Method and system for intelligently optimizing parameters of air source heat pump unit in refrigeration system

Abstract

The present invention discloses a method and system for intelligently tuning the parameters of an air source heat pump unit in a refrigeration system, belonging to the field of big data model application technology, comprising a data source acquisition module for acquiring a first type of data source and a second type of data source; a first model for predicting a base load and performing preliminary control on the air source heat pump unit; a second model for predicting a load difference using a PSO-LSTM algorithm model and the second type of data source; a parameter tuning module for first superimposing the base load and the load difference to obtain the cooling capacity required by the air source heat pump unit; then calculating the standard water supply temperature corresponding to the cooling capacity based on the relationship between the cooling capacity and the water supply temperature, the return water temperature, and the flow rate; finally, regulating the water supply temperature using a regulation condition, and adjusting the parameters of the preliminary control according to the regulation result. The present invention adopts a step-by-step control method, which can not only avoid large fluctuations in the refrigeration system, but also ensure the efficient operation of the air source heat pump unit.

Description

Technical Field

[0001] The present invention belongs to the technical field of big data model application, and in particular relates to a method and system for intelligently adjusting parameters of an air source heat pump unit in a refrigeration system. Background Art

[0002] As we all know, with the rapid development of society, refrigeration systems have become essential basic equipment for high-rise or large buildings. Currently, in order to meet the cooling needs in buildings, the following two methods are mainly used:

[0003] Method 1 involves manual control of cooling system parameters by staff based on their personal experience. This approach relies on the operator's expertise and experience, requiring them to manually adjust the cooling system's operating parameters based on factors such as ambient temperature and humidity, and changes in the building's internal heat load. While this approach can meet basic cooling needs to a certain extent, its level of automation is limited, primarily relying on on-site control. Due to the lack of unified quantitative standards, system operation and management are relatively crude, often resulting in low thermal efficiency and poor cooling quality. Especially when users implement proactive energy-saving measures, fluctuations in heat load inevitably cause fluctuations in pipe network pressure, further increasing the safety control requirements of the energy station. As a primary energy source, the operation and regulation of air-source heat pumps rely primarily on manual experience and the unit's own control logic. In practice, staff must rely on experience to determine when to start and stop the heat pump and how to adjust its operating parameters to suit varying environmental conditions and load demands.

[0004] The second method is preset automatic control, that is, when the room temperature reaches the preset value, the refrigeration system automatically adjusts its parameters; this method uses pre-set control logic and parameters to enable the refrigeration system to automatically adjust its parameters when the room temperature reaches the preset value, thereby achieving control of the indoor temperature. Preset automatic control has improved the automation level of the refrigeration system to a certain extent and reduced the need for manual intervention. However, this method also has certain limitations. In actual applications, on-site operators usually set some fixed thresholds, and these thresholds often fail to fully consider the actual needs of indoor loads or real-time outdoor meteorological conditions, resulting in the air source heat pump being unable to operate under optimal conditions. In addition, due to the inability to monitor indoor terminal energy consumption, even when there is no load or extremely low load at the terminal, the circulating water pump and unit continue to operate ineffectively. The lack of an effective automation solution results in serious energy waste.

[0005] In summary, an operation model that relies on manual experience makes it difficult to ensure that air-source heat pump units consistently operate optimally under varying environmental conditions. This leads to overall system inefficiency, significant energy waste, and a poor user experience. Furthermore, manual experience is difficult to replicate and has significant limitations due to the varying energy demands of different building types. As the number of projects increases, labor costs also increase significantly.

[0006] In the field of air-source heat pumps used in refrigeration systems, recent research progress both domestically and internationally has focused on improving system energy efficiency, optimizing control strategies, and implementing intelligent regulation, all with the goal of increasing the energy efficiency ratio (COP) of air-source heat pumps. This is achieved through improved heat pump design, the selection of efficient compressors and heat exchangers, and optimized refrigerant use, among other approaches.

[0007] Some researchers have explored the performance of heat pumps under different operating conditions by combining experiments and simulations, a key area of ​​technological innovation. Several patents propose multi-parameter experimental platforms that utilize precision measurement equipment to comprehensively test the energy efficiency ratio (COP) and cooling capacity of heat pump systems under varying ambient temperature, humidity, and load conditions. Other researchers have described a heat pump performance testing device equipped with multiple sensors that can collect system operating parameters in real time and build a detailed performance database, providing an important experimental foundation for subsequent system optimization. Computational fluid dynamics (CFD) simulation technology has also played a significant role in heat pump performance research. Several patents utilize CFD technology to precisely simulate the fluid flow and heat exchange processes within heat pumps. Numerical simulations analyze the impact of different heat exchanger structures on system performance, enabling precise design and performance prediction of key heat pump components. Other researchers have proposed a comprehensive heat pump performance evaluation system that compares experimental test data with CFD simulation results to establish a more accurate performance prediction model, significantly improving the accuracy of system design.

[0008] Although some progress has been made in the research of air source heat pump refrigeration systems, existing technologies still have limitations. First, existing control algorithms are usually based on fixed models and lack the ability to adapt to environmental changes and load fluctuations in real time, resulting in poor energy efficiency performance under complex working conditions. Secondly, intelligent adjustment methods usually rely on a large amount of historical data for training and optimization, but in practical applications, data acquisition and processing may be limited, resulting in insufficient model generalization capabilities. In addition, the integration between different technologies and equipment remains a challenge, especially when multiple energy systems operate in coordination. How to achieve efficient energy management and scheduling still requires further research. Finally, although intelligent control and optimization algorithms can improve system performance, their implementation and maintenance costs may be high, limiting their application in a wider range of fields. Summary of the Invention

[0009] In response to the above-mentioned technical problems, the present invention provides a method and system for intelligently adjusting the parameters of an air source heat pump unit in a refrigeration system. First, a first type of data source is analyzed, and then the air source heat pump unit is preliminarily controlled according to the outdoor temperature. Then, an algorithm model is used to analyze a second type of data source, and then, based on the preliminary control, the parameters of the preliminary control are optimized. Through a step-by-step control approach, the present invention can ensure the efficient operation of the air source heat pump unit while avoiding large fluctuations in the system.

