment of the application, the hysteresis factor is the water supply temperature and cumulative flow required by the air conditioning system when the user's room temperature changes from a first room temperature to a second room temperature within a set time period.

[0052] It should be noted that the embodiments of the present invention aim to model the user-side demand end-to-end, that is, to predict the cumulative flow and water temperature required by the energy station of the air conditioning system when the user's room temperature changes from the first room temperature T1 to the second room temperature T2 after a set time t under the current operating conditions.

[0053] As one possible implementation, step S102 can be specifically implemented as steps S1021-S1023.

[0054] S1021. Perform data cleaning on the monitoring data to obtain cleaned data.

[0055] For example, embodiments of the present invention can perform anomaly detection on each data in the monitoring data according to anomaly rules, identify abnormal data in the monitoring data, and remove and fill abnormal data.

[0056] The anomaly rules include upper and lower limit constraints and rate of change constraints. For example, monitored data should be less than or equal to the upper limit value and greater than or equal to the lower limit value. The rate of change of monitored data should be less than or equal to the upper limit of the rate of change.

[0057] The upper limit, lower limit, and upper limit of change rate for various monitoring data are shown in Table 1.

[0058] Table 1

[0059]

[0060]

[0061] For example, for abnormal data, embodiments of the present invention can fill the abnormal data by means of linear interpolation.

[0062] It should be noted that, in order to ensure the validity of the monitoring data, the number of consecutive abnormal data in each monitoring data should be less than 3.

[0063] S1022. Based on the cleaned data, perform pattern recognition to determine the operating data of the air conditioning system under stable operating conditions.

[0064] In some embodiments, the operating modes of the air conditioning system include cooling mode, heating mode, and alternating cooling and heating mode.

[0065] In some embodiments, if the air conditioning system operates in cooling mode for a duration longer than a first duration, it is determined that the air conditioning system is operating continuously and is in a stable operating condition. If the air conditioning system operates in heating mode for a duration longer than the first duration, it is determined that the air conditioning system is operating continuously and is in a stable operating condition.

[0066] It should be noted that this invention involves the identification of both cooling and heating operation modes. To ensure compatibility with both cooling and heating modes of the air conditioning system, the monitoring data of the air conditioning system may simultaneously include both cooling and heating data. Therefore, this invention can perform pattern recognition and extraction on historical data for the current required mode based on the system's water supply temperature (extraction is performed at the daily data granularity).

[0067] In addition, embodiments of the present invention can also use the water pump operating status as the basis for calibrating whether the system is operating, and divide the cleaning data obtained in step S1021 into multiple operating segments to obtain the operating data of the air conditioning system under stable operating conditions, that is, multiple continuous operating data segments in the cooling (heating) mode.

[0068] It should be noted that for air conditioning systems, the system is in an unstable operating condition when it is turned on and off. The data required by the embodiments of the present invention is data under stable operating conditions. Therefore, in the process of segmenting the cleaned data, data segments after the third period of system operation and data segments before the fourth period of system shutdown can be cut.

[0069] S1023. Based on the operating data of the air conditioning system under stable operating conditions, identify the circulating water temperature and calculate the lag factor.

[0070] It should be noted that, in order to be applicable to both low and high thermal response delay systems, this invention requires the identification of hysteresis factors affecting the delay. Theoretically, this delay factor can be calculated using data such as pipe diameter and pipe length, but this calculation method undoubtedly increases labor costs and the difficulty of obtaining project data. Furthermore, the liquid level in the pipeline network also has an impact; therefore, the algorithm for conventional hysteresis factors is quite complex.

[0071] This invention first identifies changes in the supply water temperature. Based on the operating data of the air conditioning system under stable conditions, and using abrupt changes in the main supply water temperature as a marker, the location of these changes is identified, thus obtaining the supply water temperature abrupt change nodes. Next, changes in the return water temperature are identified. Based on the identified locations of these supply water temperature changes, the response points of the main return water temperature are identified, thus obtaining the return water temperature abrupt change nodes. Finally, the instantaneous flow rates within the response interval between the supply and return water temperature abrupt change nodes are accumulated to obtain the hysteresis factor.

[0072] For example, step S1023 can be specifically implemented as steps A1-A7.

