A plant cold source station control method and device and computer equipment

Abstract

The application relates to a plant cold source station control method and device and computer equipment, the method comprising: obtaining target meteorological data of the plant in a future first preset period, the target meteorological data comprising temperature data; determining a target load model from a plurality of typical load models pre-constructed according to a parameter range of the target meteorological data, inputting the target meteorological data into the target load model to obtain a first target load of the cold source station; the typical load model is trained based on historical meteorological data of the plant and historical load data of the cold source station, wherein the plurality of typical load models are constructed according to the temperature data of different parameter ranges; determining a pre-control parameter of the cold source station in the first preset period based on the first target load, and controlling the cold source station to work based on the pre-control parameter. The method can improve the accuracy of load prediction.

Description

Technical Field

[0001] This application relates to the field of industrial control technology, specifically to a control method, device, and computer equipment for a factory cold source station. Background Technology

[0002] Energy demand in automobile manufacturing spans the entire process, including stamping, welding, painting, final assembly, storage, and transportation. Air conditioning, boiler, and process systems account for 40%–60% of the plant's total energy consumption. Cold storage stations centrally provide cooling capacity, supplying the necessary cryogenic media (usually chilled water or other refrigerants) for production. Under the "dual carbon" target (carbon reduction and emission reduction), effective management of energy systems in the manufacturing process is crucial for energy conservation and emission reduction, thus placing higher demands on the control of cold storage stations.

[0003] In related technologies, the main control of cold source stations often uses the combination of components with rated cooling capacity to match load demand. For plant-level cold source stations, the actual load demand may change in real time during the production process and is affected by various environmental factors. Related technologies struggle to accurately predict the load demand of cold source stations. Therefore, the control of cold source stations results in low load prediction accuracy, making it difficult to achieve refined and forward-looking energy-saving control and effectively meet the energy conservation, emission reduction, and efficient operation requirements of industrial plants. Summary of the Invention

[0004] In view of the above problems, this disclosure provides a control method, device and computer equipment for a plant cold source station.

[0005] The first aspect of this disclosure provides a method for controlling a plant cooling station, comprising:

[0006] Acquire target meteorological data for the factory area in the first preset period in the future, the target meteorological data including temperature data;

[0007] Based on the parameter range of the target meteorological data, a target load model is determined from multiple pre-built typical load models. The target meteorological data is then input into the target load model to obtain the first target load of the cold source station. The typical load model is trained based on the historical meteorological data of the plant area and the historical load data of the cold source station. The multiple typical load models are constructed based on the temperature data with different parameter ranges.

[0008] Based on the first target load, the pre-control parameters of the cold source station are determined in the first preset cycle, and the operation of the cold source station is controlled based on the pre-control parameters.

[0009] The embodiments provided in this disclosure acquire target meteorological data for a future first preset period and match the corresponding specific typical load model according to the parameter range of the temperature data to predict the load for the first preset period. This can fully adapt to the load change patterns under different meteorological conditions and significantly improve the accuracy and reliability of long-term load prediction. Based on the more accurate predicted first target load, pre-control parameters are determined to achieve forward-looking and refined operation control of the cold source station. This ensures a high degree of matching between cooling supply and plant load demand, effectively reducing the frequency of problems such as response lag, coarse regulation, and excessive cooling in traditional control methods. Thus, while ensuring the plant's production and environmental needs, the overall operating efficiency of the cold source station is improved, energy consumption is reduced, and the cold source station achieves high-efficiency, energy-saving, stable, reliable, and low-carbon operation.

[0010] In an optional embodiment of the first aspect, the typical load model is trained based on historical meteorological data of the plant area and historical load data of the cooling source station, including:

[0011] Acquire historical meteorological data and seasonal load data of the plant area, including outdoor dry-bulb temperature and outdoor wet-bulb temperature;

[0012] The outdoor dry-bulb temperature and outdoor wet-bulb temperature are converted into the average dry-bulb temperature and the average wet-bulb temperature of the first preset period, respectively.

[0013] Clustering the average dry-bulb temperature and the average wet-bulb temperature of the first preset cycle yields M typical load intervals for the plant area, where M > 1;

[0014] For each typical load range of the plant area, an initial typical load model is constructed. The outdoor dry-bulb temperature and outdoor wet-bulb temperature corresponding to the typical load range of the plant area are used as input features, and the load of the first preset period is used as output features. The initial typical load model corresponding to the typical load range of the plant area is trained to obtain the trained typical load model.

[0015] The embodiments provided in this disclosure construct typical load models for multiple intervals using historical meteorological data and seasonal loads, which can improve the model's adaptability to different meteorological intervals, thereby improving the accuracy and reliability of load forecasting.

[0016] In an optional embodiment of the first aspect, after determining the pre-control parameters of the cooling station based on the first target load for the first preset period, the method further includes:

[0017] Obtain the second target load of the cold source station for a future second preset period; wherein the second preset period is shorter than the first preset period;

[0018] If the second target load is inconsistent with the load data corresponding to the first target load, the pre-control parameters are corrected by the second target load to obtain the corrected pre-control parameters. The control of the cold source station based on the pre-control parameters includes controlling the cold source station based on the corrected pre-control parameter table.

[0019] The embodiments provided in this disclosure modify the pre-control parameters by using a short-cycle second target load, thereby improving the cooling station's response to real-time load fluctuations, reducing the impact of deviations between long-cycle forecasts and actual short-term demand, and enhancing the matching degree between the control strategy and the actual load.

[0020] In an optional embodiment of the first aspect, the step of obtaining the second target load for a future second preset period of the cooling station includes:

[0021] A local load feature set is obtained based on local load association data. The local load feature set is input into the constructed local load model to obtain the predicted second target load of the cold source station for the second preset period. The local load association data includes the historical local load of the cold source station obtained at the second preset period as the collection interval, the historical production cycle information corresponding to the historical local load, and the first target load of the first preset period. The local load association data is spliced ​​to obtain the local load feature set.

[0022] The embodiments provided in this disclosure construct a local load feature set by integrating historical local load, production cycle information and the first target load. This enriches the feature dimensions of short-term load forecasting, improves the fit between the second target load and the actual production and operation status of the plant, and thus improves the accuracy of short-term forecasting.

[0023] In an optional embodiment of the first aspect, the local load model is a bidirectional long short-term memory network, and obtaining the second target load of the cold source station for a future second preset period includes:

[0024] The local load model is controlled to predict the load using a prediction period of 1 / K of the second preset period to obtain a candidate second target load, where K is a natural number greater than 1.

[0025] The second target load is determined based on K candidate second target loads within the second preset period.

[0026] The embodiments provided in this disclosure employ a bidirectional long short-term memory network to perform multiple predictions and then combine the results of multiple predictions to determine the output of the second target load. This can improve the smoothness and stability of short-term load prediction, reduce the error impact caused by single prediction fluctuations, and improve the reliability of correction control.