[0010] The operating parameters of the air source heat pump unit are automatically adjusted through intelligent means to achieve the goals of both satisfying user comfort and achieving energy conservation and emission reduction.

[0011] To achieve the above-mentioned object of the invention, the first object of the present invention is to provide a system for intelligently adjusting the parameters of an air source heat pump unit in a refrigeration system, comprising:

[0012] A data source acquisition module acquires a first-class data source and a second-class data source, wherein the first-class data source includes the operating status, supply water temperature, return water temperature, instantaneous flow rate, and outdoor temperature of the air source heat pump unit at different times; and the second-class data source includes the first-class data source, outdoor humidity, outdoor wind direction, outdoor wind speed, outdoor radiation, and the base load output by the first model at different times;

[0013] The first model is used to predict the base load and perform preliminary control of the air source heat pump unit; the first model includes:

[0014] Unit 1, using the first type of data source, constructs a corresponding relationship between the water supply temperature and the outdoor temperature;

[0015] Unit 2 calculates the base load at different times using the corresponding relationship, and performs preliminary control of the air source heat pump unit based on the base load;

[0016] The second model uses the PSO-LSTM algorithm model and two types of data sources to predict load differences;

[0017] The parameter tuning module first superimposes the basic load and the load difference to obtain the cooling capacity required by the air source heat pump unit; then calculates the standard water supply temperature corresponding to the cooling capacity based on the relationship between the cooling capacity and the supply water temperature, return water temperature, and flow rate; finally, uses the regulation conditions to regulate the water supply temperature, and adjusts the preliminary control parameters according to the regulation results.

[0018] Preferably, the first model includes a first preprocessing module, and the preprocessing of the first preprocessing module includes:

[0019] Delete the data within the preset time period after the air source heat pump unit is turned on;

[0020] Delete the data of the air source heat pump unit not starting up;

[0021] Filter out abnormal data and filter outliers for supply water temperature, return water temperature, flow rate and outdoor temperature.

[0022] Preferably, the use of the first type of data source to construct a corresponding relationship between the water supply temperature and the outdoor temperature includes:

[0023] Using the mean aggregation and quadratic polynomial fitting method, the mean of the corresponding water supply temperature is calculated according to the outdoor temperature range. The calculation formula is:

[0024] T _供水温度 = a • (T _室外温度 )^2 + b•T _室外温度 + c;

[0025] Among them, T _室外温度 is the outdoor temperature, T _供水温度 is the water supply temperature, a, b, c are constant terms in the curve equation,

[0026] Output outdoor temperature and water supply temperature reference curves.

[0027] A second object of the present invention is to provide a method for intelligently optimizing parameters of an air source heat pump unit in a refrigeration system, comprising:

[0028] S1. Obtain a first data source and a second data source, wherein the first data source includes the operating status, supply water temperature, return water temperature, instantaneous flow rate, and outdoor temperature of the air source heat pump unit at different times; and the second data source includes the first data source, outdoor humidity, outdoor wind direction, outdoor wind speed, outdoor radiation, and base load output by the first model at different times.

[0029] S2. Importing a type of data source into a first model to predict a base load; the first model is used to predict the base load and perform preliminary control of the air source heat pump unit; the first model first uses the type of data source to establish a correspondence between the water supply temperature and the outdoor temperature; then, using the correspondence, calculates the base load at different times, and performs preliminary control of the air source heat pump unit based on the base load;

[0030] S3. Import the two types of data sources into a second model to predict the load difference; the second model includes a PSO-LSTM algorithm model;

[0031] S4. Optimize the parameters of the air source heat pump unit. First, superimpose the base load and the load difference to obtain the cooling capacity required by the air source heat pump unit. Then, calculate the standard water supply temperature corresponding to the cooling capacity based on the relationship between the cooling capacity and the supply water temperature, return water temperature, and flow rate. Finally, use the regulation conditions to regulate the water supply temperature. According to the regulation results, adjust the parameters of the preliminary control.

[0032] Preferably, the first model includes a first preprocessing module, and the preprocessing of the first preprocessing module includes:

[0033] Delete the data within the preset time period after the air source heat pump unit is turned on;

[0034] Delete the data of the air source heat pump unit not starting up;

[0035] Filter out abnormal data and filter outliers for supply water temperature, return water temperature, flow rate and outdoor temperature.

[0036] Preferably, the use of the first type of data source to construct a corresponding relationship between the water supply temperature and the outdoor temperature includes:

[0037] Using the mean aggregation and quadratic polynomial fitting method, the mean of the corresponding water supply temperature is calculated according to the outdoor temperature range. The calculation formula is:

[0038] T _供水温度 = a • (T _室外温度 )^2 + b•T _室外温度 + c;

[0039] Among them, T _室外温度 is the outdoor temperature, T _供水温度 is the water supply temperature, a, b, c are constant terms in the curve equation,

[0040] Output outdoor temperature and water supply temperature reference curves.

[0041] A third object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for intelligently adjusting the parameters of the air source heat pump unit in the above-mentioned refrigeration system.

[0042] A fourth object of the present invention is to provide a computer program product, comprising a computer program, which, when executed by a processor, implements the above-mentioned method for intelligently adjusting the parameters of an air source heat pump unit in a refrigeration system.