[0073] A1. Based on stable operating data, identify multiple abrupt changes in water supply temperature over historical periods.

[0074] In some embodiments, the abrupt change node of the water supply temperature can be represented as Tsi, where i = 1, 2, 3, ..., n.

[0075] A2. Based on stable operating data, identify multiple abrupt changes in return water temperature over historical periods.

[0076] In some embodiments, the abrupt change node of the water supply temperature can be represented as Tri, where i = 1, 2, 3, ..., n.

[0077] A3. Match multiple abrupt change nodes of supply water temperature and multiple abrupt change nodes of return water temperature to obtain multiple pairs of abrupt change nodes of supply and return water temperature.

[0078] A4. Calculate the time interval between each pair of supply and return water temperature abrupt change nodes.

[0079] A5. Based on the instantaneous flow rate of each pair of supply and return water temperature change nodes and the time interval of each pair of supply and return water temperature change nodes, determine the cumulative flow rate of each pair of supply and return water temperature change nodes.

[0080] For example, in embodiments of the present invention, the cumulative flow can be calculated according to the following formula.

[0081]

[0082] Where Vi is the cumulative flow rate of the i-th pair of supply and return water temperature change nodes; Qi is the instantaneous flow rate of the i-th pair of supply and return water temperature change nodes, in cubic meters per hour (m3 / h); N is the time interval between the i-th pair of supply and return water temperature change nodes, in seconds (s).

[0083] A6. Query the operating data under stable conditions to determine the first room temperature of the user corresponding to the supply water temperature change node and the second room temperature of the user corresponding to the return water temperature change node in each pair of supply and return water temperature change nodes.

[0084] A7. Based on the user's first room temperature and second room temperature, as well as the cumulative flow and supply water temperature at each pair of supply and return water temperature abrupt change nodes, determine the lag factor.

[0085] For example, embodiments of the present invention can calculate the lag factor based on the following formula.

[0086]

[0087] Where S is the lag factor, Vi is the cumulative flow of the i-th pair of supply and return water temperature abrupt change nodes, and n is the number of pairs of supply and return water temperature abrupt change nodes.

[0088] For example, Figure 2 This is a schematic diagram of a sudden temperature change node in the supply and return water provided in an embodiment of the present invention. Figure 2 The time interval between the supply water temperature mutation node and the return water temperature mutation node is the time interval between the supply and return water temperature mutation nodes.

[0089] S103. Based on the monitoring data of the air conditioning system, conduct an impact factor analysis to obtain the impact characteristics of the air conditioning system.

[0090] In this embodiment of the application, the influencing feature is the feature in the monitoring data that has an impact on the lag factor.

[0091] As one possible implementation, step S103 can be specifically implemented as steps S1031-S1032.

[0092] S1031. Based on multiple pairs of supply and return water temperature change nodes, the monitoring data in the air conditioning system is segmented to obtain monitoring data for multiple time periods in multiple response intervals.

[0093] S1032. Based on monitoring data from multiple time periods within multiple response intervals, conduct influencing factor analysis to obtain the influence characteristics of the air conditioning system.

[0094] It should be noted that the embodiments of the present invention can be based on first principles to analyze the room temperature response factors and obtain the basic influencing factors, that is, the influencing characteristics.

[0095] In some embodiments, such as Figure 3 As shown, the influencing characteristics include building thermal inertia characteristics, circulating water characteristics, internal disturbance characteristics, and external disturbance characteristics.

[0096] For example, building thermal inertia characteristics include continuous cooling duration, outdoor temperature during cooling, and indoor temperature during cooling.

[0097] For example, circulating water characteristics include instantaneous circulating water flow rate, cumulative circulating water flow rate, supply water temperature, and return water temperature.

[0098] For example, internal disturbance characteristics include room temperature data and the intensity of indoor human activity.

[0099] For example, external disturbance characteristics include outdoor temperature, outdoor wind speed, and irradiance.

[0100] S104. Based on the impact characteristics and lag factors, machine learning model training is performed to obtain the dynamic load prediction model.

[0101] In this embodiment of the application, the dynamic load forecasting model uses the lag factor as the output and the influencing characteristics as the input.

[0102] As one possible implementation, step S104 can be specifically implemented as steps S1041-S1042.

[0103] S1041. Using the influence feature as the input feature and the lag factor as the output feature, generate training samples.