[0027] In an optional embodiment of the first aspect, the first preset period is 1 day, and the input features of the initial typical load model are hourly dry-bulb temperature and hourly wet-bulb temperature.

[0028] The embodiments provided in this disclosure set the first preset period to 1 day and use hourly temperature as input, which can improve the accuracy of the load change characterization of each hourly interval within the 1-day period, making the prediction results more consistent with the daily work and operation regularity of the plant's cold source station and operators, and improving the rationality of scheduling and control.

[0029] In an optional embodiment of the first aspect, acquiring historical meteorological data and seasonal load data of the plant area includes:

[0030] By loading annual meteorological data into a white-box model based on building thermophysical mechanisms, the periodic load of each workshop is obtained. The annual meteorological data includes typical annual meteorological data, historical 1-year meteorological data, and historical 2-year meteorological data. The annual meteorological data includes meteorological data of the area where the factory is located, factory building data, factory personnel data, and factory equipment load data.

[0031] The seasonal load data of the plant area is obtained by summing the periodic loads of each workshop.

[0032] This embodiment employs a building thermophysical white-box model to calculate seasonal loads, effectively reducing load data distortion caused by abnormal operating conditions such as sensor failures, equipment start-up and shutdown fluctuations, and transmission noise, which are common in traditional methods. This reduces the interference of abnormal load points on subsequent model training from the source. Furthermore, by combining typical meteorological year data with multi-year historical meteorological data, it takes into account both long-term meteorological patterns and short-term meteorological changes, making the calculated seasonal load data more closely match the actual cooling needs of the plant. This provides high-quality, less biased benchmark data support for the training of subsequent typical load models, thereby improving the overall accuracy and stability of load forecasting and providing a reliable data foundation for the forward-looking control of the cooling plant.

[0033] In an optional embodiment of the first aspect, controlling the operation of the cold source station based on the pre-control parameters includes:

[0034] During the process of controlling the cold source station through the parameters in the pre-control parameters, the adjustment range of the chilled water outlet temperature and the condensate return temperature of the cold source station host within the preset temperature range is greater than or equal to the preset adjustment step size. The adjustment step size is determined based on the control algorithm of the cold source station, the stable operating conditions of the cold source station control system, and energy consumption constraint information.

[0035] The embodiments provided in this disclosure, by setting a reasonable adjustment step size, can effectively reduce the frequency of problems such as control system oscillation, frequent actuator operation, and decreased operational stability caused by excessively small or frequent parameter adjustments. Simultaneously, it can reduce the increase in equipment wear and energy consumption caused by repeated adjustments due to minor fluctuations, significantly improving the system's operational stability and reliability while ensuring the adjustment accuracy of the cold source station.

[0036] A second aspect of this disclosure provides a control device for a plant cooling station, comprising: a data acquisition module, a first load prediction module, and a control processing module, wherein:

[0037] The data acquisition module is used to acquire target meteorological data of the factory area in the first preset period in the future, and the target meteorological data includes temperature data;

[0038] The first load forecasting module is used to determine a target load model from a number of pre-built typical load models based on the parameter range of the target meteorological data, and input the target meteorological data into the target load model to obtain the first target load of the cold source station; the typical load model is trained based on the historical meteorological data of the plant area and the historical load data of the cold source station, wherein the multiple typical load models are constructed based on the temperature data with different parameter ranges;

[0039] The control processing module is used to determine the pre-control parameters of the cold source station in the first preset period based on the first target load, and control the operation of the cold source station based on the pre-control parameters.

[0040] A third aspect of this disclosure provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a control method for a plant cold source station according to any one of the claims of this disclosure.

[0041] The fourth aspect of this disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a plant cold source station control method according to any one of the present disclosures.

[0042] The fifth aspect of this disclosure provides a computer program product, including a computer program that, when executed by a processor, implements a plant cold source station control method according to any one of the present disclosures.

[0043] Regarding the beneficial effects of any of the technical solutions in the second to fifth aspects mentioned above, refer to the beneficial effects of the corresponding technical solutions in the first aspect; repeated examples will not be listed here. Attached Figure Description

[0044] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0045] Figure 1 This is a schematic diagram of an optional application environment for a plant cold source station control method in one embodiment;

[0046] Figure 2 This is a schematic diagram of an optional process for controlling a plant cold source station in one embodiment;

[0047] Figure 3 This is a schematic diagram of an optional process for step S22 in one embodiment;

[0048] Figure 4 This is a schematic diagram of an optional process for controlling a plant cold source station in one embodiment;

[0049] Figure 5 This is a schematic diagram of an optional structure of a factory cold source station control device in one embodiment;

[0050] Figure 6 This is a schematic diagram of an optional structure of an electronic device in one embodiment. Detailed Implementation

[0051] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0052] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data used can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims. The terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, product, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, product, or apparatus. Without further limitation, the presence of other identical or equivalent elements in a process, method, product, or apparatus that includes said elements is not excluded. For example, the use of terms such as "first," "second," etc., is to denote names and does not indicate any specific order.

[0053] The control method for a plant cooling station provided in this disclosure can be applied to, for example... Figure 1 The control device for the cold source station is shown. Figure 1 This is a schematic diagram of an optional application environment for a factory cold source station control method in one embodiment, where 10 is the intelligent control platform of the cold source station, 101 is the control processing device of the cold source station, 20 is the cold source station, 201 is the control execution device in the cold source station, and data storage system 30. The factory area may include multiple cold source stations, such as... Figure 1 The diagram may include three cold source stations. In some embodiments, the control processing device 101 can process meteorological data, predict target loads, and generate pre-control parameters. When adjustments to the cold source station's chilled water outlet temperature are required, the pre-control parameters are sent to the control execution device 201 of the cold source station via a data transmission network. The control execution device 201 then executes the corresponding parameter actions to achieve control processing of the cold source station. In some embodiments of this disclosure, the control processing device may include one or more processing units. These processing units may include servers on the control platform where the control processing device is located, such as servers that centrally control the cold source stations, or remote servers, such as servers on the cold source station side that can communicate with local servers. The control processing device described in this disclosure may include, but is not limited to, various in-vehicle devices, personal computers, laptops, smartphones, tablets, wearable devices, medical devices, VR (Virtual Reality) devices, etc., or it may be a single server, server cluster, distributed subsystem, cloud processing platform, server containing blockchain nodes, and combinations thereof. The processing unit described in this disclosure may include various control units capable of performing logic processing functions, including but not limited to CPU (Central Processing Unit), PLC (Programmable Logic Controller), ECU (Electronic Control Unit), MCU (Microcontroller Unit), and controllers composed of one or more logic function units, chips, etc.

[0054] Figure 2 This is a schematic diagram of an optional flow chart for a plant cooling station control method in one embodiment, such as... Figure 2 As shown, the method can be implemented in Figure 1 The control processing device 201 shown may include the following method:

[0055] S20: Obtain target meteorological data for the factory area in the first preset period in the future, the target meteorological data including temperature data.

