Prediction method and device for power regulation space of cold station and electronic equipment

Abstract

The invention relates to the technical field of cold station regulation and control, artificial intelligence and the like, and discloses a prediction method and device for a power regulation space of a cold station and electronic equipment, and the method comprises the steps: obtaining future weather data, chilled water temperature data and chilled water flow data of the cold station; on the basis of the future weather data, the chilled water temperature data and the chilled water flow data, demand cooling capacity data of the cold station is predicted; obtaining cold station power data based on the required cold capacity data and the system efficiency data of the cold station; and obtaining a power regulation space of the cold station based on the cold station power data. Through the method, the prediction accuracy of the power regulation space of the cold station can be improved, and the utilization rate of energy is effectively improved on the premise that the refrigeration requirement is guaranteed.

Description

Technical Field

[0001] This application relates to the technical fields of cold station regulation, artificial intelligence, etc., and particularly relates to a method, device, and electronic device for predicting the power adjustment space of a cold station. Background Art

[0002] With the continuous increase in the demand for air-conditioning refrigeration in large public buildings and industrial production, the proportion of chiller units and their auxiliary equipment (cooling towers, water pumps, etc.) in the total energy consumption of the air-conditioning system has been continuously rising. Parameters such as the temperature and flow rate on the chilled water side and the cooling water side directly determine the operating load and operating efficiency of the units.

[0003] However, in actual engineering, the regulation of the cold station system power is affected by various complex factors, including weather data (such as ambient temperature, humidity, air pressure, etc.), building load fluctuations (such as personnel density, equipment heat dissipation, and diurnal cycle changes), the system efficiency of the cold station (such as the efficiency of cooling towers and water pumps, chiller efficiency), and the characteristics of the units themselves (such as part-load characteristics, start-stop strategies). The dynamic changes of these factors make the regulation requirements of the cold station power complex and variable. If the adjustable power space of the entire cold station cannot be evaluated in real time and accurately, the energy-saving potential cannot be fully exploited. Therefore, there is an urgent need for a method for predicting the power adjustment space of a cold station, which can predict and evaluate the accuracy of the power regulation space of chillers and supporting facilities in real time, and improve the energy utilization efficiency on the premise of ensuring the refrigeration demand. Summary of the Invention

[0004] The embodiments of this application aim to at least solve one of the technical problems in the related art to some extent. For this purpose, the embodiments of this application propose a method, device, electronic device, computer product, and medium for predicting the power adjustment space of a cold station.

[0005] The embodiments of this application provide a method for predicting the power adjustment space of a cold station. The method includes: obtaining the future weather data, chilled water temperature data, and chilled water flow rate data of the cold station; predicting the required cooling capacity data of the cold station based on the future weather data, chilled water temperature data, and chilled water flow rate data; obtaining the cold station power data based on the required cooling capacity data and the system efficiency data of the cold station; and obtaining the power adjustment space of the cold station based on the cold station power data.

[0006] In some embodiments, the required cooling capacity data includes the maximum required cooling capacity and the minimum required cooling capacity; the cold station power data includes the maximum cold station power and the minimum cold station power; obtaining the cold station power data based on the required cooling capacity data and the system efficiency data of the cold station includes: obtaining the maximum cold station power based on the maximum required cooling capacity and the system efficiency data; obtaining the minimum cold station power based on the minimum required cooling capacity and the system efficiency data; wherein, obtaining the power adjustment space of the cold station based on the cold station power data includes: determining the difference between the maximum cold station power and the minimum cold station power as the power adjustment space.

[0007] In some embodiments, the method further includes: obtaining the system efficiency data of the cold station; obtaining the system efficiency data of the cold station includes: obtaining historical required cooling capacity data and cooling water temperature data; querying a table based on the historical required cooling capacity data to obtain initial system efficiency data, wherein the table includes multiple system efficiency values corresponding to multiple cooling capacity gradient values; adjusting the initial system efficiency data based on future weather data and cooling water temperature data to obtain the system efficiency data.

[0008] In some embodiments, adjusting the initial system efficiency data based on future weather data and cooling water temperature data to obtain the system efficiency data includes: inputting the future weather data into a cooling water temperature prediction model to obtain cooling water temperature data; inputting the cooling water temperature data into a system efficiency prediction model to obtain adjustment parameters; adjusting the initial system efficiency data based on the adjustment parameters to obtain the system efficiency data.

[0009] In some embodiments, predicting the required cooling capacity data of the cold station based on future weather data, chilled water temperature data, and chilled water flow data includes: inputting the future weather data, chilled water temperature data, and chilled water flow data into a cooling capacity prediction model for prediction, and outputting the required cooling capacity data of the cold station; the cooling capacity prediction model is trained in the following manner: obtaining first sample data, wherein the first sample data includes weather sample data, chilled water supply temperature sample data, chilled water return temperature sample data, and chilled water flow sample data; training the cooling capacity prediction model based on the first sample data.

[0010] In some embodiments, the cooling capacity prediction model includes a reinforcement learning model; training the cooling capacity prediction model based on the first sample data includes: constructing a state space based on the first sample data; making a prediction based on the state space to obtain an action space, wherein the action space includes the predicted required cooling capacity data; determining a reward parameter for the action space based on a reward function, wherein the reward parameter is associated with the chilled water return temperature; updating the model parameters of the reinforcement learning model based on the reward parameter to obtain a trained reinforcement learning model.

[0011] In some embodiments, based on the reward parameter, the model parameters of the reinforcement learning model are updated to obtain a trained reinforcement learning model, including: caching the current state, the required cooling capacity data corresponding to the current action, the current reward parameter, and the next state; sampling from the cached data, and updating the model parameters of the reinforcement learning model according to the temporal difference error method.