[0104] S1042. Based on the training samples, train the machine model to obtain the initial model.

[0105] S1043. Remove the input features from the training samples with replacement, perform feature ablation, retrain the machine model, optimize the initial model, and obtain the dynamic load prediction model.

[0106] It should be noted that the purpose of feature ablation is mainly to remove redundant features. These redundant features not only affect the modeling accuracy but also increase the difficulty of subsequent model applications. Therefore, the model needs to be simplified.

[0107] For example, step S1043 can be implemented as steps one through eight.

[0108] Step 1: Calculate the determination coefficients of the initial model and initialize the initial model as the global optimal solution.

[0109] Step 2: Initialize the number of features removed from the input features to 1.

[0110] Step 3: Based on the number of features removed, remove the features that have been removed from the input features to obtain multiple optimized input features.

[0111] Step 4: Based on each optimized input feature and lag factor, train the machine model to obtain multiple optimized models.

[0112] Step 5: Calculate the coefficient of determination for each optimization model, and the maximum coefficient of determination among multiple optimization models.

[0113] Step 6: If the largest determination coefficient among multiple optimization models is greater than the determination coefficient of the global optimal solution, then the optimization model corresponding to the largest determination coefficient is determined as the global optimal solution; if the largest determination coefficient among multiple optimization models is less than or equal to the determination coefficient of the global optimal solution, then it remains unchanged.

[0114] Step 7: Increase the number of features removed by 1, and repeat steps 3 to 7 until the number of features removed is greater than the number of features in the input features. Exit the iteration process and proceed to step 8.

[0115] Step 8: Determine the global optimal solution as the dynamic load prediction model.

[0116] S105. Based on the dynamic load forecasting model, predict the dynamic load demand of the air conditioning system.

[0117] As one possible implementation, step S105 can be specifically implemented as steps S1051-S1054.

[0118] S1051. Obtain real-time monitoring data and target room temperature of the air conditioning system.

[0119] S1052. Based on real-time monitoring data and target room temperature, generate n-dimensional input features.

[0120] S1053. Input the n-dimensional input features into the dynamic load prediction model to obtain the water supply temperature and lag factor.

[0121] S1054. Based on the water supply temperature and hysteresis factor, determine the water supply flow rate and control duration of the air conditioning system.

[0122] This invention provides a method for predicting dynamic load demand in an air conditioning system. By acquiring historical monitoring data such as building thermal inertia data, circulating water data, internal disturbance data, and external disturbance data, the method identifies the circulating water temperature and calculates the lag factor of the air conditioning system. This enables the calculation of the required supply water temperature and cumulative flow rate of the air conditioning system when the user's room temperature changes from a first room temperature to a second room temperature within a set time period, comprehensively considering the heat exchange performance and response delay factors when the user's temperature changes. Subsequently, the method incorporates the influence characteristics of the lag factor from the monitoring data to train a machine model, resulting in a dynamic load prediction model. Based on this model, the dynamic load demand of the air conditioning system is predicted, comprehensively considering the heat exchange performance and response delay factors when the temperature changes at the air conditioning system's terminals, thus improving the accuracy of dynamic load demand prediction and the control precision of the air conditioning system's terminals.

[0123] It should be noted that the embodiments of the present invention take the temperature change at the user end as the target, realize the end-to-end room temperature change target as the terminal load demand of the air conditioning system, and improve the accuracy of dynamic load demand prediction of the air conditioning system.

[0124] Optionally, the dynamic load demand prediction method for air conditioning systems provided in this embodiment of the invention further includes steps S201-S204 before step S105.

[0125] S201. Based on the set interval of each feature in the training samples, divide each feature into intervals to obtain multiple value intervals for each feature.

[0126] For example, the features in the training samples are shown in Table 2, which gives the set intervals for each feature.

[0127] Table 2

[0128] 1 flow m3 / h (maximum value - minimum value) / 10 2 water supply temperature ℃ 1 Refrigeration conditions 3 water supply temperature ℃ 2 Heating mode 4 Indoor temperature ℃ 1 5 outdoor temperature ℃ 1 6 Number of end openings (maximum value - minimum value) / 10 7 Hour 1 8 Daily Type 1

[0129] S202. Iterate through each training sample and determine the value range corresponding to each training sample.