[0043] The advantages and positive effects of this application are:

[0044] The present invention first uses a first model to analyze a first type of data source, and then performs preliminary control of the air source heat pump unit based on the corresponding relationship between outdoor temperature and water supply temperature. Then, a second model is used to analyze a second type of data source, and then, based on the preliminary control, the parameters of the preliminary control are optimized. Through a step-by-step control approach, the present invention can ensure the efficient operation of the air source heat pump unit while avoiding large fluctuations in the system. Specifically:

[0045] The present invention first uses the first model to perform predictions, and can obtain the basic load required for unit cooling at different outdoor temperatures. Based on the basic load, the basic operating parameters of the air source heat pump unit are obtained. Since the first model can directly reflect the corresponding relationship between the water supply temperature and the outdoor temperature, the water supply temperature can be quickly obtained based on the outdoor temperature, and the basic operating parameters of the air source heat pump unit can be obtained based on the water supply temperature.

[0046] The present invention then uses a second model to predict the cooling load at the next moment and calculate the load difference between the predicted load and the base load. The second model comprehensively considers parameters such as the building's target indoor temperature, historical indoor temperatures, supply water temperature, return water temperature, and the impact of outdoor humidity, outdoor wind direction, and outdoor radiation on the building. Through real-time prediction using the PSO-LSTM algorithm model, the load difference can be quickly determined.

[0047] Finally, the present invention superimposes the basic load and the load difference to obtain the cooling capacity that the air source heat pump unit needs to provide; calculates the water supply temperature corresponding to the cooling capacity based on the relationship between the load and the supply water temperature, return water temperature, and flow rate, and uses the regulation conditions to regulate the water supply temperature. According to the regulation results, the parameters of the air source heat pump unit are adjusted. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0049] Figure 1 Shows a system block diagram of a preferred embodiment of the present invention;

[0050] Figure 2 A flow chart showing a preferred embodiment of the present invention is shown;

[0051] Figure 3 The curve constructed according to Table 1 in the preferred embodiment of the present invention is shown;

[0052] Figure 4shows an architectural diagram of the second model in a preferred embodiment of the present invention;

[0053] Figure 5 A comparison diagram of the unit performance coefficient before and after using the technical solution of the present invention is shown in the preferred embodiment of the present invention;

[0054] Figure 6 The figure shows the indoor temperature change diagram before and after the technical solution of the present invention is used in the preferred embodiment of the present invention;

[0055] Figure 7 A point cloud diagram showing energy consumption changes before and after the present invention is applied to a sports venue is shown;

[0056] Figure 8 The energy consumption curve of the present invention before and after being applied to a sports venue is shown;

[0057] Figure 9 The figure shows a point cloud diagram and a histogram of room temperature changes before and after the present invention is applied to a sports stadium;

[0058] Figure 10 A point cloud diagram showing energy consumption changes before and after the present invention is applied to a hospital is shown;

[0059] Figure 11 The figure shows the energy consumption change curve before and after the present invention is applied to a hospital;

[0060] Figure 12 A curve diagram showing room temperature changes before and after the present invention is applied to a hospital is shown;

[0061] Figure 13 A point cloud diagram of room temperature changes before and after the present invention is applied to a hospital is shown. DETAILED DESCRIPTION

[0062] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0063] See also Figure 1 ; A first embodiment, a system for intelligently adjusting parameters of an air source heat pump unit in a refrigeration system, mainly comprising:

[0064] The data source acquisition module acquires a first-class data source and a second-class data source. The first-class data source includes the operating status, water supply temperature, return water temperature, instantaneous flow rate, and outdoor temperature of the air source heat pump unit at different times. The second-class data source includes the first-class data source, indoor measured temperature, indoor target temperature, outdoor humidity, outdoor wind direction, outdoor wind speed, outdoor radiation, and the basic load output by the first model at different times. The indoor measured temperature refers to the temperature actually measured indoors by a temperature sensor. The indoor target temperature refers to the ideal indoor temperature in the building (the indoor target temperature to be achieved by the cooling unit to make the room temperature suitable, such as 20°C). It should be noted that the data source includes a historical data source and a current-time prediction data source. The historical data source is used to establish the first model and the second model. The current-time prediction data source is used to input the first model and the second model to obtain the prediction results.

[0065] The first model is used to predict the base load and perform preliminary control of the air source heat pump unit; the first model is designed to capture the cooling demand characteristics of the building under different outdoor temperature conditions. The first model reflects the thermal response characteristics of the building through a curve. For example, the horizontal axis of the curve represents the outdoor temperature, and the vertical axis represents the water supply temperature (such as the unit water supply set temperature or the unit's actual water supply temperature). The construction of this curve is based on a large number of measured historical data sources, which is used to depict the water supply temperature required to meet indoor comfort requirements under various outdoor temperature conditions. The first model not only reflects the physical properties of the building itself, but also takes into account the user's comfort preferences. It can serve as an important basis for formulating water supply temperature strategies. Through this basis, it can be ensured that the final water supply temperature not only meets the actual needs of the building, but also avoids excessive deviation from the ideal value;

[0066] The first model mainly includes:

[0067] Unit 1 uses the first type of data source to construct a corresponding relationship between the water supply temperature and the outdoor temperature. The corresponding relationship can be constructed in the following manner:

[0068] First, obtain the required historical data sources, including the operating status, supply water temperature, return water temperature, flow rate, outdoor temperature, etc. of the air source heat pump unit at different times.

[0069] Then, the historical data source is preprocessed using the first preprocessing module, including removing data within a preset time (e.g., 30 minutes) after the air source heat pump unit is turned on (e.g., if it is turned on at 2 p.m., data before 2:30 p.m. is removed), deleting data when the air source heat pump unit is not turned on, filtering out abnormal data (e.g., data where the supply water temperature is lower than the return water temperature), and filtering outliers for the supply water temperature, return water temperature, flow rate, and outdoor temperature. It should be noted that preprocessing is not required, but is preferably performed to improve the accuracy of the first model.

