Method and device for determining cold quantity, electronic equipment and readable storage medium

Abstract

The present disclosure relates to the technical field of cold station intelligent prediction, and provides a cold quantity determination method and device, electronic equipment and a readable storage medium. The method comprises: inputting at least one cold quantity one-hot vector into a trained cold quantity determination model, performing linear transformation processing on the at least one cold quantity one-hot vector to obtain each cold quantity feature vector; processing each cold quantity feature vector through a gated residual network to obtain each regulation vector; performing splicing processing on each cold quantity feature vector to obtain a spliced vector; processing the spliced vector through the gated residual network to obtain a weight vector; performing exponential normalization processing on the weight vector to obtain each variable selection weight value; according to a preset mapping relationship, performing splicing processing on each regulation vector and each variable selection weight value to obtain a target feature vector; and processing the target feature vector through a full connection layer to obtain target cold quantity data.

Description

Technical Field

[0001] This disclosure relates to the field of intelligent forecasting technology for cooling plants, and in particular to a method, apparatus, electronic device, and readable storage medium for determining cooling capacity. Background Technology

[0002] As the core facility for regulating indoor temperature and humidity, the operating efficiency and control precision of the chiller plant system directly affect indoor comfort and energy utilization. Traditional chiller plant control systems use experience-based setpoint control or feedback control, which is difficult to adapt to environmental changes, has a slow response, and consumes a lot of energy. In order to improve control precision and energy saving, modern chiller plants have begun to introduce computer-aided control and automation technology. Currently, the system usually collects environmental parameters and equipment operating data, trains machine learning models with historical data, and inputs real-time collected environmental parameters into the model to predict cooling demand in the future. However, because traditional machine learning algorithms cannot understand the complex nonlinear relationships between multi-source heterogeneous data, the prediction error of cooling demand is large. The model cannot automatically identify the key features that have the greatest impact on the prediction results, resulting in the omission or redundancy of feature data.

[0003] It is evident that existing technologies suffer from problems such as large errors in determining cooling capacity and data redundancy or omissions due to insufficient understanding of the complex nonlinear relationships between multi-source heterogeneous data and the inability to identify key features that affect the prediction results. Summary of the Invention

[0004] In view of this, the present disclosure provides a method, apparatus, electronic device and readable storage medium for determining cooling capacity, in order to solve the problems in the prior art that lead to large errors in determining cooling capacity and data redundancy or omissions due to insufficient understanding of the complex nonlinear relationships between multi-source heterogeneous data and the inability to identify key features that affect the prediction results.

[0005] A first aspect of this disclosure provides a method for determining cooling capacity, comprising: inputting at least one unique cooling capacity vector into a trained cooling capacity determination model; performing a linear transformation on the at least one unique cooling capacity vector to obtain cooling capacity feature vectors corresponding to each unique cooling capacity vector; processing the cooling capacity feature vectors corresponding to each unique cooling capacity vector through a gated residual network to obtain control vectors corresponding to each cooling capacity feature vector; concatenating the cooling capacity feature vectors corresponding to each unique cooling capacity vector to obtain a concatenated vector corresponding to the cooling capacity feature vector; processing the concatenated vector corresponding to the cooling capacity feature vector through a gated residual network to obtain a weight vector corresponding to the cooling capacity feature vector; performing exponential normalization on the weight vector corresponding to the cooling capacity feature vector to obtain variable selection weight values ​​corresponding to each cooling capacity feature vector; concatenating the control vectors and variable selection weight values ​​corresponding to each cooling capacity feature vector according to a preset mapping relationship to obtain a target feature vector corresponding to the cooling capacity feature vector; and processing the target feature vector corresponding to the cooling capacity feature vector through a fully connected layer to obtain target cooling capacity data.