[0130] S203. Based on the value range corresponding to each training sample and the multiple value ranges of each feature, determine the feature coverage of the training samples.

[0131] For example, in this embodiment of the invention, when traversing training samples, if the N-dimensional input features of a training sample fall into a certain value interval, the representation value of that value interval is set to 1. The representation value of each value interval is initially 0. After traversal is completed, the ratio of the sum of the representation values ​​of each value interval to the number of value intervals is determined as the feature coverage.

[0132] S204. Based on the feature coverage of the training samples, determine the evaluation results of the dynamic load prediction model. The evaluation results include evaluation passed and evaluation failed.

[0133] For example, if the feature coverage of the training samples is greater than or equal to the set coverage, the evaluation result is determined to be passed. If the feature coverage of the training samples is less than the set coverage, the evaluation result is determined to be failed.

[0134] In this way, after training the dynamic load prediction model, the feature coverage of the training samples can be evaluated. When the feature coverage is high, it means that the dynamic load prediction model can perform dynamic load prediction for more operating conditions, thereby improving the accuracy of the dynamic load prediction model.

[0135] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0136] The following are embodiments of the apparatus of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0137] Figure 4 A schematic diagram of a dynamic load demand forecasting device for an air conditioning system provided in an embodiment of the present invention is shown. The forecasting device 300 includes a communication module 301 and a processing module 302.

[0138] The communication module 301 is used to acquire monitoring data of the air conditioning system during historical periods. The monitoring data includes building thermal inertia data, circulating water data, internal disturbance data, and external disturbance data.

[0139] The processing module 302 is used to identify the circulating water temperature based on monitoring data of the air conditioning system over a historical period and calculate the lag factor of the air conditioning system. The lag factor is the required water supply temperature and cumulative flow of the air conditioning system when the user's room temperature changes from a first room temperature to a second room temperature within a set time period. Based on the monitoring data of the air conditioning system, it performs influence factor analysis to obtain the influence characteristics of the air conditioning system. The influence characteristics are the features in the monitoring data that have an impact on the lag factor. Based on the influence characteristics and the lag factor, it performs machine model training to obtain a dynamic load prediction model. The dynamic load prediction model takes the lag factor as the output and the influence characteristics as the input. Based on the dynamic load prediction model, it predicts the dynamic load demand of the air conditioning system.

[0140] In one possible implementation, the communication module 301 is specifically used to acquire real-time monitoring data and target room temperature of the air conditioning system; the processing module 302 is specifically used to generate n-dimensional input features based on the real-time monitoring data and target room temperature; input the n-dimensional input features into the dynamic load prediction model to obtain the water supply temperature and lag factor; and determine the water supply flow rate and control duration of the air conditioning system based on the water supply temperature and lag factor.

[0141] In one possible implementation, the processing module 302 is specifically used to clean the monitoring data to obtain cleaned data; based on the cleaned data, perform pattern recognition to determine the operating data of the air conditioning system under stable operating conditions; and based on the operating data of the air conditioning system under stable operating conditions, perform circulating water temperature identification and calculate the hysteresis factor.

[0142] In one possible implementation, the processing module 302 is specifically used to: identify multiple abrupt changes in supply water temperature over a historical period based on stable operating data; identify multiple abrupt changes in return water temperature over a historical period based on stable operating data; match the multiple abrupt changes in supply water temperature and return water temperature to obtain multiple pairs of abrupt changes in supply and return water temperature; calculate the time interval between each pair of abrupt changes in supply and return water temperature; determine the cumulative flow of each pair of abrupt changes in supply and return water temperature based on the instantaneous flow rate of each pair of abrupt changes in supply and return water temperature and the time interval between each pair of abrupt changes in supply and return water temperature; query the stable operating data to determine the user's first room temperature corresponding to the supply water abrupt change in each pair of abrupt changes in supply water temperature and the user's second room temperature corresponding to the return water abrupt change in supply water temperature; and determine the lag factor based on the user's first room temperature and second room temperature, as well as the cumulative flow rate and supply water temperature of each pair of abrupt changes in supply and return water temperature.