[0070] Finally, using the aforementioned data source, a corresponding relationship between the water supply temperature and the outdoor temperature is constructed; specifically:

[0071] At each outdoor temperature, the average supply and return water temperature difference is calculated based on the corresponding supply water temperature, return water temperature, and instantaneous flow rate;

[0072] Average supply and return water temperature difference = return water temperature - supply water temperature = cooling capacity / [(specific heat capacity of water) • instantaneous flow rate];

[0073] Based on the water supply temperature and outdoor temperature data, the corresponding relationship between the water supply temperature and the outdoor temperature is established:

[0074] Using the mean aggregation and quadratic polynomial fitting method, the mean of the corresponding water supply temperature is calculated according to the outdoor temperature range. The calculation formula is:

[0075] T _供水温度 = a • (T _室外温度 )^2 + b•T _室外温度 + c;

[0076] Among them, T _室外温度 is the outdoor temperature, T _供水温度 is the water supply temperature, a, b, c are constant terms in the curve equation,

[0077] Output outdoor temperature and water supply temperature reference curves.

[0078] See also Figure 3 ,The following uses a specific case to elaborate on the construction process of the corresponding relationship; First, obtain a set of data, as shown in Table 1:

[0079] Table 1 shows the actual water supply temperature and set water supply temperature of the unit at different outdoor temperatures

[0080]

[0081] The data in Table 1 are all historical data; then, based on the data in Table 1, we can construct Figure 3The curve shown, where a = 0.01691, b = -1.276, c = 35.42;

[0082] The curve is updated daily in actual application, incorporating historical data from the previous day into the model.

[0083] Unit 2 uses the corresponding relationship to calculate the base load at different times and performs preliminary control of the air source heat pump unit based on the base load; specifically:

[0084] First, different outdoor temperatures are introduced into the corresponding relationship (the first model curve after integration) to obtain the water supply temperature under different outdoor temperatures;

[0085] Then calculate the base load at the corresponding time, such as cooling capacity demand, power consumption, etc., where:

[0086] The calculation formula of the cooling capacity is: cooling capacity = specific heat capacity of water • instantaneous flow rate • average supply and return water temperature difference;

[0087] The calculation formula for the power consumption is: power consumption = air source unit power consumption + water pump power consumption.

[0088] Finally, the data of the first model is stored in the database, which may include time, outdoor temperature, supply water temperature, return water temperature, load and other information to facilitate subsequent query and analysis.

[0089] Table 2 shows the first model data stored in the database.

[0090]

[0091] The first model primarily reflects the characteristic curves of the building's outdoor temperature and supply water temperature, without factoring in other key environmental factors such as humidity, wind speed, and solar radiation. Therefore, the supply water temperature calculated solely by the first model cannot fully meet the requirements for precise control. Therefore, the second model is introduced to accurately predict the supply water temperature and load.

[0092] In order to improve the accuracy of the first model, it is necessary to perform model update training regularly. For example, at 0:00 every day, the data of the previous day can be imported into the first model for training, thereby updating the first model.

[0093] Using the first model, the correspondence between the outdoor temperature and the water supply temperature can be obtained. In actual use, when the outdoor temperature change is detected, the water supply temperature corresponding to each outdoor temperature can be quickly obtained, and then the working state of the unit can be controlled according to the water supply temperature and the water supply temperature provided according to the corresponding demand.

[0094] The second model uses the PSO-LSTM algorithm model and two types of data sources to predict load differences;

[0095] In the second model of the air-source heat pump system, outdoor temperature, outdoor humidity, outdoor wind speed, outdoor wind direction, and outdoor radiation are selected as features based on the complex mechanisms of system thermodynamics and energy transfer. Solar radiation intensity directly affects the building's heat exchange process, altering the building's exterior surface temperature and, in turn, affecting the heat pump system's operating load. Outdoor wind direction and speed significantly influence the heat pump system's energy efficiency and heat transfer by regulating the convective heat transfer coefficient. These multidimensional features are not simply independent indicators but rather interconnected, interdependent, and interoperable systemic elements. By incorporating these features and utilizing the PSO-LSTM algorithm, we can more comprehensively and accurately capture the inherent laws of system operation, significantly improving the accuracy and adaptability of the load forecasting model and providing more precise decision-making support for the intelligent control of air-source heat pump systems.

[0096] The second model includes:

[0097] The second preprocessing module, the second model, focuses on missing values ​​and outliers. Because the data is all minute-level time series data and the data volume is abundant, the data with missing features is filtered. For outliers in the data after filtering missing values, the following processing is performed:

[0098] a. Filter abnormal values ​​of supply water temperature, return water temperature, flow rate and outdoor temperature;

[0099] b. Filter and remove data where the supply water temperature is greater than or equal to the return water temperature;

[0100] c. Filter and remove data in the unpowered state;

[0101] d. Filter and remove samples within 30 minutes after startup (in the event of abnormal water supply temperature rise, which adds noise interference to the second model training);

[0102] e. Exclude features from columns with the same value.

[0103] Actual data examples (mainly showing some abnormal data);

[0104] Table 3 is an example table of missing values ​​and outliers that the second model focuses on

[0105]

[0106] Training target and feature selection modules,

[0107] One of the key steps in modeling the second model is determining the load difference ΔQ learned by the regression model. First, the average load (average cooling capacity) for each outdoor temperature range at the target room temperature is used. Then, polynomial regression is used to determine the relationship between outdoor temperature and average load. This relationship is applied to all samples to obtain the standard load (standardQ) for all samples (each sample corresponds to an average load). Finally, the load difference ΔQ for all samples is calculated by subtracting the unit's current load (calculated using Q = cmΔt) from the standard load. Regarding feature selection and processing, the second model, considering the small number of total features, does not require screening. Only the indoor temperature sensor (some projects also include the indoor humidity sensor) is averaged, which in turn derives features such as whether it is a weekday, X days, and X hours.

[0108] In this embodiment, the training target refers to the difference between the standard load and the current load, that is, the target load difference.