[0006] A second aspect of this disclosure provides a cooling capacity determination apparatus, comprising: a first processing module, configured to input at least one cooling capacity unique heat vector into a trained cooling capacity determination model, and perform linear transformation processing on the at least one cooling capacity unique heat vector to obtain cooling capacity feature vectors corresponding to each cooling capacity unique heat vector; a second processing module, configured to process the cooling capacity feature vectors corresponding to each cooling capacity unique heat vector through a gated residual network to obtain control vectors corresponding to each cooling capacity feature vector; a third processing module, configured to concatenate the cooling capacity feature vectors corresponding to each cooling capacity unique heat vector to obtain concatenated vectors corresponding to the cooling capacity feature vectors; and a fourth processing module, configured to use... The first processing module processes the concatenated vector corresponding to the cooling capacity feature vector through a gated residual network to obtain the weight vector corresponding to the cooling capacity feature vector; the second processing module performs exponential normalization on the weight vector corresponding to the cooling capacity feature vector to obtain the variable selection weight value corresponding to each cooling capacity feature vector; the third processing module concatenates the control vector and the variable selection weight value corresponding to each cooling capacity feature vector according to a preset mapping relationship to obtain the target feature vector corresponding to the cooling capacity feature vector; the fourth processing module processes the target feature vector corresponding to the cooling capacity feature vector through a fully connected layer to obtain the target cooling capacity data.

[0007] A third aspect of this disclosure provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method described above.

[0008] A fourth aspect of this disclosure provides a readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method.

[0009] The beneficial effects of this embodiment compared to the prior art are as follows: By using a trained cooling capacity determination model, at least one unique cooling capacity vector is linearly transformed into a cooling capacity feature vector. A gated residual network can then be used to process the cooling capacity feature vector corresponding to the unique cooling capacity vector to obtain the control vector corresponding to the cooling capacity feature vector. Furthermore, the cooling capacity feature vectors corresponding to each unique cooling capacity vector can be concatenated to obtain a concatenated vector corresponding to the cooling capacity feature vector. By processing this concatenated vector using a gated residual network, a weight vector corresponding to the cooling capacity feature vector can be obtained. This weight vector contains the weight values ​​corresponding to each control vector. Finally, a normalized exponential function can be used to process the cooling capacity feature vector. The weight vectors corresponding to the cold load feature vectors are exponentially normalized to obtain the variable selection weight values ​​corresponding to each cold load feature vector. According to the preset mapping relationship, the control vectors corresponding to the cold load feature vectors and the variable selection weight values ​​are multiplied and concatenated to obtain the target feature vectors corresponding to the cold load feature vectors. The target feature vectors corresponding to the cold load feature vectors can be processed through fully connected layers to obtain the target cold load data. In this way, the cold load determination model enhances its ability to control the cold load feature vectors and the accuracy of weight allocation through a gated residual network. The processing of the fully connected layer improves the accuracy and reliability of the target cold load data, providing strong technical support for the accurate prediction and control of cold load demand, and improving the efficiency and accuracy of cold load determination. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this disclosure, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic flowchart of a method for determining cooling capacity provided in an embodiment of this disclosure;

[0012] Figure 2 This is a schematic diagram of the structure of a gated residual network provided in an embodiment of this disclosure;

[0013] Figure 3 This is a schematic diagram of the structure of a cooling capacity determination device provided in an embodiment of this disclosure;

[0014] Figure 4This is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. Detailed Implementation

[0015] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, so as to provide a thorough understanding of the embodiments of this disclosure. However, those skilled in the art will understand that this disclosure may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this disclosure with unnecessary detail.

[0016] A method and apparatus for determining cooling capacity according to an embodiment of the present disclosure will now be described in detail with reference to the accompanying drawings.

[0017] Figure 1 This is a flowchart illustrating a method for determining cooling capacity provided in an embodiment of this disclosure. Figure 1 The method for determining the cooling capacity can be performed by the server. For example... Figure 1 As shown, the method for determining this cooling capacity includes:

[0018] Step 101: Input at least one unique cold quantity vector into the trained cold quantity determination model, and perform a linear transformation on the at least one unique cold quantity vector to obtain the cold quantity feature vector corresponding to each unique cold quantity vector.

[0019] Specifically, at least one unique cold vector can be input into a trained cold volume determination model. The cold volume determination model can be a convolutional neural network, a long short-term memory network, or a Transformer model, etc., without limitation here. The cold volume determination model can be used to perform linear transformation processing on the input unique cold vector. In this embodiment, the cold volume determination model can be used to extract feature information from the unique cold vector to obtain the cold volume feature vector corresponding to each unique cold vector. By performing linear transformation processing on the unique cold vector through the trained cold volume determination model, the accuracy of cold volume determination is improved, the processing efficiency is improved, and the scalability of the cold volume determination model is enhanced.