[0143] In one possible implementation, the processing module 302 is specifically used to segment the monitoring data of the air conditioning system based on multiple pairs of supply and return water temperature change nodes to obtain monitoring data for multiple time periods in multiple response intervals; based on the monitoring data for multiple time periods in multiple response intervals, it performs influence factor analysis to obtain the influence characteristics of the air conditioning system; the influence characteristics include building thermal inertia characteristics, circulating water characteristics, internal disturbance characteristics, and external disturbance characteristics; the building thermal inertia characteristics include continuous cooling duration, outdoor temperature during cooling, and indoor temperature during cooling; the circulating water characteristics include instantaneous circulating water flow rate, cumulative circulating water flow rate, supply water temperature, and return water temperature; the internal disturbance characteristics include room temperature data and indoor occupant activity intensity; the external disturbance characteristics include outdoor temperature, outdoor wind speed, and irradiance.

[0144] In one possible implementation, the processing module 302 is specifically used to generate training samples with the impact features as input features and the lag factor as output features; to train a machine model based on the training samples to obtain an initial model; to remove the input features from the training samples with replacement, to perform feature ablation, to retrain the machine model, to optimize the initial model, and to obtain a dynamic load prediction model.

[0145] In one possible implementation, the processing module 302 is specifically used to perform the following steps: Step 1: Calculate the determination coefficient of the initial model and initialize the initial model as the global optimum; Step 2: Initialize the number of features removed from the input features to 1; Step 3: Based on the number of features removed, remove features from the input features that have been removed, obtaining multiple optimized input features; Step 4: Based on each optimized input feature and the lag factor, perform machine model training to obtain multiple optimized models; Step 5: Calculate the determination coefficient of each optimized model, and the maximum determination coefficient among the multiple optimized models; Step 6: If the maximum determination coefficient among the multiple optimized models is greater than the determination coefficient of the global optimum, then the optimized model corresponding to the maximum determination coefficient is determined as the global optimum; if the maximum determination coefficient among the multiple optimized models is less than or equal to the determination coefficient of the global optimum, then it remains unchanged; Step 7: Increase the number of features removed by 1, repeat steps 3 to 7 until the number of features removed is greater than the number of features in the input features, exit the iteration process, and execute step 8; Step 8: Determine the global optimum as the dynamic load prediction model.

[0146] In one possible implementation, the processing module is further configured to divide each feature into intervals based on a set interval in the training samples, thereby obtaining multiple value intervals for each feature; traverse each training sample to determine the value interval corresponding to each training sample; determine the feature coverage of the training samples based on the value interval corresponding to each training sample and the multiple value intervals for each feature; and determine the evaluation result of the dynamic load prediction model based on the feature coverage of the training samples, wherein the evaluation result includes evaluation passed and evaluation failed.

[0147] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. For example... Figure 5 As shown, the electronic device 400 of this embodiment includes: a processor 401, a memory 402, and a computer program 403 stored in the memory 402 and executable on the processor 401. When the processor 401 executes the computer program 403, it implements the steps in the above-described method embodiments, for example... Figure 1 The steps S101-S105 are shown. Alternatively, when the processor 401 executes the computer program 403, it implements the functions of each module / unit in the above-described device embodiments, for example... Figure 4The functions of the communication module 301 and the processing module 302 shown are illustrated.

[0148] For example, the computer program 403 can be divided into one or more modules / units, which are stored in the memory 402 and executed by the processor 401 to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program 403 in the electronic device 400. For example, the computer program 403 can be divided into... Figure 4 The communication module 301 and the processing module 302 are shown.

[0149] The processor 401 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0150] The memory 402 can be an internal storage unit of the electronic device 400, such as a hard disk or memory of the electronic device 400. The memory 402 can also be an external storage device of the electronic device 400, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device 400. Furthermore, the memory 402 can include both internal and external storage units of the electronic device 400. The memory 402 is used to store the computer program and other programs and data required by the terminal. The memory 402 can also be used to temporarily store data that has been output or will be output.