[0109] The results of feature selection include outdoor temperature, unit supply water temperature, unit return water temperature, instantaneous flow rate, room temperature sensor data (after mean processing), outdoor meteorological parameters (outdoor temperature, outdoor humidity, outdoor wind speed, outdoor wind direction, outdoor radiation, etc.), whether it is a weekday, and other time-related features.

[0110] Model selection, see Figure 4 Regarding model selection, first, and then by comparing the performance of various models, we found that PSO-LSTM outperformed other alternative models in terms of performance indicators. The PSO-LSTM algorithm, which combines Particle Swarm Optimization (PSO) with Long Short-Term Memory (LSTM), has significant advantages in time series forecasting tasks. By combining the global optimization capabilities of PSO with the time series processing capabilities of LSTM, the PSO-LSTM algorithm significantly improves the performance of time series forecasting tasks. Its advantages mainly focus on hyperparameter optimization, adaptive learning rate adjustment, overfitting avoidance, and parallel computing. These advantages include improved prediction accuracy, faster convergence, enhanced robustness, and reduced manual intervention.

[0111] The model training and evaluation module first sets the training sample sampling duration. For example, the second model uses 10-minute training samples (the specific duration can be adjusted as needed). Time series aggregation is then performed. Finally, the dataset is split into training and testing at a ratio of 4:1. After deployment, all data (excluding the current time) is used for training, and the current time data is used for prediction.

[0112] Model evaluation is a critical step in ensuring the reliability and generalization capabilities of machine learning models. This model uses a multi-dimensional evaluation method, using k-fold cross-validation. The dataset is randomly divided into k subsets, with k-1 subsets selected for training each time, and the remaining subset used for validation. This reduces accidental errors and improves the stability of model evaluation.

[0113] In addition to the R² coefficient, we also used the root mean square error (RMSE) and mean absolute error (MAE) as supplementary evaluation metrics. R² reflects the model's ability to explain variance, while RMSE and MAE directly reflect the absolute error in the prediction results. Preliminary experimental results show that the model achieved an R² of over 0.97 on the test set, with an RMSE within ±0.3 degrees and a MAE of approximately 0.2 degrees, demonstrating high prediction accuracy.

[0114] The parameter tuning module first superimposes the basic load and the load difference to obtain the cooling capacity required by the air source heat pump unit; then calculates the standard water supply temperature corresponding to the cooling capacity based on the relationship between the cooling capacity and the supply water temperature, return water temperature, and flow rate; finally, uses the regulation conditions to regulate the water supply temperature, and adjusts the preliminary control parameters according to the regulation results.

[0115] In the parameter tuning module, the base load and the load difference are first superimposed to calculate the cooling capacity required by the air source heat pump unit. Next, based on the relationship between cooling capacity and supply water temperature, return water temperature, and flow rate, the corresponding supply water temperature is further calculated. After obtaining this supply water temperature, it is regulated using established regulation conditions. Based on the results of the regulation, the initially set control parameters are adjusted accordingly to ensure the efficient and accurate operation of the air source heat pump unit.

[0116] The second model is used for prediction and post-processing of the prediction results. The prediction result of the second model is the load difference △Q. The predicted cooling capacity required by the unit is equal to the base load + load difference △Q. Then, through the relationship between load and supply water temperature, return water temperature and flow rate Q=cmΔt, T 供 =T 回 -(Q / cm), calculate the unit water supply temperature corresponding to the cooling capacity, T 供 is the water supply temperature, T 回 is the return water temperature, c is the specific heat capacity of water, m is the instantaneous flow rate, Δt is the difference between the supply water temperature and the return water temperature, Δt=T 回 -T 供 According to the specific project operation situation, the calculated water supply temperature is regulated and restricted. For example, the upper and lower limits of the water supply temperature have certain constraints, such as (9, 14);

[0117] The absolute value of the difference between the water supply temperature predicted by the second model and the water supply temperature predicted by the first model does not exceed 2 degrees;

[0118] Time-sharing water supply temperature range limit;

[0119] In the present invention, the first model outputs a correlation curve between outdoor temperature and water supply temperature, reflecting the characteristics of the building itself. This curve, obtained through regression fitting and smoothing of the sample distribution, can provide a corresponding relationship between various outdoor temperature ranges and water supply temperatures. However, under the various outdoor meteorological parameters and indoor temperature and humidity conditions in actual business scenarios, the water supply temperature it provides cannot reflect the real-time changes in various parameters (such as outdoor humidity, outdoor wind speed, outdoor wind direction, outdoor radiation, and other key environmental factors), and its accuracy needs to be calibrated. The second model is used to compensate for or reduce the insufficient or excessive cooling capacity predicted by the first model. By combining the first and second models, the base load forecast provided by the first model and the load difference predicted by the second model based on real-time meteorological parameters and room temperature parameters are superimposed. It is expected that a more accurate load forecast can be obtained for each characteristic parameter at that moment.

[0120] The second model can compensate for the insufficient or excessive cooling capacity predicted by the first model due to the current outdoor meteorological conditions.

[0121] For the second embodiment, please refer to Figure 2 A method for intelligently optimizing parameters of an air source heat pump unit in a refrigeration system, using the system of the first embodiment, performs the following steps:

[0122] S1. Obtain a first data source and a second data source, wherein the first data source includes the operating status, supply water temperature, return water temperature, instantaneous flow rate, and outdoor temperature of the air source heat pump unit at different times; and the second data source includes the first data source, outdoor humidity, outdoor wind direction, outdoor wind speed, outdoor radiation, and base load output by the first model at different times.

[0123] In this example, we aim to acquire and analyze two important data sources for further data processing and analysis. One source primarily involves the operating status of the air-source heat pump unit at different points in time. This includes key parameters such as the unit's operating status, supply water temperature, return water temperature, instantaneous flow rate, and outdoor temperature. This data is crucial for understanding the unit's real-time performance and efficiency, helping us monitor and optimize the operation of the heat pump system.