[0056] The factory area mentioned in this disclosure typically refers to the geographical location of the cold source station, which can be represented by the actual land area of ​​the factory. The resolution of the area can be expressed as the actual land area, or by district, county, city, etc. The factory area may include one or more cold source stations. The cold source station in the factory area is mainly used to provide the required cooling capacity support for different functional areas such as production workshops, office areas, and R&D laboratories within the factory area, in order to meet the strict requirements of production processes for ambient temperature and humidity, as well as the comfort needs of personnel. The cold source station typically includes a main unit and supporting equipment such as chilled water pumps, cooling water pumps, and cooling towers.

[0057] The target meteorological data typically refers to meteorological parameters that may affect the cooling load of the plant area within a future first preset period. During the operation of the cooling station in the plant area, temperature is a crucial factor affecting its operational efficiency. Therefore, in some embodiments of this disclosure, the target meteorological data may include temperature data. In practical applications, it may also include, but is not limited to, humidity data, light intensity data, and wind speed data. The target meteorological data can be acquired based on meteorological monitoring equipment or publicly available meteorological information, or it can receive meteorological data transmitted from other terminal devices. For example, it can be obtained in real time through data interaction with an external meteorological service platform, acquiring high-precision meteorological forecast data for the future first preset period.

[0058] In some embodiments disclosed herein, the temperature data may include outdoor dry-bulb temperature and outdoor wet-bulb temperature. Outdoor dry-bulb temperature is a fundamental temperature parameter characterizing the temperature of the outdoor environment, while outdoor wet-bulb temperature can be directly measured or calculated from outdoor dry-bulb temperature and relative humidity, reflecting the outdoor air's moisture content and heat dissipation capacity. In some embodiment scenarios, if the typical ranges corresponding to outdoor dry-bulb temperature and outdoor wet-bulb temperature are inconsistent or the difference is greater than a preset range, outdoor dry-bulb temperature can be selected for data processing or as a basis for judgment and selection to ensure the stability and consistency of the load forecasting logic.

[0059] The target meteorological data or historical meteorological data described in this disclosure (collectively referred to as meteorological data for ease of description) may include various types of data and various data formats. Taking temperature data as an example, the resolution density of meteorological data may be 30 minutes, 1 hour, or 1 day, etc. For example, it may include specific values ​​such as the highest temperature, lowest temperature, and average temperature of the next hour, providing basic data support for the selection of the subsequent target load model and the calculation of the first target load.

[0060] The first preset period described in some embodiments of this disclosure typically refers to the time span of the load of the plant's cooling station that needs to be predicted. For example, the time span of the first preset period can be 24 hours (hour, h), such as a whole day starting from the current moment, so as to plan and schedule the cooling station in advance. Of course, it can also be set to other durations, such as 48 hours or 12 hours, according to the plant's production schedule and the operating characteristics of the cooling station.

[0061] S22: Determine the target load model from multiple pre-built typical load models based on the parameter range of the target meteorological data, input the target meteorological data into the target load model, and obtain the first target load of the cold source station; the typical load model is trained based on the historical meteorological data of the plant area and the historical load data of the cold source station, wherein the multiple typical load models are constructed based on the temperature data with different parameter ranges.

[0062] In some embodiments of this disclosure, the multiple typical load models can be pre-built load prediction models. Different typical load models correspond to different parameter ranges of temperature data, enabling each model to adapt to the load variation patterns under specific meteorological conditions. For example, when the temperature data in the target meteorological data is within a preset parameter range of -5℃ to 10℃, the corresponding target load model is A; when the temperature data is within a preset parameter range of 10℃ to 25℃, the corresponding target load model is B; when the temperature data is within a preset parameter range of 25℃ to 40℃, the corresponding target load model is C, and so on. In this way, matching and determining the corresponding target load model based on the actual parameter range of the temperature data in the target meteorological data makes the prediction process more consistent with the current meteorological conditions and load characteristics. Inputting the target meteorological data into the target load model outputs the first target load of the cooling station within the first preset period. This target load can characterize the total cooling demand required by the plant in the future period.

[0063] In some embodiments of this disclosure, the different parameter ranges of the temperature data can be set manually in advance or using a preset algorithm or model. For example, several temperature parameter ranges can be divided based on the latitude information of the factory location and historical meteorological data, and a corresponding typical load model can be constructed for each temperature parameter range. Alternatively, the variation patterns of temperature and load, or the distribution information of temperature and load, can be analyzed, and the temperature parameter ranges can be further divided based on the analysis results. Or, some algorithms can be used, such as clustering the load based on temperature as the dividing standard to obtain the temperature parameter ranges divided by the algorithm, and a typical load model can be set based on the temperature parameter ranges obtained by the algorithm.

[0064] In some embodiments of this disclosure, the typical load model can be trained based on historical meteorological data of the plant area and historical load data of the cooling source station. It can learn the mapping relationship between meteorological parameters and cooling source load, thereby achieving accurate prediction of cooling demand. The typical load model can use one or more neural networks, such as RNN (Recurrent Neural Network), LSTM (Long Short-Term Memory), BPNN (Back Propagation Neural Network), etc. It can use one type of neural network or an improved neural network, or combine multiple neural networks to construct the typical load model. During the training process, the input to the typical load model is historical meteorological data of the plant area, or a dataset obtained by processing historical meteorological data. The output of the typical load model is the target load (in this embodiment, the target load for the first preset period output based on the target meteorological data can be called the first target load; similarly, in other embodiments, it can also output the second target load for the second preset period, the third target load for the third preset period, etc.).

[0065] It should be understood that the degree and pattern of influence of meteorological factors on cooling load vary within different preset parameter ranges of target meteorological data. The division of preset parameter ranges can be flexibly adjusted based on the climate characteristics of the plant location, the distribution patterns of historical meteorological data, and the actual changes in the plant's cooling load. For example, if the plant is located in a temperate region, the cooling load is usually high and significantly affected by temperature during the high-temperature period in summer. The preset parameter range for temperature data can be divided more finely, such as setting the intervals of 25℃-30℃, 30℃-35℃, and above 35℃ as independent parameter ranges. Conversely, for situations where winter temperatures are low and the cooling load is small, the temperature parameter range can be divided more broadly, such as using -10℃ to 5℃ as a unified interval. This embodiment, through this setting method, matches targeted typical load models for different parameter ranges, enabling the model to more accurately capture the intrinsic correlation between meteorological data and cooling load within that range, thereby achieving accurate prediction of cooling demand.

[0066] S24: Determine the pre-control parameters of the cold source station in the first preset cycle based on the first target load, and control the operation of the cold source station based on the pre-control parameters.