[0151] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0152] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0153] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0154] In the embodiments provided by this invention, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0155] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0156] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0157] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0158] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for dynamic load demand forecasting of an air conditioning system, characterized in that, include: Acquire monitoring data of the air conditioning system during historical periods, including building thermal inertia data, circulating water data, internal disturbance data, and external disturbance data; Based on the monitoring data of the air conditioning system during the historical period, circulating water temperature is identified, and the lag factor of the air conditioning system is calculated. This includes: cleaning the monitoring data to obtain cleaned data; performing pattern recognition based on the cleaned data to determine the operating data of the air conditioning system under stable operating conditions; and performing circulating water temperature identification based on the operating data of the air conditioning system under stable operating conditions to calculate the lag factor. The lag factor is the required water supply temperature and cumulative flow rate of the air conditioning system when the user's room temperature changes from a first room temperature to a second room temperature within a set time period. Based on the monitoring data of the air conditioning system, an impact factor analysis is performed to obtain the impact characteristics of the air conditioning system; the impact characteristics are the features in the monitoring data that have an impact on the lag factor. Based on the aforementioned impact characteristics and the aforementioned lag factor, a machine learning model is trained to obtain a dynamic load forecasting model; the dynamic load forecasting model takes the lag factor as output and the aforementioned impact characteristics as input. Based on the dynamic load forecasting model, the dynamic load demand of the air conditioning system is predicted; The process of identifying circulating water temperature and calculating the lag factor based on the stable operating data of the air conditioning system includes: identifying multiple abrupt changes in supply water temperature over a historical period based on the stable operating data; identifying multiple abrupt changes in return water temperature over a historical period based on the stable operating data; matching the multiple abrupt changes in supply water temperature and the multiple abrupt changes in return water temperature to obtain multiple pairs of abrupt changes in supply and return water temperature; calculating the time interval between each pair of abrupt changes in supply and return water temperature; determining the cumulative flow of each pair of abrupt changes in supply and return water temperature based on the instantaneous flow rate and the time interval between each pair of abrupt changes in supply and return water temperature; querying the stable operating data to determine the user's first room temperature corresponding to the supply water abrupt change and the user's second room temperature corresponding to the return water abrupt change in each pair of abrupt changes in supply and return water temperature; and determining the lag factor based on the user's first room temperature and second room temperature, as well as the cumulative flow rate and supply water temperature of each pair of abrupt changes in supply and return water temperature.

2. The method for dynamic load demand forecasting of an air conditioning system according to claim 1, characterized in that, The prediction of dynamic load demand for the air conditioning system based on the dynamic load prediction model includes: Obtain real-time monitoring data and target room temperature of the air conditioning system; Based on the real-time monitoring data and the target room temperature, n-dimensional input features are generated; The n-dimensional input features are input into the dynamic load prediction model to obtain the water supply temperature and lag factor. Based on the water supply temperature and hysteresis factor, the water supply flow rate and control duration of the air conditioning system are determined.

3. The method for dynamic load demand forecasting of an air conditioning system according to claim 1, characterized in that, The influence factor analysis, based on the monitoring data of the air conditioning system, yields the influence characteristics of the air conditioning system, including: Based on the multiple pairs of supply and return water temperature change nodes, the monitoring data in the air conditioning system is segmented to obtain monitoring data for multiple time periods in multiple response intervals; Based on monitoring data from multiple time periods within the multiple response intervals, an impact factor analysis was performed to obtain the impact characteristics of the air conditioning system. The impact characteristics include building thermal inertia characteristics, circulating water characteristics, internal disturbance characteristics, and external disturbance characteristics. The building thermal inertia characteristics include continuous cooling duration, outdoor temperature during cooling, and indoor temperature during cooling. The circulating water characteristics include instantaneous circulating water flow rate, cumulative circulating water flow rate, supply water temperature, and return water temperature. The internal disturbance characteristics include room temperature data and indoor occupant activity intensity. The external disturbance characteristics include outdoor temperature, outdoor wind speed, and irradiance.

4. The method for dynamic load demand forecasting of an air conditioning system according to claim 1, characterized in that, The step of training a machine model based on the influencing characteristics and the lag factor to obtain a dynamic load prediction model includes: Using the aforementioned impact features as input features and the aforementioned lag factor as output features, training samples are generated; Based on the training samples, machine model training is performed to obtain an initial model; The input features in the training samples are removed with replacement, feature ablation is performed, the machine model is retrained, the initial model is optimized, and the dynamic load prediction model is obtained.