[0020] The unique thermal vector of cooling capacity can be a vector representation obtained by performing unique thermal encoding on environmental monitoring data and operating data of refrigeration station equipment after cleaning. It can be used to characterize different cooling capacity types or states. In the unique thermal vector of cooling capacity, the value of one position is 1 and the values ​​of the other positions are 0. The position with a value of 1 corresponds to a specific cooling capacity type or state, and the position with a value of 0 can be used to characterize a state or category that does not belong to the position. The cooling capacity determination model can be a machine learning model trained with training sample data, which can be used to determine the target cooling capacity data. The cooling capacity feature vector corresponding to the unique thermal vector of cooling capacity can be a feature vector obtained by performing a linear transformation on the unique thermal vector of cooling capacity through the cooling capacity determination model. The cooling capacity feature vector corresponding to the unique thermal vector of cooling capacity can contain time series information, category information, semantic information, and / or contextual information, etc., which are not limited here.

[0021] For example, the cold quantity one-hot vector corresponding to the cold quantity data processed by one-hot encoding can be input into the cold quantity determination model trained based on the Transformer model. The cold quantity determination model can then perform a linear transformation on the cold quantity one-hot vector to obtain the cold quantity feature vector corresponding to the cold quantity one-hot vector.

[0022] Step 102: Process the cooling feature vectors corresponding to each cooling feature vector using a gated residual network to obtain the control vectors corresponding to each cooling feature vector.

[0023] Specifically, the residual connection structure of a gated residual network can retain information from previous layers while learning new features. Furthermore, the gating mechanism can filter and adjust the cold quantity feature vector corresponding to the one-hot vector. This allows the gated residual network to select features that significantly influence the determination of the target cold quantity data. The processed cold quantity feature vector can be mapped to a corresponding control vector, thereby improving the expressive and learning capabilities of the cold quantity determination model. The combination of residual connections and gating mechanisms in the gated residual network eliminates the gradient vanishing problem in deep networks, improving the training efficiency and stability of the cold quantity determination model. By obtaining the control vector corresponding to the cold quantity feature vector, the accuracy and efficiency of the control are improved.

[0024] Among them, the gated residual network can be a deep neural network that combines residual connections and gating mechanisms. The residual connections can skip layers and directly pass information to deeper layers. The gating mechanism can filter and adjust information through gating units, and can be used to support the network to focus on features that have a greater impact on the results. The gating mechanism can be composed of gating linear units and activation functions. The activation functions can be Sigmoid functions, Tanh functions, or ReLU functions, etc., without limitation here.

[0025] For example, the gate mechanism in the gated residual network can be used to filter the cold energy feature vectors corresponding to the cold energy unique heat vectors, and 16 cold energy feature vectors that have a great impact on the generation of target cold energy data can be selected. Then, combined with the residual network, 8 cold energy feature vectors that have a great impact on the generation of target cold energy data can be selected from the 16 cold energy feature vectors.

[0026] Step 103: Concatenate the cold feature vectors corresponding to each cold feature vector to obtain the concatenated vector corresponding to the cold feature vector.

[0027] Specifically, according to a preset order, the cold energy feature vectors corresponding to each unique cold energy vector can be concatenated in the same dimension to obtain a concatenated vector corresponding to the cold energy feature vector. The preset order includes, but is not limited to, the logical order of features or the order of feature importance. The concatenated vector corresponding to the cold energy feature vector can be used to represent the feature data that affects the determination of the target cold energy data, including but not limited to environmental monitoring data and / or cooling plant equipment operation data. In this way, the data of each cold energy feature is integrated through the concatenated vector corresponding to the cold energy feature vector, which improves the information integration effect of the cold energy determination model. By integrating feature information, the performance of the cold energy determination model is improved, the processing efficiency of feature engineering is improved, and the quality of feature information is improved.

[0028] Among them, the spliced ​​vector corresponding to the cold quantity feature vector can be a long vector formed by splicing multiple cold quantity independent heat vectors, which can contain the information of all cold quantity independent heat vectors. The length of the spliced ​​vector corresponding to the cold quantity feature vector is equal to the sum of the lengths of all cold quantity independent heat vectors.

[0029] For example, based on the importance of the features, the cold feature vectors corresponding to the 16 cold and hot vectors can be concatenated in the same dimension to obtain the concatenated vector corresponding to the cold feature vector.

[0030] Step 104: Process the concatenated vector corresponding to the cooling capacity feature vector through a gated residual network to obtain the weight vector corresponding to the cooling capacity feature vector.