[0012] In some embodiments, the chilled water return temperature associated with the reward parameter is obtained by: inputting the weather sample data of the chilled water station, the chilled water supply temperature sample data, and the chilled water flow sample data into an environmental simulator for prediction to obtain the chilled water return temperature.

[0013] In some embodiments, the cooling water temperature prediction model includes a multi-layer fully connected neural network, and the cooling water temperature model is trained by: obtaining the second sample data and the cooling water temperature label corresponding to the second sample data, where the second sample data includes historical outdoor temperature, historical outdoor humidity, historical wind speed, and historical solar radiation; processing the cooling water temperature sample data using a multi-layer fully connected neural network to obtain the cooling water temperature data; determining the error between the cooling water temperature prediction data and the cooling water temperature label by the root mean square error method, and adjusting the model parameters of the cooling water temperature prediction model based on the error to obtain a trained cooling water temperature prediction model.

[0014] In some embodiments, the system efficiency prediction model includes a multi-layer fully connected neural network, and the system efficiency prediction model is trained by: obtaining the third sample data and the system efficiency label corresponding to the third sample data, where the third sample data includes historical cooling water temperature data; processing the third sample data using a multi-layer fully connected neural network to obtain the system efficiency prediction data; determining the error between the system efficiency prediction data and the system efficiency label by the root mean square error method, and adjusting the model parameters of the system efficiency model based on the error to obtain a trained system efficiency prediction model.

[0015] An embodiment of the present application provides a prediction device for the power adjustment space of a chilled water station. The device includes: an acquisition module for acquiring the future weather data, chilled water temperature data, and chilled water flow data of the chilled water station; a prediction module for predicting the required cooling capacity data of the chilled water station based on the future weather data, chilled water temperature data, and chilled water flow data; a first acquisition module for obtaining the chilled water station power data based on the required cooling capacity data and the system efficiency data; and a second acquisition module for obtaining the power adjustment space of the chilled water station based on the chilled water station power data.

[0016] Embodiments of the present application provide an electronic device, which includes: a memory, and one or more processors communicatively connected to the memory; instructions executable by the one or more processors are stored in the memory, and when the instructions are executed by the one or more processors, the one or more processors are caused to implement the steps of the method according to any one of the above embodiments.

[0017] Embodiments of the present application provide a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method according to any one of the above embodiments are implemented.

[0018] Embodiments of the present application provide a computer program product, including a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of the above embodiments are implemented.

[0019] Through the prediction method for the power regulation space of the cold station provided by the present application, multi-feature modeling and algorithm prediction can be combined. For example, future weather data from multiple platforms, chilled water temperature data, and demand cooling capacity data predicted by a reinforcement learning model are fused to achieve dynamic evaluation of the power space of the cold station and accurately quantify the upper and lower limits of the adjustable power of the cold station. Based on the boundary calculation of system efficiency data and cold station power data, the scheduling accuracy of the cold station power can be improved, providing an adjustable power range for the chillers and related auxiliary equipment in the cold station system to avoid blind startup or shutdown. At the same time, through the collaborative optimization of demand cooling capacity data and cold station power extreme values, the energy-saving potential is maximized. On the premise of ensuring load demand satisfaction, the unit load distribution and operation strategy are optimized, effectively reducing the overall energy consumption of the cold station system. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic flowchart of a prediction method for the power regulation space of a cold station provided by an embodiment of the present application;

[0021] Figure 2 It is a schematic flowchart of an example of a prediction method for the power regulation space of a cold station provided by an embodiment of the present application;

[0022] Figure 3 It is a schematic diagram of a prediction device for the power regulation space of a cold station provided by an embodiment of the present application;

[0023] Figure 4 It is a block diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] Embodiments of the present application will be described in detail below. Examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present application, and should not be construed as limiting the present application.

[0025] With the continuous increase in the demand for air-conditioning refrigeration in large public buildings and industrial production, the proportion of chiller units and their auxiliary equipment (cooling towers, water pumps, etc.) in the total energy consumption of the air-conditioning system has been continuously rising. Parameters such as temperature and flow rate on the chilled water side and the cooling water side directly determine the operating load and operating efficiency of the units.

[0026] However, in actual engineering, the regulation of the cold station system power is affected by various complex factors, including weather data (such as ambient temperature, humidity, air pressure, etc.), building load fluctuations (such as personnel density, equipment heat dissipation, and diurnal cycle changes), the system efficiency of the cold station (such as the efficiency of cooling towers and water pumps, chiller efficiency), and other characteristics of the units themselves (such as part-load characteristics, start-stop strategies, etc.). The dynamic changes of these factors make the regulation demand of the cold station power complex and variable, and the usual scheduling methods are difficult to evaluate the overall adjustable power space of the cold station in real time and accurately, resulting in the inability to fully exploit the energy-saving potential.

[0027] In the usual operation scheduling, operation and maintenance personnel mainly rely on existing experience or simple rules, such as "front priority" or "alternate switching", etc. to regulate the operation of chiller units and auxiliary equipment. This extensive scheduling mode lacks the prediction of the power regulation space, is difficult to optimize the operation state of the units in real time, often leads to poor operation efficiency of the units under part-load conditions, and even over-regulation or under-regulation occurs, affecting the energy utilization efficiency. In addition, the usual scheduling methods have low prediction accuracy for the power regulation space and cannot comprehensively consider the mutual influence and dynamic characteristics of the various components of the cold station system, resulting in insufficient accuracy and flexibility of power regulation.

[0028] In summary, in view of the defects existing in the usual scheduling methods for cold station systems, the present application provides a method for predicting the power regulation space of a cold station. This method can combine multi-feature modeling and algorithm prediction to quantitatively evaluate the power regulation space of the cold station.