5. The method for dynamic load demand forecasting of an air conditioning system according to claim 4, characterized in that, The process of removing input features from the training samples with replacement, performing feature ablation, retraining the machine model, optimizing the initial model, and obtaining the dynamic load prediction model includes: Step 1: Calculate the determination coefficients of the initial model and initialize the initial model as the global optimal solution; Step 2: Initialize the number of features removed from the input features to 1; Step 3: Based on the number of features removed, remove the features with the number of features removed from the input features to obtain multiple optimized input features; Step 4: Based on each optimized input feature and the hysteresis factor, train the machine model to obtain multiple optimized models; Step 5: Calculate the coefficient of determination for each optimization model, and the maximum coefficient of determination among multiple optimization models; Step Six: If the largest determination coefficient among the plurality of optimization models is greater than the determination coefficient of the global optimal solution, then the optimization model corresponding to the largest determination coefficient is determined as the global optimal solution; if the largest determination coefficient among the plurality of optimization models is less than or equal to the determination coefficient of the global optimal solution, then it remains unchanged. Step 7: Increase the number of features removed by 1, and repeat steps 3 to 7 until the number of features removed is greater than the number of features in the input features. Exit the iteration process and proceed to step 8. Step 8: Determine the global optimal solution as the dynamic load prediction model.

6. The method for dynamic load demand forecasting of an air conditioning system according to claim 4, characterized in that, The process of predicting the dynamic load demand of the air conditioning system based on the dynamic load prediction model also includes, prior to: Based on the set interval of each feature in the training sample, each feature is divided into intervals to obtain multiple value intervals for each feature. Iterate through each training sample to determine the value range corresponding to each training sample; Based on the value range corresponding to each training sample and the multiple value ranges of each feature, the feature coverage of the training sample is determined. Based on the feature coverage of the training samples, the evaluation result of the dynamic load prediction model is determined, and the evaluation result includes evaluation passed and evaluation failed.

7. A dynamic load demand forecasting device for an air conditioning system, characterized in that, include: The communication module is used to acquire monitoring data of the air conditioning system during historical periods. The monitoring data includes building thermal inertia data, circulating water data, internal disturbance data, and external disturbance data. The processing module is used to identify the circulating water temperature based on monitoring data from the air conditioning system during the historical period, and calculate the lag factor of the air conditioning system. The lag factor is the required supply water temperature and cumulative flow rate of the air conditioning system when the user's room temperature changes from a first room temperature to a second room temperature within a set time period. Based on the monitoring data from the air conditioning system, it performs influence factor analysis to obtain the influence characteristics of the air conditioning system. The influencing features are the features in the monitoring data that affect the lag factor; based on the influencing features and the lag factor, a machine model is trained to obtain a dynamic load prediction model; the dynamic load prediction model takes the lag factor as the output and the influencing features as the input. Based on the dynamic load forecasting model, the dynamic load demand of the air conditioning system is predicted; The processing module is specifically used to clean the monitoring data to obtain cleaned data; based on the cleaned data, perform pattern recognition to determine the operating data of the air conditioning system under stable operating conditions; based on the operating data of the air conditioning system under stable operating conditions, perform circulating water temperature identification and calculate the hysteresis factor. The processing module is specifically used to: identify multiple abrupt changes in supply water temperature over a historical period based on the operating data of the stable operating condition; identify multiple abrupt changes in return water temperature over a historical period based on the operating data of the stable operating condition; match the multiple abrupt changes in supply water temperature and the multiple abrupt changes in return water temperature to obtain multiple pairs of abrupt changes in supply and return water temperature; calculate the time interval of each pair of abrupt changes in supply and return water temperature; determine the cumulative flow of each pair of abrupt changes in supply and return water temperature based on the instantaneous flow rate of each pair of abrupt changes in supply and return water temperature and the time interval of each pair of abrupt changes in supply and return water temperature; query the operating data of the stable operating condition to determine the user's first room temperature corresponding to the supply water abrupt change in each pair of abrupt changes in supply water temperature and the user's second room temperature corresponding to the return water abrupt change in each pair of abrupt changes in supply and return water temperature; and determine the hysteresis factor based on the user's first room temperature and second room temperature, as well as the cumulative flow rate and supply water temperature of each pair of abrupt changes in supply and return water temperature.

8. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor being configured to invoke and run the computer program stored in the memory to perform the steps of the method as described in any one of claims 1 to 6.

Images