[0067] In this embodiment, the pre-control parameters can be determined based on the first target load and may include control parameters for the operation of the cold source station within a first preset period, used to characterize the operating strategy required by the cold source station to meet the predicted load. The pre-control parameters may correspond to the first target load. After obtaining the first target load of the cold source station, a specific operating strategy for the cold source station within the first preset period can be formulated based on this load, and the pre-control parameters are determined based on the operating strategy. The pre-control parameters can form a pre-control parameter table. The pre-control parameter table may include the operating parameters of each core device of the cold source station at different times, such as the number of chiller units started and stopped, operating load rate, flow rate and head setpoint of chilled water pumps and cooling water pumps, and speed of cooling tower fans. In some embodiments, after determining the pre-control parameters, the control processing device of the cold source station can automatically adjust the control parameters of each device according to the pre-control parameters, adjust the operating status of each device, and realize the forward-looking and refined control of the cold source station, thereby improving energy utilization efficiency while meeting the cooling demand of the plant area.

[0068] By controlling the cooling station based on pre-controlled parameters, the main unit and supporting equipment of the cooling station can be adjusted according to the preset operating strategy. This allows the cooling station to operate proactively according to the predicted load, reducing the occurrence of over-cooling or under-cooling due to load fluctuations. It achieves matching of cooling supply and load demand, thereby improving the accuracy of load forecasting and effectively achieving energy conservation, emission reduction, and carbon emission reduction.

[0069] Figure 3 This is a schematic diagram of an optional process for step S22 in one embodiment. For example... Figure 3 As shown, in some other embodiments of the method, the typical load model is trained based on historical meteorological data of the plant area and historical load data of the cooling station, including:

[0070] S30: Acquire historical meteorological data and seasonal load data of the plant area, wherein the historical meteorological data includes outdoor dry-bulb temperature and outdoor wet-bulb temperature;

[0071] S32: Convert the outdoor dry-bulb temperature and outdoor wet-bulb temperature into the average dry-bulb temperature and the average wet-bulb temperature of the first preset period;

[0072] S34: Cluster the average dry-bulb temperature and the average wet-bulb temperature of the first preset cycle to obtain M typical load intervals of the plant area, where M>1;

[0073] S36: For each typical load interval of the plant area, an initial typical load model is constructed. The outdoor dry-bulb temperature and outdoor wet-bulb temperature corresponding to the typical load interval of the plant area are used as input features, and the load of the first preset period is used as output features. The initial typical load model corresponding to the typical load interval of the plant area is trained to obtain the trained typical load model.

[0074] In this embodiment, historical meteorological data of the district and seasonal load data of the plant area can be obtained. The seasonal load data of the plant area includes the load data of the cold source station under the meteorological conditions corresponding to the historical meteorological year data. The seasonality or preset period mentioned in some embodiments of this disclosure can usually refer to the change period of environmental factors that affect the load change of the cold source station, such as 24 hours, where the hourly changes throughout the day have a significant impact on the load (morning and evening), and exhibit periodic changes throughout the day (e.g., large temperature differences between morning and evening).

[0075] In some embodiments of this disclosure, the historical meteorological data may include meteorological data from past periods at the cold source station, with the time span being year, quarter, or other set time spans. The resolution of the meteorological data may be half-hour, hour, day, month, quarter, etc. For example, in some application scenarios of this disclosure, the historical annual meteorological data of the cold source station is a record of annual meteorological data for the location of the cold source station, such as hourly, daily, and monthly meteorological parameters for the past 1 year, 2 years, or even longer, specifically including detailed data such as hourly dry-bulb temperature, wet-bulb temperature, relative humidity, atmospheric pressure, precipitation, wind direction, and wind speed. This historical annual meteorological data can comprehensively reflect the long-term climate characteristics and seasonal variation patterns of the region, providing rich meteorological background information for the training and establishment of subsequent target load models. In this embodiment, the historical meteorological data includes outdoor dry-bulb temperature and outdoor wet-bulb temperature.

[0076] The acquired hourly outdoor dry-bulb and wet-bulb temperatures are periodically statistically processed and converted into average dry-bulb and wet-bulb temperatures for a first preset period. This reduces the interference of short-term temperature fluctuations on model training and makes the input features better reflect the overall meteorological variation patterns. The average dry-bulb and wet-bulb temperatures corresponding to the first preset period are then clustered, grouping data with similar meteorological characteristics and load patterns into one class, resulting in M ​​typical load intervals for the plant area, where M is an integer greater than 1. In some application scenarios, K-means clustering can be used; the number of clusters can be preset or automatically determined by the clustering algorithm.

[0077] In this way, each typical load range of the plant area corresponds to a relatively stable relationship between weather and load.

[0078] For each typical load range of the plant area, an initial typical load model is constructed. The outdoor dry-bulb temperature and outdoor wet-bulb temperature corresponding to the typical load range of the plant area are used as the input features of the model, and the cold source load corresponding to the first preset cycle is used as the output feature of the model. The initial typical load models are trained using historical data within the corresponding range, thereby obtaining a trained typical load model applicable to different temperature parameter ranges.

[0079] The target meteorological data from historical meteorological year data is divided into a target number of data sets according to a preset parameter range. Each data set and its corresponding cold source load are used to train an initial typical model to obtain a corresponding typical load model. In one specific embodiment, for the initial typical load model A1, the data set D1 corresponding to the typical load interval of the initial typical model A1 is used for training. The outdoor dry-bulb temperature and outdoor wet-bulb temperature of the training data set D1 are used as input features, and the load of the first preset period is used as the output feature. After training, typical load model A is obtained. Similarly, the initial typical load model B1 is trained using another typical load interval data set D2 to obtain typical load model B, and so on, to obtain various typical load models, providing a model basis for selecting the target load model based on the temperature parameter range of the target meteorological data.

[0080] In a specific implementation example, a backpropagation neural network (BPNN) can be used to train an initial typical load model. For example, 16 typical load intervals for each plant area can be divided by co-clustering, constructing 16 typical load models SAM1-SAM16. The corresponding typical load model for each plant area's typical interval can then be trained using the datasets associated with those intervals. In some embodiments of this disclosure, the BPNN can employ a 3-layer neural network with 280 neurons per layer, an optimizer such as Adam, a loss function such as MSE, and 1000 iterations.

[0081] As mentioned above, in some other embodiments of this disclosure, the first preset period is 1 day, and the input features of the initial typical load model are hourly dry-bulb temperature and hourly wet-bulb temperature. Thus, this embodiment uses a 1-day prediction period to fully cover the typical daily load variation patterns of the plant area. Using hourly dry-bulb temperature and hourly wet-bulb temperature as input features can capture the hourly fluctuations in cooling load due to changes in meteorological conditions, enabling the typical load model to more accurately learn the hourly mapping relationship between meteorological parameters and cooling demand, thereby improving the precision of the first target load prediction.

[0082] In some application scenarios, these seasonal load data of the plant area can be directly collected by the energy metering device of the cold source station, such as the cumulative operating cooling capacity of the chiller unit, the instantaneous cooling capacity calculated by the temperature difference and flow rate of the chilled water supply and return water, etc. The time granularity can be consistent with the historical meteorological data and plant area data, such as recording the cold source load value once an hour, so as to carry out subsequent data matching and model training.