[0029] Figure 1 It is a schematic flow diagram of a method for predicting the power regulation space of a cold station provided by an embodiment of the present application.

[0030] As Figure 1 shown, an embodiment of the present application provides a method 100 for predicting the power regulation space of a cold station. The method 100 includes:

[0031] Step 110: Obtain the future weather data, chilled water temperature data, and chilled water flow rate data of the cold station.

[0032] Exemplarily, the obtained future weather data of the cold station may include, for example, the weather forecast data every 15 minutes for the next 24 hours based on multiple platforms, such as outdoor temperature, humidity, wind speed, and solar radiation intensity. For the weather forecast data of multiple platforms, for example, the Kalman filter can be used to fuse the data of multiple platforms to improve the reliability of the data. Specifically, the prediction variances of each prediction index data of each platform in the past 24 hours can be calculated:

[0033]

[0034] where r j is the actual weather value such as temperature, humidity, etc., and p j is the predicted weather value such as predicted temperature, humidity, etc., is the prediction variance; here, N is generally 96.

[0035] Based on the different indexes of the prediction variance (outdoor temperature, humidity, wind speed, solar radiation intensity), select the data of different platforms, that is, select the corresponding platform index data with a low prediction variance.

[0036] Exemplarily, the chilled water temperature data may include the data designed by the system. Deeply fuse the weather forecast data with the operation data of the cold station to improve the prediction accuracy of the cooling load

[0037] Step 120: Predict the required cooling capacity data of the cold station based on the future weather data, chilled water temperature data, and chilled water flow rate data.

[0038] Exemplarily, the required cooling capacity data of the cold station can be predicted through a reinforcement learning model. Specifically, under the condition of given future weather data and the temperature of the terminal equipment of the cold station, for example, the future weather data is obtained based on Step 110, and the temperature of the terminal equipment can be the full-load opening state of the terminal equipment, so as to predict the required cooling capacity data.

[0039] Step 130: Obtain the cold station power data based on the required cooling capacity data and the system efficiency data of the cold station.

[0040] Exemplarily, the obtained cold station power data may include the maximum cold station power and the minimum cold station power.

[0041] Step 140: Obtain the power adjustment space of the cold station based on the cold station power data.

[0042] Exemplarily, the difference between the maximum cold station power and the minimum cold station power obtained in Step 130 can be the power adjustment space of the cold station.

[0043] Through the prediction method of the power regulation space of the cold station provided in this application, multi - feature modeling and algorithm prediction can be combined. For example, by fusing future weather data from multiple platforms, chilled water temperature data, and using a reinforcement learning model to predict the required cooling capacity data, the dynamic evaluation of the power space of the cold station can be realized, and the upper and lower limits of the adjustable power of the cold station can be accurately quantified. Based on the boundary calculation of the system efficiency data and the cold station power data, the scheduling accuracy of the cold station power can be improved, providing an adjustable power range for the chillers and related auxiliary equipment in the cold station system, and avoiding blind startup or shutdown. At the same time, through the collaborative optimization of the required cooling capacity data and the extreme values of the cold station power, the energy - saving potential can be maximized. On the premise of meeting the load demand, the unit load distribution and operation strategy can be optimized, effectively reducing the overall energy consumption of the cold station system.

[0044] In another embodiment of this application, the required cooling capacity data includes the maximum required cooling capacity and the minimum required cooling capacity; the cold station power data includes the maximum cold station power and the minimum cold station power. Obtaining the cold station power data based on the required cooling capacity data and the system efficiency data of the cold station includes: obtaining the system efficiency data based on the future weather data and the chilled water temperature data; obtaining the maximum cold station power based on the maximum required cooling capacity and the system efficiency data; obtaining the minimum cold station power based on the minimum required cooling capacity and the system efficiency data.

[0045] Exemplarily, the required cooling capacity data includes two thresholds, namely the maximum required cooling capacity and the minimum required cooling capacity. Given the future weather data and in the operating state of the cold station, the maximum required cooling capacity can make the chilled water temperature closest to the lowest designed chilled water return temperature of the system. When the return water temperature is close to and not lower than this lowest chilled water return temperature, the indoor temperature can be maintained stable within the human comfort range. The minimum required cooling capacity can make the chilled water return temperature closest to the highest designed chilled water return temperature of the system. When the return water temperature is close to and not higher than this lowest chilled water return temperature, the indoor temperature can be maintained stable within the human - acceptable range. By distinguishing the maximum required cooling capacity and the minimum required cooling capacity, the power regulation requirements under different comfort requirements can be accurately met.

[0046] Exemplarily, in the process of obtaining the cold station power data based on the required cooling capacity data and the system efficiency data, since the system efficiency of the cold station is closely related to the cooling water temperature, the system efficiency data can be dynamically corrected through a neural network model. The dynamically corrected system efficiency can reflect the influence of the cooling water temperature change on the system efficiency. Thus, by combining the maximum required cooling capacity and the minimum required cooling capacity, the maximum cold station power and the minimum cold station power can be calculated more accurately.

[0047] In the embodiments of the present application, by defining the maximum demand cooling capacity and the minimum demand cooling capacity, respectively corresponding to ensuring that the return water temperature of chilled water is close to the lower limit of the system design to maintain the indoor comfortable temperature, and allowing the return water temperature to be close to the upper limit of the system design to meet the basic refrigeration demand, the cooling capacity adjustment range of the chilled water station is quantified. Considering the influence of the change of cooling water temperature on the system efficiency, a neural network model is used to dynamically correct the efficiency parameters, and the maximum chilled water station power and the minimum chilled water station power are accurately calculated. Through the embodiments of the present application, a refined prediction of the power adjustment space of the chilled water station can be realized, the unit operation strategy of the chilled water station can be optimized on the premise of ensuring user comfort, the energy efficiency can be improved and the flexible load response of the power grid can be supported, effectively solving the problems of adjustment deviation and adjustment lag caused by static estimation of system efficiency in traditional methods.