[0124] The second type of data source contains more comprehensive information. It not only includes all the data from the first type of data source, but also covers other important parameters related to the outdoor environment. These parameters include outdoor humidity, outdoor wind direction, outdoor wind speed, outdoor radiation intensity, and base load data calculated by the first model. These environmental parameters are crucial for evaluating the performance of heat pump units under different environmental conditions. They help the system more accurately predict and adjust the heat pump system's operating strategy to adapt to changing external environmental conditions.

[0125] S2. Importing a type of data source into a first model to predict a base load; the first model is used to predict the base load and perform preliminary control of the air source heat pump unit; the first model first uses the type of data source to establish a correspondence between the water supply temperature and the outdoor temperature; then, using the correspondence, calculates the base load at different times, and performs preliminary control of the air source heat pump unit based on the base load;

[0126] S3. Import the two types of data sources into a second model to predict the load difference; the second model includes a PSO-LSTM algorithm model;

[0127] S4. Optimize the parameters of the air source heat pump unit. First, superimpose the base load and the load difference to obtain the cooling capacity required by the air source heat pump unit. Then, calculate the standard water supply temperature corresponding to the cooling capacity based on the relationship between the cooling capacity and the supply water temperature, return water temperature, and flow rate. Finally, use the regulation conditions to regulate the water supply temperature. According to the regulation results, adjust the parameters of the preliminary control.

[0128] A third embodiment is a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned method for intelligently adjusting parameters of an air source heat pump unit in a refrigeration system.

[0129] A fourth embodiment is a computer program product, comprising a computer program, which, when executed by a processor, implements the above-mentioned method for intelligently adjusting parameters of an air source heat pump unit in a refrigeration system.

[0130] See also Figure 5After using the technical solution of the present application, the coefficient of performance (COP) of the air source heat pump unit is significantly improved; the present invention first provides a benchmark water supply temperature (the water supply temperature obtained by the basic load of the first model), which can accurately depict the water supply temperature corresponding to each outdoor temperature range. This provides a scientific basis for the operation of the air source heat pump unit, and then dynamically adjusts the model (the second model) according to the basic water supply temperature provided by the first model, comprehensively considering environmental factors such as outdoor humidity, outdoor wind direction, outdoor wind speed and solar radiation, thereby further improving the adaptability and response speed of the system. Through this load matching method, the system can maintain the best cooling effect under different meteorological conditions, avoiding the problem of insufficient or excessive cooling capacity supply due to environmental changes, thereby significantly reducing power consumption and improving the unit's coefficient of performance;

[0131] See also Figure 6 After using the technical solution of this application, the comfort of indoor occupants can be improved. The combination of the two models allows users to feel higher comfort while enjoying a stable supply of cooling air, significantly reducing the indoor temperature during the cooling season. By real-time monitoring and dynamic adjustment of the water supply temperature, the system can better adapt to changes in user needs and ensure the comfort of the indoor environment. This flexibility not only improves user satisfaction, but also improves the intelligence level of the system. During use, users can experience more precise temperature control effects, avoiding the temperature fluctuations and discomfort common in traditional systems, thereby improving the comfort of living and working environments.

[0132] By improving the energy efficiency of air-source heat pump units, the system reduces energy consumption and carbon emissions, promoting sustainable development. It enables the system to maintain efficient operation even in complex weather conditions, reducing energy waste caused by over- or under-cooling. This invention effectively reduces operating costs. By optimizing cooling capacity and improving energy efficiency, the system's overall operating expenses are lowered, allowing users to enjoy a comfortable environment while also reducing their financial burden. During operation, the refrigeration system intelligently adjusts its operating strategy based on real-time meteorological data and changes in the indoor environment, avoiding unnecessary energy consumption and resource waste. This economical approach makes this method highly competitive in the market, attracting greater user interest and adoption, particularly in commercial buildings and large public facilities, significantly improving return on investment. This invention can adapt to the application needs of different regions and climate conditions. It can adjust based on real-time data to ensure optimal cooling performance at all times. This flexibility makes this technology suitable not only for new buildings but also for retrofitting the cooling systems of existing buildings, improving their energy efficiency and comfort.

[0133] See also Figure 10 , Figure 10The energy consumption point cloud of a hospital before and after the algorithm was used is shown. The blue point cloud represents the energy consumption point cloud after the algorithm was used under different temperature conditions, and the red point cloud represents the energy consumption point cloud before the algorithm was used under different temperature conditions.

[0134] See also Figure 11 , Figure 11 The figure shows the energy consumption changes of a hospital before and after using the algorithm. The blue curve is the energy consumption curve after using the algorithm at different times, and the red curve is the energy consumption curve before using the algorithm at different times.

[0135] See also Figure 12 , Figure 12 The figure shows the room temperature changes in a hospital before and after the algorithm was used. The blue curve shows the room temperature curve after the algorithm was used at different times, and the red curve shows the room temperature curve before the algorithm was used at different times.

[0136] See also Figure 13 , Figure 13 The figure shows the changes in room temperature in a hospital before and after the algorithm was used. The blue point cloud represents the indoor temperature point cloud after the algorithm was used under different outdoor temperature conditions, and the red point cloud represents the indoor temperature point cloud before the algorithm was used under different temperature conditions.

[0137] Experimental environment

[0138] The example used in this paper is a hospital project with an actual energy consumption area of ​​42,000 square meters, encompassing the inpatient building, outpatient building, and ancillary buildings. The project employed 18 Midea air source units, model DNL-E1550 / NSN1-H2. The air source units were located in the hospital's ground floor space, while two circulating water pumps and ancillary equipment were housed in a basement-level machine room. Fan coil units were installed at the end of the energy supply area.