[0083] In some embodiments, historical meteorological data corresponding to the plant area data can also be obtained, and the seasonal load of the cold source station corresponding to the historical meteorological data and the plant area data can be obtained. That is, the load data of the cold source station generated during actual operation under the meteorological conditions corresponding to the historical meteorological data and the actual conditions of the plant area reflected by the plant area data. The plant area data can refer to various relevant data of the plant area itself under the corresponding historical meteorological conditions. These data directly or indirectly affect the cold source load of the cold source station. Specifically, the plant area data may include, but is not limited to, the building structure parameters of each building in the plant area (such as building area, building height, thermal conductivity of wall materials, heat transfer coefficient of roof, window type and area, etc.), the number of personnel and work and rest patterns of each functional area (production workshop, office, laboratory, etc.), the operating parameters of production equipment (such as equipment power, running time, heat dissipation, etc.), the specific requirements of the production process (such as the ambient temperature and humidity range required for specific production links), the power and running time of the lighting system, and the ventilation rate in the plant area, etc.

[0084] In other application scenarios, historical meteorological data and plant area data can be input into the energy consumption analysis software DOE-2 (a white-box model of building thermophysics) to obtain corresponding seasonal load data of the plant area through simulation calculations. DOE-2 is a building energy consumption simulation software mainly used to predict the energy consumption and related costs of various buildings under hourly conditions throughout the year. In some embodiments of the method provided in this disclosure, obtaining historical meteorological data and seasonal load data of the plant area includes:

[0085] S302: Using a white-box model based on building thermophysical mechanisms, meteorological annual data is loaded to obtain the periodic load of each workshop. The meteorological annual data includes typical meteorological annual data, historical 1-year meteorological annual data, and historical 2-year meteorological annual data. The meteorological annual data includes meteorological data of the area where the factory is located, factory building data, factory personnel data, and factory equipment load data.

[0086] S304: The seasonal load data of the plant area is obtained by adding up the periodic loads of each workshop.

[0087] In the specific implementation process, a white-box model based on building thermophysics mechanisms can be used to load annual meteorological data to calculate the periodic load of each workshop. The annual meteorological data includes typical meteorological year data, historical 1-year meteorological year data, and historical 2-year meteorological year data. The annual meteorological data can include meteorological data of the plant area, plant building data, plant personnel data, and plant equipment load data. The typical meteorological year data is obtained by statistically analyzing nearly 30 years of meteorological data (including dry-bulb temperature, wet-bulb temperature, relative humidity, wind speed, solar radiation, etc.) of the cold source station location to select representative year meteorological data. This data reflects the long-term average climate conditions of the area and reduces data bias caused by extreme weather conditions in a single year. The historical 1-year meteorological year data is the actual meteorological observation data of the most recent complete year, reflecting recent meteorological trends. The historical 2-year meteorological year data is the actual meteorological observation data of the most recent two consecutive years, used to further enrich the meteorological data sample and capture meteorological fluctuation characteristics over a longer period. The DOE-2 was used to simulate the heating and cooling loads of each workshop under different weather conditions hourly, and the periodic load data of each workshop was obtained.

[0088] The periodic load data of each workshop are summed and aggregated to obtain the overall seasonal load data of the plant area. This data can reflect the overall cooling demand pattern of the plant area under different periodic time periods and different meteorological conditions.

[0089] This embodiment calculates the seasonal load of the plant area using DOE-2, effectively avoiding the load data distortion problems caused by abnormal operating conditions such as sensor failure, equipment start-up and shutdown fluctuations, and transmission noise in traditional methods. It eliminates the interference of abnormal load points on subsequent model training from the source. Simultaneously, by combining typical meteorological year data with multi-year historical meteorological data, it takes into account both long-term meteorological patterns and short-term meteorological changes, making the calculated seasonal load data more closely match the actual cooling demand of the plant area. This provides high-quality, unbiased benchmark data support for the training of subsequent typical load models, thereby improving the overall accuracy and stability of load forecasting and providing a reliable data foundation for the forward-looking control of the cooling plant.

[0090] Figure 4 This is a schematic flowchart of an optional process for controlling a factory cold source station in one embodiment. After determining the pre-control parameters of the cold source station for the first preset period based on the first target load, the method further includes:

[0091] S40: Obtain the second target load of the cold source station in the second preset period of the future; wherein, the second preset period is shorter than the first preset period;

[0092] S42: If the second target load is inconsistent with the load data corresponding to the first target load, the pre-control parameters are corrected by the second target load to obtain the corrected pre-control parameters. The control of the cold source station based on the pre-control parameters includes controlling the cold source station based on the corrected pre-control parameter table.

[0093] In some embodiments provided in this disclosure, the first target load can be a single data point representing the load status of the cooling station within a future first preset period, and can be a pre-control parameter determined based on the first target load of the first preset period. In some embodiments, the first target load can also be multiple load data points divided into a second preset period under the first preset period, and each load data point can have pre-control parameters corresponding to a time period. For example, if the first target load is a 24-hour load, the overall pre-control parameters of the cooling station for the next 24 hours are determined based on the 24-hour load. Alternatively, the first target load can be a 24-hour load with corresponding load data for each time period, in which case the pre-control parameters can be hourly pre-control parameters determined based on the load data for each time period. In the embodiments of this disclosure, after determining the pre-control parameters of the cooling station for the first preset period based on the first target load, a second target load of the cooling station within a future second preset period can also be obtained.

[0094] The second preset period is a shorter time period than the first preset period. Its purpose is to allow for more immediate and dynamic adjustments to the operation of the chiller station to cope with potential short-term load fluctuations. For example, if the first preset period is 24 hours, the second preset period can be set to 1 hour, 2 hours, or 4 hours. The second target load refers to the actual cooling capacity demand that the chiller station needs to meet within this shorter second preset period. Compared to the first target load, the second target load has a shorter prediction period and requires higher real-time prediction accuracy to quickly respond to subtle changes in load. The method for obtaining the second target load can be similar to that for obtaining the first target load, but may employ a prediction method that focuses more on recent data.

[0095] The second target load represents the actual cooling demand required by the cooling station in the short term. Its prediction process focuses more on the real-time operating status of the plant and short-term production changes, thus exhibiting higher real-time performance and sensitivity compared to the first target load. This embodiment sets a second preset period and calculates the target load for that period. The second preset period reflects the real-time changing trend of the plant's cooling load over a shorter timeframe, capturing instantaneous load changes caused by factors such as production rhythm and environmental fluctuations, thereby supplementing and calibrating the first target load obtained based on long-term meteorological forecasts.