[0048] In another embodiment of the present application, the above method further includes: obtaining the system efficiency data of the chilled water station; obtaining the system efficiency data of the chilled water station, including: obtaining historical demand cooling capacity data and cooling water temperature data; based on the historical demand cooling capacity data, querying a table to obtain the initial system efficiency data, where the table includes multiple system efficiency values corresponding to multiple cooling capacity gradient values; based on the future weather data and the cooling water temperature data, adjusting the initial system efficiency data to obtain the system efficiency data.

[0049] Exemplarily, an isolation forest algorithm can be used to remove abnormal data (such as sensor anomalies, power mutations) in historical data (such as historical demand cooling capacity data). Based on the historical demand cooling capacity data, when querying the table to obtain the initial system efficiency data, the table is as follows:

[0050]

[0051]

[0052] According to the historical demand cooling capacity data and the system efficiency (COP), find the highest system efficiency value corresponding to different cooling capacities in the past 365 days. Taking 10% of the design cooling capacity as the gradient, for every 10% of the system design cooling capacity C design corresponds to a highest COP max . Generally, the demand cooling capacity data will not be lower than 20% of the system design cooling capacity. For too low cooling capacity demand, the chilled water station system should be shut down and fresh air cooling should be adopted.

[0053] After obtaining the initial system efficiency data through the above table, the system efficiency values corresponding to the maximum demand cooling capacity and the minimum demand cooling capacity are obtained by using the table lookup method. Then, the ratio of the maximum demand cooling capacity to the corresponding system efficiency value is determined as the maximum chilled water station power, and the ratio of the minimum demand cooling capacity to the corresponding system efficiency value is determined as the minimum chilled water station power.

[0054] In another example, based on future weather data and cooling water temperature data, the initial system efficiency data is adjusted to obtain system efficiency data, including: inputting the future weather data into a cooling water temperature prediction model to obtain cooling water temperature data; inputting the cooling water temperature data into a system efficiency prediction model to obtain an adjustment parameter; and adjusting the initial system efficiency data based on the adjustment parameter to obtain system efficiency data.

[0055] Exemplarily, the cooling water temperature prediction model can be a 3-layer fully connected neural network model: T cwr = f(outdoor temperature, humidity, wind speed, solar radiation), and the cooling water temperature prediction model is trained based on historical data, representing the correlation between future weather data and cooling water temperature data (cooling water return temperature), and can predict the cooling water temperature data (T cwr ) that dynamically changes in the next 24 hours. The system efficiency prediction model can be a 5-layer fully connected neural network model: COP adjust = g(T cwr ), and the system efficiency prediction model can learn the non-linear mapping between the cooling water temperature data and the system efficiency data, and output the dynamically adjusted system efficiency data (COP adjust ). Finally, the adjusted system efficiency data for the given future weather data is obtained:

[0056] COP adjust = COP max × g(f(outdoor temperature, humidity, wind speed, solar radiation))

[0057] In the embodiments of the present application, the cooling water return temperature data in the next 24 hours is dynamically predicted through the cooling water temperature prediction model, and the mapping relationship between the future weather data and the cooling water temperature data is established; the system efficiency prediction model learns the dynamic influence law of the cooling water temperature data on the system efficiency data based on the predicted cooling water temperature data, so as to generate the adjusted system efficiency data that matches the actual working conditions. The training process driven by historical data ensures that the cooling water temperature prediction model and the system efficiency prediction model fully explore the dynamic change characteristics of the weather data and the operation of the cold station, avoid the static deviation of relying on existing experience or simple rules, and finally, based on the predicted cooling water temperature data and the adjusted system efficiency data, achieve high-precision calculation of the system power data, providing reliable data support for the power adjustment space of the cold station.

[0058] In another embodiment of the present application, based on the cold station power data, the power adjustment space of the cold station is obtained, including: determining the difference between the maximum cold station power and the minimum cold station power as the power adjustment space.

[0059] Exemplarily, the calculation formula for the maximum cold station power is:

[0060] Pmax = Q max / COP adjust

[0061] The calculation formula for the minimum value of the cooling station power is:

[0062] P min = Q min / COP adjust

[0063] Wherein, P max is the maximum value of the cooling station power, P min is the maximum value of the cooling station power, Q max is the maximum value of the required cooling capacity, Q min is the minimum value of the required cooling capacity, COP adjust is the system efficiency data.

[0064] The difference between the maximum value and the minimum value of the cooling station power is determined as the power adjustment space ΔP:

[0065] ΔP = P max - P min

[0066] The power adjustment space ΔP of the cooling station can be provided to the third-party electricity price platform. Based on the real-time electricity price signal and the ΔP range, the platform dynamically formulates and issues an instruction for the target control power adapted to the current power grid demand (such as peak shaving, valley filling or demand response), guiding the cooling station to adjust the operating power within the safe adjustable range to achieve the optimization of electricity cost and the coordinated control of the power grid. Based on the physical constraints of the equipment and the reverse derivation of the cooling capacity demand, the power feasible region is realized, and the visualization expression of the energy-saving potential is achieved.

[0067] In another embodiment of the present application, based on future weather data, chilled water temperature data and chilled water flow data, the required cooling capacity data of the cooling station is predicted, including: inputting the future weather data, chilled water temperature data and chilled water flow data into the cooling capacity prediction model for prediction, and outputting the required cooling capacity data of the cooling station; the cooling capacity prediction model is trained in the following manner: obtaining first sample data, wherein the first sample data includes weather sample data, chilled water supply temperature sample data, chilled water return temperature sample data and chilled water flow sample data; training the cooling capacity prediction model based on the first sample data.