[0139] Experimental setup

[0140] In this example, the algorithm uses the unit's cooling water supply temperature and indoor temperature as key parameters. The target cooling water supply temperature range is set at 7°C to 13°C. The average terminal room temperature is maintained between 25°C and 27°C. The number of circulating water pumps is controlled based on the supply and return water temperature differential and the predicted load. The unit start / stop temperature differential is set at 2°C.

[0141] Experimental process

[0142] To monitor indoor temperature and humidity at the end point, 30 indoor temperature and humidity sensors were installed in representative areas. In this example, the cooling period in 2024 is from May 15th to October 13th. To compare the effectiveness of the algorithm control, manual control was performed from May 15th to July 31st, and the algorithm was applied from August 1st to October 13th. The main strategies in the algorithm control process are as follows:

[0143] Temperature setting: after taking over, the water supply temperature is adjusted in real time according to the outdoor temperature and indoor temperature;

[0144] Adjust the number of units and gradually change the number of units turned on according to the outdoor temperature;

[0145] Adjust the number of water pumps and start different numbers and frequencies of water pumps according to the energy consumption characteristics during the day and at night.

[0146] Result analysis: This example compares the effects of algorithm control and manual control from several aspects, such as energy consumption data, indoor temperature comfort, and unit performance coefficient. Data analysis and comparison show that the algorithm is superior to manual control in these aspects.

[0147] Compared with manual experience, the algorithm control can set the water outlet temperature higher under the same outdoor temperature while ensuring that the indoor temperature is within a comfortable range;

[0148] Compared with manual experience, the total power consumption of algorithm control is relatively lower under the same outdoor temperature conditions;

[0149] Compared with manual experience, the algorithm control can slightly improve the unit COP under the same outdoor temperature conditions.

[0150] See also Figure 7 , Figure 7 The energy consumption point cloud of a sports venue before and after the algorithm was used is shown. The blue point cloud represents the energy consumption point cloud after the algorithm was used under different temperature conditions, and the red point cloud represents the energy consumption point cloud before the algorithm was used under different temperature conditions.

[0151] See also Figure 8 , Figure 8 The figure shows the energy consumption changes of a sports venue before and after the algorithm is used. The blue curve is the energy consumption curve after the algorithm is used, and the red curve is the energy consumption curve before the algorithm is used.

[0152] See also Figure 9 , Figure 9 The figure shows the changes in room temperature in a sports venue under different outdoor temperature conditions before and after the algorithm is used. In the figure on the left, the blue point cloud is the room temperature point cloud under different outdoor temperature conditions after the algorithm is used, and the red point cloud is the room temperature point cloud under different outdoor temperature conditions before the algorithm is used; the figure on the right is the histogram corresponding to the figure on the left, showing the concentration of room temperature distribution after the algorithm is used. It can be seen that after the algorithm is used, the room temperature is mostly distributed in a more comfortable range.

[0153] Experimental environment:

[0154] This invention was tested at a large sports venue using the DNL-E1550 / NSN1-H2 series air-source heat pump unit. This unit offers high efficiency and adaptability, making it suitable for the cooling and heating needs of large venues. Experimental conditions included an outdoor temperature range of 18°C ​​to 37°C, an indoor setpoint temperature of 26°C, and a humidity range of 30% to 70%. During the experiment, the venue's usage simulated peak attendance to ensure representative results.

[0155] Experimental setup

[0156] Parameter settings: During the algorithm application process, the key parameters and their value ranges are as follows:

[0157] Outdoor temperature: 18℃~37℃;

[0158] Indoor set temperature: 26℃;

[0159] Water supply temperature: calculated based on the initial load matching model, ranging from 8°C to 14°C;

[0160] Humidity (H): 30% to 70%;

[0161] Wind speed (V): 0.5 m / s to 5 m / s;

[0162] Solar radiation (SR): 0 W / m² to 800 W / m²;

[0163] Model baseline water supply temperature: obtained through polynomial regression analysis of historical data;

[0164] Load matching compensation: Dynamically adjust according to real-time environmental parameters, use the LSTM algorithm to predict the load required at the next moment, and comprehensively consider multi-dimensional factors such as room temperature to infer the corresponding water supply temperature.

[0165] Experimental process:

[0166] 1. Experimental Preparation:

[0167] Sensors are installed inside and outside the sports venues to monitor real-time environmental parameters such as indoor and outdoor temperature, humidity, wind speed, and solar radiation. These sensors provide highly accurate data to ensure the accuracy of the experiments.

[0168] Configure the air source heat pump unit and ensure its normal operation, and carry out necessary debugging and testing to ensure the stability of the system during the experiment.

[0169] 2. Data Collection:

[0170] Under different outdoor temperature conditions, record the system operation data such as water supply temperature, indoor temperature and humidity.

[0171] A two-month experiment was conducted to collect environmental data and system operation data at different time periods.

[0172] 3. Algorithm application:

[0173] First, based on historical operating data such as outdoor temperature and water supply temperature, the model is used to calculate the benchmark water supply temperature.

[0174] Then, the load matching compensation mechanism is utilized to dynamically adjust the water supply temperature of the baseline model according to the real-time monitored environmental parameters using a machine learning prediction algorithm.

[0175] 4. Comparative experiment:

[0176] A comparative test was conducted using the traditional manual experience-based strategy to adjust the water supply temperature, and the changes in indoor temperature and humidity under the same environmental conditions were recorded.

[0177] Result analysis: The experimental results show that after adopting the load matching algorithm model of the present invention, the cooling effect of the sports venue is significantly better than the traditional method.

[0178] Energy efficiency ratio (COP):

[0179] The air-source heat pump system using the method of the present invention achieved an average COP of 3.98 throughout the experimental period, compared to 3.23 for the traditional method, representing an improvement of over 23.2%. This significant improvement demonstrates the effectiveness of the method in improving system energy efficiency, enabling it to provide higher cooling capacity at the same energy consumption.