[0096] In this embodiment, the second target load corresponding to the second preset period is compared with the load data of the first target load for the same time period. If the two are inconsistent, it indicates that there is a deviation between the short-term actual load demand and the long-term predicted load. In this case, the second target load is used to correct the pre-control parameters to obtain corrected pre-control parameters that better fit the current actual operating conditions. In some embodiments, the second target load of one second preset period after the current time node can be obtained and compared with the load of the time node (or interval) corresponding to the first target load. Alternatively, multiple second target loads can be obtained and compared with the corresponding time nodes (or intervals) of the first target load. In some embodiments of this disclosure, the inconsistency between the second target load and the load data corresponding to the first target load may include actual differences in the load data of the two, or the data deviation between the two may exceed a preset range, such as the difference between the two being greater than a preset difference or the ratio between the two being greater than a preset ratio difference.

[0097] For example, if the first preset cycle is 24 hours, the average first target load during the first 1-3 hours is 1000 RT, while the average second target load obtained through the second preset cycle (e.g., the first 1-3 hours) is 1200 RT, this is considered inconsistent. In this case, the pre-control parameter table is corrected using the second target load to obtain a corrected pre-control parameter table, and the chiller station is controlled based on the corrected pre-control parameter table.

[0098] The specific correction method for adjusting the pre-control parameters based on the second target load can be preset and can include various implementation methods. For example, a new pre-control parameter can be determined based on the second target load, and the pre-control parameter for the corresponding time period in the first target load can be replaced with the new pre-control parameter. Of course, in other embodiments, the average value of the first target load and the second target load can be taken, or one of them can be used to meet energy-saving conditions or stable operation guarantee conditions, or other preset methods can be used for correction.

[0099] In one implementation scenario of the aforementioned correction method, an implementation method may also be included in which modifications are made to some time periods or some parameters while simultaneously modifying other time periods or parameters.

[0100] This embodiment introduces a shorter second preset period for predicting the second target load and uses this second target load to correct the pre-control parameter table generated based on the first target load. Its innovation lies in constructing a two-layer load decision-making mechanism of long-cycle load prediction and short-cycle load calibration, rather than relying solely on a single long-cycle prediction result for control. This approach effectively solves the technical problems of relying solely on long-cycle weather forecasts being unable to adapt to real-time fluctuations in plant production and the decrease in control accuracy due to the accumulation of load deviations over time. It also reduces the phenomenon of insufficient or excessive cooling caused by the lag of a single prediction model. This embodiment's solution, through short-time, high-precision load calibration of the long-cycle pre-control parameter table, allows the cooling station's operation strategy to better match actual load demand, improves the matching degree between cooling supply and load demand, further enhances control accuracy and energy-saving effects, and ensures more stable, efficient, and reliable operation of the cooling station.

[0101] In some other embodiments of the plant cold source station control method provided in this disclosure, the step of obtaining the second target load of the cold source station for a future second preset period includes:

[0102] S402: Obtain a local load feature set constructed based on local load association data, input the local load feature set into the constructed local load model, and obtain the predicted second target load of the cold source station for the second preset period. The local load association data includes the historical local load of the cold source station acquired at the second preset period as the collection interval, the historical production cycle information corresponding to the historical local load, and the first target load of the first preset period. The local load association data is spliced ​​together to obtain the local load feature set.

[0103] Acquire local load correlation data for constructing short-term load forecasts. The local load correlation data includes three types of information: first, the historical local load of the cold source station collected in real time at a second preset period; second, the historical production cycle information corresponding to the time dimension of the historical local load; and third, the first target load of the first preset period that has been predicted.

[0104] In some embodiments, the aforementioned historical local load, historical production cycle information, and first target load can be aligned along the time axis and then spliced ​​together to form a multi-dimensional fused local load feature set, so that the feature set can simultaneously reflect historical operating status, actual production intensity, and long-term load trends.

[0105] The completed local load feature set is input into the pre-trained local load model. The local load model learns and infers the multi-dimensional features and outputs the second target load of the cold source station in the second preset period in the future, so that the second target load is more in line with the actual cooling demand of the plant area and short-term production changes.

[0106] Historical local load can be determined based on the supply and return water temperatures and flow rates of the chiller station's outlet main pipe. For example, based on the system platform, hourly historical local load load_local(1X24XN) can be obtained according to the supply and return water temperatures and flow rates of the chiller station's outlet main pipe, where 1 represents the feature dimension, 24 represents the time dimension, and N represents the number of samples containing features such as 24 hours in a day. Historical production cycle information reflects the rhythm and intensity of production activities in the plant area, and can be reflected through production task plans, equipment operation logs, production line start-up and shutdown status, processing time of each process, and product output. For example, the historical production cycle information JPH(1X24XN) corresponding to the historical local load is obtained. The load data of the first preset period obtained above is load_season(1X24).

[0107] The historical local load, historical production cycle information, and the first target load are aligned along the time axis and then spliced ​​together to form a multi-dimensional fused local load feature set. This feature set can simultaneously reflect historical operating status, actual production intensity, and long-term load trends. For example, load_local, load_season, and JPH are spliced ​​together to obtain a local load feature set (3 x 24 x N). Of course, in other embodiments, the local load features can be further processed, such as slicing the load feature set with a time step of 8 (the working time of one shift) to obtain a new local load feature set (3 x 8 x N).

[0108] The constructed local load feature set is input into a pre-trained local load model. For example, the input features are the local load feature set (3X8XN), and the output features are the predicted load (1X1XN). Here, the predicted load can be the actual local load data of the cooling station in the second preset period (e.g., 1 hour). In this way, by learning and inferring from the multi-dimensional features through the local load model, the second target load of the cooling station in the second preset period is output, so that the second target load is more in line with the actual cooling demand of the plant and short-term production changes.

[0109] It should be understood that the historical local load of the cold source station and the historical production cycle information corresponding to the historical local load can reflect the close correlation between the cold source load and the real-time production activities of the plant. In some application scenarios, the local load model can be trained and predicted using GRU (Gated Recurrent Unit, an improved recurrent neural network) or other machine learning algorithms (such as LSTM neural network) to obtain the second target load for the second preset period in the future. For example, in some embodiments, a load prediction model can be built with Bi-LSTM (bidirectional long short-term memory) as the core. The most advanced model can adopt a 2-layer structure with 256 neurons per layer, Adam as the optimizer, MSE as the loss function, and average as the output operation (averaging and merging the hidden state vectors in the forward and backward directions to generate the final output representation).

[0110] In some embodiments, during the training of the local load model, the local load feature set can be divided into three subsets—training set, test set, and validation set—according to a preset ratio. The preset ratio can be set based on the data volume and model training requirements. For example, 70% of the subset can be used as the training set for learning model parameters, 20% as the test set for evaluating the model's generalization ability, and 10% as the validation set for adjusting model hyperparameters (such as learning rate and number of iterations) during training. During model training, the local load model is first trained iteratively using the training set. The training effect is then verified using the validation set. When the loss function value on the validation set no longer decreases significantly or shows signs of overfitting, training is stopped and the current model parameters are saved. Subsequently, the test set is used to perform a final performance evaluation of the trained model.