[0068] In another embodiment, the cooling capacity prediction model includes a reinforcement learning model; training the cooling capacity prediction model based on the first sample data includes: constructing a state space based on the first sample data; making a prediction based on the state space to obtain an action space, where the action space includes predicted required cooling capacity data; determining a reward parameter for the action space based on a reward function, where the reward parameter is associated with the return water temperature of the chilled water; and updating the model parameters of the reinforcement learning model based on the reward parameter to obtain a trained reinforcement learning model.

[0069] Exemplarily, the input data of the reinforcement learning model includes: historical weather data: outdoor temperature T out , humidity H out , chilled water temperature data: chilled water supply temperature T s , chilled water return temperature T r , and chilled water flow rate F. The output data of the reinforcement learning model includes: under the condition of given weather data, the maximum required cooling capacity Q max and the minimum required cooling capacity Q min . The reinforcement learning model adopts the Deep Deterministic Policy Gradient (DDPG) reinforcement learning algorithm, which is suitable for predicting continuous cooling capacity action spaces and can learn the strategy for outputting the optimal required cooling capacity data Q. The reinforcement learning model can also be a cooling load prediction module, and future weather data, chilled water temperature data, and chilled water flow rate data are input into the cooling load prediction module through an operation data input module for prediction. The static configuration data of the cooling load prediction module includes, for example, cold station design parameters: design cooling capacity, design power, chilled water supply and return temperatures, cooling water supply and return temperatures, design minimum chilled water flow rate, design maximum and minimum values of chilled water return. The training historical data of the cooling load prediction module includes, for example: outdoor temperature and humidity, cold station cooling capacity / power, system efficiency, chilled water / cooling water main pipe flow rate, pressure, temperature, and real-time equipment power. For the above data, the Isolation Forest algorithm can also be used to remove abnormal data, such as sensor anomalies and power mutations.

[0070] Specific algorithms include state space, action space, reward function, environment model, etc. The state space contains key factors affecting the return water temperature of chilled water: weather data: T out , H out , chilled water temperature data: T s , T r , F. The state is defined as:

[0071] s = (T out , H out , T supply , T return , F)

[0072] The action space is the cooling capacity Q provided by the cold station, which is a continuous value; the reward function is based on the return water temperature T of the chilled water return Design, introduce two thresholds: the lowest designed return water temperature of the chilled water in the cold station: T comfort : When the return water temperature of the chilled water is close to and not lower than this value, the room is the most comfortable. The highest designed return water temperature of the chilled water in the cold station: T acceptable : When the return water temperature of the chilled water is close to and not higher than this value, the room temperature is acceptable.

[0073] For the setting of these two design temperature thresholds, namely the lowest designed return water temperature of the chilled water in the cold station and the highest designed return water temperature of the chilled water in the cold station, in general building design documents, T comfort = 14°C, T acceptable = 18°C, which can be adjusted according to actual needs. Select an appropriate step size for the cooling capacity step ΔQ (generally select 1 - 10) to balance the prediction accuracy and calculation efficiency. Reward function tuning: Adjust the decay coefficient k and the negative reward value through experiments to ensure that the model converges to a reasonable strategy.

[0074] The definition of the reward function is:

[0075]

[0076] Among them, k is the decay coefficient, which is used to control the speed at which the reward decreases as the return water temperature of the chilled water increases. The reward gradually decreases from 1 (the most comfortable) to -1 (unacceptable).

[0077] Use a three-layer fully connected neural network trained with historical data as an environment simulator to predict the return water temperature of the chilled water: T return = f(s, a); Input of the environmental model: State s = (T out , H out , T s , F) and action a = Q. Output: Return water temperature T r .

[0078] During the training process of the cooling capacity prediction model, the first sample data can be the operation records covering 365 days, which serve as the training basis for the cooling capacity prediction model. For example, the outdoor temperature and humidity (T out , H out ) for 365 days, the parameters of the chilled water system, the supply and return water temperatures of the chilled water, and the chilled water flow rate (T s , T r , F) and the corresponding cooling capacity Q of the cold station. Then build an environmental model, train a neural network with historical data to predict the return water temperature T ro ; After training, this model serves as the environmental simulator for reinforcement learning.

[0079] First, initialize the policy network (Actor): Input the state S and the required cooling capacity data Q corresponding to the current action. Initialize the value network (Critic): Input the state S and the action Q, and output the Q value (the expected reward parameter for evaluating the action Q). Also, synchronously initialize the target networks (Target Actor and Target Critic). Then, perform experience collection. Use the Actor network to generate the cooling capacity Q and add exploration noise to enhance the exploration ability. Execute the required cooling capacity data Q in the environment simulator to obtain the return water temperature T of the chilled water r , the reward parameter r, and the next state s'. Store the current state, the required cooling capacity data corresponding to the current action, the current reward parameter, and the next state (s, Q, r, s') in the experience replay buffer. Finally, update the model parameters of the reinforcement learning model according to the temporal difference error method. Randomly sample data from the experience replay buffer. Update the Critic network: Calculate the temporal difference error (TD-error). The target value is:

[0080] y = r + γ·Q′(s′,μ′(s′)

[0081] where γ is the discount factor, and Q' and μ' are the outputs of the target network. Update the Actor network: Use policy gradient to boost the expected reward parameter. Soft-update the target network: Slowly update the target network parameters:

[0082] θ′←τθ+(1-τ)θ′

[0083] where τ is the soft-update rate.

[0084] Repeat the experience collection and network update steps until the policy converges.