[0180] Indoor temperature stability:

[0181] In the case of large fluctuations in outdoor temperature, the indoor temperature fluctuation range of the method of the present invention is 21 o C~26 o C, kept within the set range, the indoor temperature fluctuation range of the traditional method is 20 o C~29 o C, the fluctuation range is large and the comfort level decreases. This result shows that the method of the present invention can better maintain the stability of the indoor environment and improve the user experience.

[0182] Energy consumption analysis:

[0183] The method excels in energy consumption control, reducing energy consumption by 24.21% under the same outdoor temperature conditions, saving costs and bringing significant economic benefits. By optimizing the operation strategy, unnecessary energy consumption is reduced, improving the economic and sustainability of the system.

[0184] In summary, this experiment verified the universality and effectiveness of the load matching algorithm model in sports venue projects, significantly improved the energy efficiency, comfort and user satisfaction of the air source heat pump unit, and demonstrated the wide applicability and superiority of this technology in practical applications.

[0185] In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware, or any combination thereof. When implemented in whole or in part in the form of a computer program product, the computer program product includes one or more computer instructions. When the computer program instructions are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL) or wireless (e.g., infrared, wireless, microwave, etc.)) method. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)).

[0186] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A system for intelligently adjusting parameters of air source heat pump units in a refrigeration system, characterized in that: Avoid large fluctuations in the cooling system, including: A data source acquisition module acquires a first-class data source and a second-class data source, wherein the first-class data source includes the operating status, supply water temperature, return water temperature, instantaneous flow rate, and outdoor temperature of the air source heat pump unit at different times; and the second-class data source includes the first-class data source, outdoor humidity, outdoor wind direction, outdoor wind speed, outdoor radiation, and the base load output by the first model at different times; The first model is used to predict the base load and perform preliminary control of the air source heat pump unit; the first model includes: Unit 1 uses the aforementioned data source to construct a corresponding relationship between the water supply temperature and the outdoor temperature. The mean of the corresponding water supply temperature is calculated according to the outdoor temperature range using the mean aggregation and quadratic polynomial fitting method. The calculation formula is: T _供水温度 = a •(T _室外温度 )^2 + b•T _室外温度 + c; Among them, T _室外温度 is the outdoor temperature, T _供水温度 is the water supply temperature, a, b, c are constant terms in the curve equation, Output outdoor temperature and water supply temperature reference curve; Unit 2 calculates the base load at different times using the corresponding relationship, and performs preliminary control of the air source heat pump unit based on the base load; The second model uses the PSO-LSTM algorithm model and two types of data sources to predict load differences; The parameter tuning module first superimposes the basic load and the load difference to obtain the cooling capacity required by the air source heat pump unit; then calculates the standard water supply temperature corresponding to the cooling capacity based on the relationship between the cooling capacity and the supply water temperature, return water temperature, and flow rate; finally, uses the regulation conditions to regulate the water supply temperature, and adjusts the preliminary control parameters according to the regulation results.

2. The system for intelligently adjusting parameters of air source heat pump units in a refrigeration system according to claim 1, characterized in that: The first model includes a first preprocessing module, and the preprocessing of the first preprocessing module includes: Delete the data within the preset time period after the air source heat pump unit is turned on; Delete the data of the air source heat pump unit not starting up; Filter out abnormal data and filter outliers for supply water temperature, return water temperature, flow rate and outdoor temperature.

3. A method for intelligently optimizing parameters of an air source heat pump unit in a refrigeration system, characterized in that: Avoid large fluctuations in the cooling system, including: S1. Obtain a first data source and a second data source, wherein the first data source includes the operating status, supply water temperature, return water temperature, instantaneous flow rate, and outdoor temperature of the air source heat pump unit at different times; and the second data source includes the first data source, outdoor humidity, outdoor wind direction, outdoor wind speed, outdoor radiation, and base load output by the first model at different times. S2. Import a type of data source into the first model to predict the base load; the first model is used to predict the base load and perform preliminary control of the air source heat pump unit; the first model first uses the type of data source to establish a corresponding relationship between the water supply temperature and the outdoor temperature; using the mean aggregation and quadratic polynomial fitting method, the mean of the corresponding water supply temperature is calculated according to the outdoor temperature range, and the calculation formula is: T _供水温度 = a •(T _室外温度 )^2 + b•T _室外温度 + c; Among them, T _室外温度 is the outdoor temperature, T _供水温度 is the water supply temperature, a, b, c are constant terms in the curve equation, Output the outdoor temperature and water supply temperature reference curves; then use the corresponding relationship to calculate the base load at different times, and perform preliminary control of the air source heat pump unit based on the base load; S3. Import the two types of data sources into a second model to predict the load difference; the second model includes a PSO-LSTM algorithm model; S4. Optimize the parameters of the air source heat pump unit. First, superimpose the base load and the load difference to obtain the cooling capacity required by the air source heat pump unit. Then, calculate the standard water supply temperature corresponding to the cooling capacity based on the relationship between the cooling capacity and the supply water temperature, return water temperature, and flow rate. Finally, use the regulation conditions to regulate the water supply temperature. According to the regulation results, adjust the parameters of the preliminary control.

4. The method for intelligently optimizing parameters of an air source heat pump unit in a refrigeration system according to claim 3, characterized in that: The first model includes a first preprocessing module, and the preprocessing of the first preprocessing module includes: Delete the data within the preset time period after the air source heat pump unit is turned on; Delete the data of the air source heat pump unit not starting up; Filter out abnormal data and filter outliers for supply water temperature, return water temperature, flow rate and outdoor temperature.

5. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method for intelligently adjusting parameters of an air source heat pump unit in a refrigeration system according to any one of claims 3 to 4 is implemented.

6. A computer-readable storage medium storing a computer program, characterized in that: When the program is executed by a processor, the method for intelligently adjusting parameters of an air source heat pump unit in a refrigeration system as described in any one of claims 3 to 4 is implemented.

Images