[0111] In some other embodiments of the plant cold source station control method provided in this disclosure, the local load model is a bidirectional long short-term memory network, and obtaining the second target load of the cold source station for a future second preset period includes:

[0112] The local load model is controlled to predict the load using a prediction period of 1 / K of the second preset period to obtain a candidate second target load, where K is a natural number greater than 1.

[0113] The second target load is determined based on K candidate second target loads within the second preset period.

[0114] In this embodiment of the disclosure, the local load model is a bidirectional long short-term memory network (Bi-LSTM). This model can fully learn the dependencies between time series data and adapt to the rhythmic characteristics of the plant load changing over time, providing a reliable model basis for short-term load forecasting.

[0115] When obtaining the second target load for the second preset period of the cooling station, the local load model can be controlled to predict the load using a prediction period of 1 / K of the second preset period, where K is a natural number greater than 1, thereby obtaining K candidate second target loads within the second preset period. For example, when the second preset period is 1 hour and K=2, the local load model performs prediction every 30 minutes, obtaining two candidate second target loads.

[0116] Based on K candidate second target loads obtained within the second preset period, the final second target load is determined comprehensively. Specifically, methods such as averaging, weighted averaging, or median filtering can be used to reduce the impact of single forecast fluctuations on the results, making the predicted second target load smoother and more stable.

[0117] In other embodiments of the method provided in this disclosure, controlling the operation of the cold source station based on the pre-control parameters includes:

[0118] During the process of controlling the cold source station through the parameters in the pre-control parameters, the adjustment range of the chilled water outlet temperature and the condensate return temperature of the cold source station host within the preset temperature range is greater than or equal to the preset adjustment step size. The adjustment step size is determined based on the control algorithm of the cold source station, the stable operating conditions of the cold source station control system, and energy consumption constraint information.

[0119] In this embodiment, a temperature adjustment step size is set. A minimum adjustment step size is set for the chilled water outlet temperature and condensate return temperature of the chiller station main unit, ensuring that the temperature adjustment is maintained within a range greater than or equal to the preset step size. The adjustment step size can be a predetermined fixed value or a dynamically adjustable value. It can be determined comprehensively based on the control algorithm characteristics of the chiller station, the stable operating conditions of the chiller station control system, and energy consumption constraints, so as to ensure that the parameter adjustment can both meet the optimization requirements and adapt to the actual execution capability of the control system.

[0120] For example, in one implementation example, the pre-set control parameters include a chilled water outlet temperature range of 6-13℃ and a condensate return temperature range of 28-35℃. In the algorithm for parameter control of the chilled water station, the temperature value is adjusted in increments of 0.5, such as to 6.5℃ or 28.5℃, rather than using continuous, unconstrained fine adjustments, thus sacrificing small changes to ensure the stability of system operation.

[0121] This embodiment of the invention effectively solves problems such as control system oscillation, frequent actuator movements, and decreased operational stability caused by excessively small or frequent parameter adjustments, by setting a reasonable adjustment step size. It also reduces increased equipment wear and energy consumption caused by repeated adjustments due to minor fluctuations, significantly improving system operational stability and reliability while ensuring the adjustment accuracy of the cold source station.

[0122] It should be understood that the various embodiments of the methods described in this specification are presented in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on its differences from other embodiments. Relevant details can be found in the descriptions of other method embodiments.

[0123] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0124] Based on the same inventive concept, this application also provides a plant cold source station control device for implementing the above-mentioned plant cold source station control method. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more plant cold source station control device embodiments provided below can be found in the limitations of the plant cold source station control method above, and will not be repeated here.

[0125] In one exemplary embodiment, such as Figure 5 As shown, a control device 500 for a factory cold source station is provided, comprising: a data acquisition module 50, a first load prediction module 52, and a control processing module 54, wherein:

[0126] Data acquisition module 50 is used to acquire target meteorological data of the factory area in the first preset period in the future, the target meteorological data including temperature data;

[0127] The first load prediction module 52 is used to determine a target load model from a number of pre-built typical load models based on the parameter range of the target meteorological data, input the target meteorological data into the target load model, and obtain the first target load of the cold source station; the typical load model is trained based on the historical meteorological data of the plant area and the historical load data of the cold source station, wherein the multiple typical load models are constructed based on the temperature data with different parameter ranges;

[0128] The control processing module 54 is used to determine the pre-control parameters of the cold source station in the first preset period based on the first target load, and control the operation of the cold source station based on the pre-control parameters.

[0129] In some embodiments of the apparatus provided in this disclosure, the typical load model is trained based on historical meteorological data of the plant area and historical load data of the cooling source station, including:

[0130] Acquire historical meteorological data and seasonal load data of the plant area, including outdoor dry-bulb temperature and outdoor wet-bulb temperature;

[0131] The outdoor dry-bulb temperature and outdoor wet-bulb temperature are converted into the average dry-bulb temperature and the average wet-bulb temperature of the first preset period, respectively.

[0132] Clustering the average dry-bulb temperature and the average wet-bulb temperature of the first preset cycle yields M typical load intervals for the plant area, where M > 1;

[0133] For each typical load range of the plant area, an initial typical load model is constructed. The outdoor dry-bulb temperature and outdoor wet-bulb temperature corresponding to the typical load range of the plant area are used as input features, and the load of the first preset period is used as output features. The initial typical load model corresponding to the typical load range of the plant area is trained to obtain the trained typical load model.

[0134] In some embodiments of the apparatus provided in this disclosure, the apparatus further includes a second load prediction module, configured to obtain a second target load for a future second preset period of the cold source station after determining the pre-control parameters of the cold source station for the first preset period based on the first target load; wherein the second preset period is shorter than the first preset period; if the second target load is inconsistent with the load data corresponding to the first target load, the pre-control parameters are corrected by the second target load to obtain corrected pre-control parameters, and controlling the operation of the cold source station based on the pre-control parameters includes controlling the operation of the cold source station based on the corrected pre-control parameter table.

[0135] In some embodiments of the apparatus provided in this disclosure, the step of obtaining the second target load of the cold source station for a future second preset period includes:

[0136] A local load feature set is obtained based on local load association data. The local load feature set is input into the constructed local load model to obtain the predicted second target load of the cold source station for the second preset period. The local load association data includes the historical local load of the cold source station obtained at the second preset period as the collection interval, the historical production cycle information corresponding to the historical local load, and the first target load of the first preset period. The local load association data is spliced ​​to obtain the local load feature set.

[0137] In some embodiments of the apparatus provided in this disclosure, the local load model is a bidirectional long short-term memory network, and obtaining the second target load of the cold source station for a future second preset period includes:

[0138] The local load model is controlled to predict the load using a prediction period of 1 / K of the second preset period to obtain a candidate second target load, where K is a natural number greater than 1.