[0085] In the technical solution provided by this application, the learning rates of the Actor network and the Critic network are set to 0.0001 and 0.001 respectively, and the discount factor is set to γ = 0.99. By differentiating the learning rates, the stability of policy optimization and value evaluation is balanced. By controlling the noise amplitude, the exploration ability of the action space is enhanced in the initial stage of training, and gradually converges to the optimal policy in the later stage. The Actor network and the Critic network are 3-layer fully connected networks, with 128 or 256 neurons in each layer. The shallow network (such as the 128-neuron layer) extracts the basic features of the input state (weather conditions, chilled water system parameters), and the deep network (such as the 256-neuron layer) captures the non-linear associations, and finally outputs high-precision required cooling capacity data.

[0086] In an embodiment of the present application, an enhanced learning model is trained using historical weather data, chilled water temperature data, and chilled water flow data. By combining with an environment simulator, the maximum required cooling capacity (the most comfortable required cooling capacity) and the minimum required cooling capacity (the acceptable required cooling capacity) of the chilled water station are dynamically predicted, achieving high-precision quantification of the cooling capacity adjustment range. Based on the design parameters and weather data, the operation strategy of the chilled water station is adaptively optimized to maximize energy efficiency while ensuring indoor comfort, effectively solving the problems of large prediction deviation and response lag in traditional static models.

[0087] In another embodiment of the present application, the chilled water return temperature associated with the reward parameter is obtained by: inputting the weather sample data of the chilled water station, the chilled water supply temperature sample data, and the chilled water flow sample data into the environment simulator for prediction to obtain the chilled water return temperature.

[0088] Exemplarily, an environment model is trained based on the weather sample data of the chilled water station, the chilled water supply temperature sample data, and the chilled water flow sample data. This model is used to predict the chilled water return temperature. After training, this model serves as the environment simulator of the enhanced learning model.

[0089] In another embodiment of the present application, the cooling water temperature prediction model includes a multi-layer fully connected neural network. The cooling water temperature prediction model is trained by: obtaining the second sample data and the corresponding cooling water temperature label of the second sample data, where the second sample data includes historical outdoor temperature, historical outdoor humidity, historical wind speed, and historical solar radiation; using the multi-layer fully connected neural network to process the cooling water temperature sample data to obtain the cooling water temperature prediction data; determining the error between the cooling water temperature prediction data and the cooling water temperature label by the root mean square error method, and adjusting the model parameters of the cooling water temperature prediction model based on the error to obtain the trained cooling water temperature prediction model.

[0090] Exemplarily, the cooling water temperature data is obtained through the cooling water temperature prediction model. The cooling water temperature prediction model is established using a 3-layer fully connected neural network, for example, including an input layer, a hidden layer, and an output layer. The input layer can receive the 4-dimensional features of the second sample data, such as historical outdoor temperature, historical outdoor humidity, historical wind speed, and historical solar radiation. The hidden layer is, for example, 2 layers, and the number of neurons in each layer is, for example, 128 and 64 respectively, for extracting the features of the weather data. The output layer is, for example, 1 neuron, directly outputting the prediction result of the cooling water temperature.

[0091] Exemplarily, in the training phase of the cooling water temperature prediction model, historical outdoor temperature, historical outdoor humidity, historical wind speed, historical solar radiation, etc. of the cold station in the recent period (e.g., the previous 365 days) can be obtained from the operation log of the cold station as basic training data to train the cooling water temperature prediction model. The basic training data is input through the fully connected layer of the cooling water temperature prediction model to generate the cooling water temperature in the training phase. The root mean square error can also be used to calculate the error between the second sample data and the corresponding cooling water temperature label of the second sample data.

[0092] In another embodiment of the present application, the system efficiency prediction model includes a multi-layer fully connected neural network, and the system efficiency prediction model is trained in the following way: obtaining third sample data and the corresponding system efficiency label of the third sample data, where the third sample data includes historical cooling water temperature data; processing the third sample data by using the multi-layer fully connected neural network to obtain system efficiency prediction data; determining the error between the system efficiency prediction data and the system efficiency label by the root mean square error method, and adjusting the model parameters of the system efficiency prediction model based on the error to obtain the trained system efficiency prediction model.

[0093] Exemplarily, the system efficiency prediction model is established by using a 5-layer fully connected neural network, including an input layer, a hidden layer, and an output layer. The input layer can receive 1-dimensional features of the third sample data, and the third sample data is, for example, historical cooling water temperature data. The hidden layer is, for example, 4 layers, and the number of neurons in each layer is, for example, 64, 32, 16, and 8 respectively, which is used to capture the relationship between the cooling water temperature and the system efficiency. The output layer is, for example, 1 neuron, and outputs the dynamically adjusted system efficiency data.

[0094] In the embodiment of the present application, the cooling water temperature prediction model can be trained through the association between historical weather data and the return water temperature of the cooling water, and can dynamically predict the future cooling water temperature, replacing the traditional static design value and improving the environmental adaptability. The system efficiency prediction model can be based on the prediction result of the cooling water temperature, and establish a dynamic mapping relationship between the cooling water temperature and the system efficiency, quantify the impact of environmental changes on the energy efficiency of the cold station, and provide real-time correction parameters for the power calculation of the cold station.

[0095] By the collaborative work of the two models, a relatively accurate power adjustment space of the cold station can be provided, significantly improving the energy utilization rate, reducing the operation and maintenance cost, and at the same time enhancing the adaptive ability of the cold station system to complex weather conditions.

[0096] Figure 2 It is a schematic flow chart of an example of the prediction method for the power adjustment space of the cold station provided in an embodiment of the present application.

[0097] The present application provides an example and combines with Figure 2The content is elaborated to better illustrate the content of the technical solution provided by this application:

[0098] Taking a chilled water plant in an office building as an example: Input the historical operation data of the chilled water plant for nearly 365 days to train the reinforcement learning model, and obtain the maximum demand cooling capacity and the minimum demand cooling capacity. Use Kalman filter to fuse the future weather data predicted by the next 15-minute weather forecast obtained from multi-platform data: temperature 34 degrees, humidity 43%, and the output of the reinforcement learning model for the next 15-minute minimum demand cooling capacity of 1200 kW and the maximum demand cooling capacity of 1500 kW.

[0099] Calculate that the power adjustment range of the chilled water plant for the next 15 minutes is 300 kW to 450 kW, and the power adjustment space of the chilled water plant is 150 kW. The technical effects that can be achieved are as follows: the prediction error of the reinforcement learning model ≤ 10%, and the calculation accuracy of the power adjustment space of the chilled water plant ≥ 90%.

[0100] Users can adjust the power consumption according to electricity prices and comfort requirements, or freely respond to the corresponding demands on the demand side of the power grid to obtain rewards accordingly.

[0101] Figure 3 It is a schematic diagram of a prediction device for the power adjustment space of a chilled water plant provided by an embodiment of this application.

[0102] An embodiment of this application also provides a prediction device 300 for the power adjustment space of a chilled water plant, characterized in that the device 300 includes:

[0103] An acquisition module 310, configured to acquire future weather data, chilled water temperature data, and chilled water flow data of the chilled water plant.

[0104] A prediction module 320, configured to predict the demand cooling capacity data of the chilled water plant based on the future weather data and the chilled water flow data.

[0105] A first acquisition module 330, configured to obtain the chilled water plant power data based on the demand cooling capacity data and the system efficiency data.

[0106] A second acquisition module 340, configured to obtain the power adjustment space of the chilled water plant based on the chilled water plant power data.

[0107] It can be understood that for the specific description of the power control device 300 of the chilled water plant, reference can be made to the description of the power control method 100 applied to the chilled water plant in the above text.

[0108] An embodiment of this application provides an electronic device, which includes: a memory, and one or more processors communicatively connected to the memory; instructions are stored in the memory and are executed by one or more processors, so that one or more processors implement the steps of the method according to any one of the above embodiments.

[0109] An embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method in any one of the above embodiments are implemented.

[0110] An embodiment of the present application provides a computer program product, which includes instructions. When the instructions are executed by a processor of a computer device, the computer device can execute the steps of the method in any one of the above embodiments.

[0111] Figure 4 It is a block diagram of the electronic device provided by the embodiment of the present application.

[0112] An embodiment of the present application provides an electronic device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the method in any one of the above embodiments is implemented.

[0113] As Figure 4 shown, for ease of understanding, an embodiment of the present application shows a specific electronic device 400.

[0114] The electronic device 400 is intended to represent various forms of digital computers, such as, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0115] As Figure 4 shown, the device 400 includes a computing unit 401, which can execute various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 402 or the computer program loaded from the storage unit 408 into the random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation of the electronic device 400 can also be stored. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other through a bus 404. The input / output (I / O) interface 405 is also connected to the bus 404.

[0116] Multiple components in the electronic device 400 are connected to the I / O interface 405. The multiple components include: an input unit 406, such as a keyboard, a mouse, etc.; an output unit 407, such as various types of displays, speakers, etc.; a storage unit 408, such as a disk, an optical disc, etc.; and a communication unit 409, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 409 allows the electronic device 400 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0117] The computing unit 401 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 401 executes the various methods described above. For example, in some embodiments, any one or more of the above methods can be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as the storage unit 408. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded into the RAM 403 and executed by the computing unit 401, one or more steps of any one or more of the above methods can be executed. Alternatively, in other embodiments, the computing unit 401 can be configured to execute any one or more of the above methods in any other suitable manner (e.g., by means of firmware).

[0118] Note that the logic and / or steps represented in the flowchart or described otherwise herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch instructions from and execute instructions of the instruction execution system, apparatus, or device), or used in combination with these instruction execution systems, apparatuses, or devices. For the purposes of this application, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection part with one or more wirings (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which a program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as appropriate, and then stored in a computer memory.

[0119] It should be understood that each part of this application can be implemented by hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0120] In the description of this application, the description referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of this application. In this application, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0121] In the description of the present application, it should be understood that the orientation or positional relationship indicated by terms such as "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to the present application.

[0122] In addition, the terms "first", "second", etc. used in the embodiments of the present application are only for descriptive purposes and should not be understood as indicating or implying relative importance or implicitly specifying the quantity of the technical features indicated in this embodiment. Thus, the features defined with terms such as "first", "second", etc. in the embodiments of the present application may clearly or implicitly indicate that at least one such feature is included in this embodiment. In the description of the present application, the meaning of the word "plurality" is at least two or more, such as two, three, four, etc., unless otherwise specifically defined in the embodiments.

[0123] In the present application, unless otherwise clearly specified or limited in the embodiments, the terms "mounted", "connected", "coupled" and "fixed" etc. appearing in the embodiments should be understood in a broad sense. For example, the connection can be a fixed connection, a detachable connection, or integrated. It can be understood that it can also be a mechanical connection, an electrical connection, etc.; of course, it can also be directly connected, or indirectly connected through an intermediate medium, or it can be the communication inside two elements, or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific implementation circumstances.

[0124] In the present application, unless otherwise clearly specified and limited, the first feature being "on" or "under" the second feature may be that the first and second features are in direct contact, or the first and second features are indirectly in contact through an intermediate medium. Moreover, the first feature being "above", "over" and "on top of" the second feature may be that the first feature is directly above or obliquely above the second feature, or merely indicates that the first feature has a higher horizontal height than the second feature. The first feature being "under", "beneath" and "underneath" the second feature may be that the first feature is directly below or obliquely below the second feature, or merely indicates that the first feature has a lower horizontal height than the second feature.