[0139] The second target load is determined based on K candidate second target loads within the second preset period.

[0140] In some embodiments of the device provided in this disclosure, the first preset period is 1 day, and the input features of the initial typical load model are hourly dry-bulb temperature and hourly wet-bulb temperature.

[0141] In some embodiments of the apparatus provided in this disclosure, acquiring historical meteorological data and seasonal load data of the plant area includes:

[0142] By loading annual meteorological data into a white-box model based on building thermophysical mechanisms, the periodic load of each workshop is obtained. The annual meteorological data includes typical annual meteorological data, historical 1-year meteorological data, and historical 2-year meteorological data. The annual meteorological data includes meteorological data of the area where the factory is located, factory building data, factory personnel data, and factory equipment load data.

[0143] The seasonal load data of the plant area is obtained by summing the periodic loads of each workshop.

[0144] In some embodiments of the apparatus provided in this disclosure, the operation of the cold source station is controlled based on the pre-control parameters, including:

[0145] During the process of controlling the cold source station through the parameters in the pre-control parameters, the adjustment range of the chilled water outlet temperature and the condensate return temperature of the cold source station host within the preset temperature range is greater than or equal to the preset adjustment step size. The adjustment step size is determined based on the control algorithm of the cold source station, the stable operating conditions of the cold source station control system, and energy consumption constraint information.

[0146] Each module in the aforementioned device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0147] In one exemplary embodiment, an electronic device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 6 As shown. The computer device includes a processor, memory, input / output interface (I / O), and communication interface. The processor, memory, and I / O interface are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interface. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores data. The I / O interface allows the processor to exchange information with external devices. The communication interface allows communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a plant cold source station control method according to any embodiment of this disclosure.

[0148] Those skilled in the art will understand that Figure 6 The structure shown is a block diagram of a partial structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0149] In one exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments. This computer device can be... Figure 6 The electronic device shown.

[0150] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above-described method embodiments.

[0151] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0152] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program mentioned can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0153] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0154] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A control method for a factory cold source station, characterized in that, include: Acquire target meteorological data for the factory area in the first preset period in the future, the target meteorological data including temperature data; Based on the parameter range of the target meteorological data, a target load model is determined from multiple pre-built typical load models. The target meteorological data is then input into the target load model to obtain the first target load of the cold source station. The typical load model is trained based on historical meteorological data of the plant area and historical load data of the cold source station. The multiple typical load models are constructed based on the temperature data with different parameter ranges. Based on the first target load, the pre-control parameters of the cold source station are determined in the first preset cycle, and the operation of the cold source station is controlled based on the pre-control parameters.

2. The plant source station control method according to claim 1, characterized by, The typical load model is trained based on historical meteorological data of the plant area and historical load data of the cooling station, and includes: Acquire historical meteorological data and seasonal load data of the plant area, including outdoor dry-bulb temperature and outdoor wet-bulb temperature; The outdoor dry-bulb temperature and outdoor wet-bulb temperature are converted into the average dry-bulb temperature and the average wet-bulb temperature of the first preset period, respectively. Clustering the average dry-bulb temperature and the average wet-bulb temperature of the first preset cycle yields M typical load intervals for the plant area, where M > 1; For each typical load range of the plant area, an initial typical load model is constructed. The outdoor dry-bulb temperature and outdoor wet-bulb temperature corresponding to the typical load range of the plant area are used as input features, and the load of the first preset period is used as output features. The initial typical load model corresponding to the typical load range of the plant area is trained to obtain the trained typical load model.

3. The plant chiller plant control method of claim 1, wherein, After determining the pre-control parameters of the cooling station for the first preset period based on the first target load, the method further includes: Obtain the second target load of the cold source station for a future second preset period; wherein the second preset period is shorter than the first preset period; If the second target load is inconsistent with the load data corresponding to the first target load, the pre-control parameters are corrected by the second target load to obtain the corrected pre-control parameters. The control of the cold source station based on the pre-control parameters includes controlling the cold source station based on the corrected pre-control parameter table.

4. The plant source station control method according to claim 3, characterized by, The step of obtaining the second target load for the second preset period of the future cooling station includes: A local load feature set is obtained based on local load association data. The local load feature set is input into the constructed local load model to obtain the predicted second target load of the cold source station for the second preset period. The local load association data includes the historical local load of the cold source station obtained at the second preset period as the collection interval, the historical production cycle information corresponding to the historical local load, and the first target load of the first preset period. The local load association data is spliced ​​to obtain the local load feature set.

5. The plant chiller plant control method of claim 4, wherein, The local load model is a bidirectional long short-term memory network. Obtaining the second target load of the cold source station for the next second preset period includes: The local load model is controlled to predict the load using a prediction period of 1 / K of the second preset period to obtain a candidate second target load, where K is a natural number greater than 1. The second target load is determined based on K candidate second target loads within the second preset period.

6. The plant chiller plant control method of claim 2, wherein, The first preset period is 1 day, and the input features of the initial typical load model are hourly dry-bulb temperature and hourly wet-bulb temperature.

7. The plant chiller plant control method of claim 2, wherein, The acquisition of historical meteorological data and seasonal load data of the plant area includes: By loading annual meteorological data into a white-box model based on building thermophysical mechanisms, the periodic load of each workshop is obtained. The annual meteorological data includes typical annual meteorological data, historical 1-year meteorological data, and historical 2-year meteorological data. The annual meteorological data includes meteorological data of the area where the factory is located, factory building data, factory personnel data, and factory equipment load data. The seasonal load data of the plant area is obtained by summing the periodic loads of each workshop.

8. The plant chiller plant control method of claim 1, wherein, Controlling the operation of the cold source station based on the pre-control parameters includes: During the process of controlling the cold source station through the parameters in the pre-control parameters, the adjustment range of the chilled water outlet temperature and the condensate return temperature of the cold source station host within the preset temperature range is greater than or equal to the preset adjustment step size. The adjustment step size is determined based on the control algorithm of the cold source station, the stable operating conditions of the cold source station control system, and energy consumption constraint information.

9. A plant chiller station control device characterized by comprising: include: The module comprises a data acquisition module, a first load forecasting module, and a control processing module, wherein: The data acquisition module is used to acquire target meteorological data of the factory area in the first preset period in the future, and the target meteorological data includes temperature data; The first load forecasting module is used to determine a target load model from a number of pre-built typical load models based on the parameter range of the target meteorological data, and input the target meteorological data into the target load model to obtain the first target load of the cold source station; the typical load model is trained based on the historical meteorological data of the plant area and the historical load data of the cold source station, wherein the multiple typical load models are constructed based on the temperature data with different parameter ranges; The control processing module is used to determine the pre-control parameters of the cold source station in the first preset period based on the first target load, and control the operation of the cold source station based on the pre-control parameters.

10. A computer device, comprising: include: A memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the control method for a plant cold source station according to any one of claims 1-8.

Images