Claims

1. A prediction method for the power regulation space of a cold station, characterized in that, The method includes: Obtaining future weather data, chilled water temperature data, and chilled water flow rate data of the chilled water station; Predicting the required cooling capacity data of the chilled water station based on the future weather data, the chilled water temperature data, and the chilled water flow rate data; Obtaining the chilled water station power data based on the required cooling capacity data and the system efficiency data of the chilled water station; Obtaining the power adjustment space of the chilled water station based on the chilled water station power data.

2. The method according to claim 1, wherein The required cooling capacity data includes the maximum required cooling capacity and the minimum required cooling capacity; the chilled water station power data includes the maximum chilled water station power and the minimum chilled water station power; Obtaining the chilled water station power data based on the required cooling capacity data and the system efficiency data of the chilled water station includes: Obtaining the maximum chilled water station power based on the maximum required cooling capacity and the system efficiency data; Obtaining the minimum chilled water station power based on the minimum required cooling capacity and the system efficiency data; Among them, obtaining the power adjustment space of the chilled water station based on the chilled water station power data includes: determining the difference between the maximum chilled water station power and the minimum chilled water station power as the power adjustment space.

3. The method according to claim 1 or 2, characterized in that, The method further includes: obtaining the system efficiency data of the chilled water station; obtaining the system efficiency data of the chilled water station includes: Obtaining historical required cooling capacity data and cooling water temperature data; Querying a table based on the historical required cooling capacity data to obtain initial system efficiency data, where the table includes multiple system efficiency values corresponding to multiple cooling capacity gradient values; Adjusting the initial system efficiency data based on the future weather data and the cooling water temperature data to obtain the system efficiency data.

4. The method according to claim 3, wherein Adjusting the initial system efficiency data based on the future weather data and the cooling water temperature data to obtain the system efficiency data includes: Inputting the future weather data into a cooling water temperature prediction model to obtain the cooling water temperature data; Inputting the cooling water temperature data into a system efficiency prediction model to obtain an adjustment parameter; Adjusting the initial system efficiency data based on the adjustment parameter to obtain the system efficiency data.

5. The method according to claim 1, wherein Predicting the required cooling capacity data of the chilled water station based on the future weather data, the chilled water temperature data, and the chilled water flow rate data includes: inputting the future weather data, the chilled water temperature data, and the chilled water flow rate data into a cooling capacity prediction model for prediction, and outputting the required cooling capacity data of the chilled water station; the cooling capacity prediction model is trained through the following method: Obtaining first sample data, where the first sample data includes weather sample data, chilled water supply temperature sample data, chilled water return temperature sample data, and chilled water flow rate sample data; Training the cooling capacity prediction model based on the first sample data.

6. The method according to claim 5, wherein The cooling capacity prediction model includes a reinforcement learning model; training the cooling capacity prediction model based on the first sample data includes: Constructing a state space based on the first sample data; Performing prediction based on the state space to obtain an action space, where the action space includes the predicted required cooling capacity data; Determine reward parameters for the action space based on a reward function, where the reward parameters are associated with the chilled water return temperature; Update the model parameters of the reinforcement learning model based on the reward parameters to obtain the trained reinforcement learning model.

7. The method according to claim 6, characterized in that, The updating the model parameters of the reinforcement learning model based on the reward parameters to obtain the trained reinforcement learning model includes: Cache the current state, the required cooling capacity data corresponding to the current action, the current reward parameters, and the next state; Sample from the cached data and update the model parameters of the reinforcement learning model according to the temporal difference error method.

8. The method according to claim 6, characterized in that, The chilled water return temperature associated with the reward parameters is obtained by the following method: Input the weather sample data, the chilled water supply temperature sample data, and the chilled water flow sample data of the chilled water station into an environment simulator for prediction to obtain the chilled water return temperature.

9. The method according to claim 4, wherein The cooling water temperature prediction model includes a multi-layer fully connected neural network, and the cooling water temperature prediction model is trained by the following method: Obtain second sample data and the cooling water temperature label corresponding to the second sample data, where the second sample data includes historical outdoor temperature, historical outdoor humidity, historical wind speed, and historical solar radiation; Process the cooling water temperature sample data using a multi-layer fully connected neural network to obtain the cooling water temperature prediction data; Determine the error between the cooling water temperature prediction data and the cooling water temperature label by the root mean square error method, and adjust the model parameters of the cooling water temperature prediction model based on the error to obtain the trained cooling water temperature prediction model.

10. The method according to claim 4, wherein The system efficiency prediction model includes a multi-layer fully connected neural network, and the system efficiency prediction model is trained by the following method: Obtain third sample data and the system efficiency label corresponding to the third sample data, where the third sample data includes historical cooling water temperature data; Process the third sample data using a multi-layer fully connected neural network to obtain the system efficiency prediction data; Determine the error between the system efficiency prediction data and the system efficiency label by the root mean square error method, and adjust the model parameters of the system efficiency model based on the error to obtain the trained system efficiency prediction model.

11. A prediction device for the power regulation space of a cold station, characterized in that, The device includes: An acquisition module for acquiring future weather data, chilled water temperature data, and chilled water flow data of the chilled water station; A prediction module for predicting the required cooling capacity data of the chilled water station based on the future weather data, the chilled water temperature data, and the chilled water flow data; A first acquisition module for obtaining the chilled water station power data based on the required cooling capacity data and the system efficiency data; A second acquisition module for obtaining the power adjustment space of the chilled water station based on the chilled water station power data.

12. An electronic device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the method according to any one of claims 1-10.

Images