[0031] Specifically, the concatenated vector corresponding to the cold quantity feature vector can be processed by the nonlinear activation function and residual structure in the gated residual network to obtain the weight vector corresponding to the cold quantity feature vector. The nonlinear activation function includes, but is not limited to, the Sigmoid function, the Tanh function, or the ReLU function. The obtained weight vector contains the weight vector corresponding to each cold quantity feature vector. The weight vector can be used to represent the weight value of each cold quantity feature vector. The weight value can be a value between 0 and 1, which is used to represent the importance of the corresponding cold quantity feature vector. This improves the accuracy of feature information selection, enhances the performance of the cold quantity determination model, and strengthens the robustness of the cold quantity determination model.

[0032] Among them, the weight vector corresponding to the cold quantity feature vector can be a vector with the same length as the concatenated vector of the cold quantity feature vector. The elements in the weight vector corresponding to the cold quantity feature vector can be used to characterize the importance of the corresponding cold quantity feature. The weight vector corresponding to the cold quantity feature vector can be obtained by processing the concatenated vector through a gated residual network.

[0033] For example, a gated residual network can be used to process the concatenated vector corresponding to the 16-dimensional cold energy feature vector to obtain the weight vector corresponding to the cold energy feature vector. The weight vector corresponding to the cold energy feature vector contains the weight values ​​corresponding to the 16 cold energy feature vectors.

[0034] Step 105: Perform exponential normalization on the weight vectors corresponding to the cooling capacity feature vectors to obtain the variable selection weight values ​​corresponding to each cooling capacity feature vector.

[0035] Specifically, the weight vectors corresponding to the coldness feature vectors can be processed using normalization functions. Exponential normalization functions include, but are not limited to, the softmax function or the sigmoid function. For example, the weights can be adjusted by multiplying by a constant or adding an offset to obtain the variable selection weight values ​​corresponding to each coldness feature vector. This improves the rationality of the weight distribution, enhances the stability of the coldness determination model, optimizes the processing flow, and improves processing efficiency.

[0036] Among them, the variable selection weight value corresponding to the cooling capacity feature vector can be the weight value corresponding to each cooling capacity feature vector after exponential normalization. The weight value corresponding to the cooling capacity feature vector can be used to characterize the importance or influence of different cooling capacity features in the cooling capacity determination model.

[0037] Step 106: According to the preset mapping relationship, the control vectors corresponding to each cooling feature vector and the variables corresponding to each cooling feature vector are selected and weighted and concatenated to obtain the target feature vector corresponding to the cooling feature vector.

[0038] Specifically, the preset mapping relationship can assign a variable selection weight value to each cold quantity feature vector. Then, the control vector corresponding to the cold quantity feature vector and the variable selection weight value corresponding to the cold quantity feature vector can be multiplied one-to-one. The multiple multiplication results are then concatenated. The concatenation method can be serialization, or combination or transformation according to the preset mapping relationship, which is not limited here. The target feature vector corresponding to the cold quantity feature vector is obtained, thereby improving the richness of the cold quantity feature vector, simplifying the processing flow, and improving the performance of the cold quantity determination model.

[0039] The preset mapping relationship can be a rule or function set in advance before the splicing process. It can be used to characterize the method of matching the control vector corresponding to the cold energy feature vector with the variable selection weight value corresponding to the cold energy feature vector and then splicing them. The target feature vector corresponding to the cold energy feature vector can be the feature vector obtained by splicing the control vector corresponding to the cold energy feature vector with the variable selection weight value corresponding to the cold energy feature vector. The target feature vector corresponding to the cold energy feature vector can be used to determine the target cold energy data.

[0040] For example, the preset mapping relationship can be a one-to-one correspondence between the control vector corresponding to the cooling feature vector and the variable selection weight value corresponding to the cooling feature vector. Then, the control vector corresponding to the cooling feature vector and the variable selection weight value corresponding to the cooling feature vector are multiplied and concatenated respectively to obtain the target feature vector corresponding to the cooling feature vector.

[0041] Step 107: Process the target feature vector corresponding to the cooling feature vector through a fully connected layer to obtain the target cooling data.

[0042] Specifically, a fully connected layer can be used to multiply the target feature vector corresponding to the coldness feature vector (for example, the target feature vector can have a dimension of n) with a weight matrix (for example, the weight matrix can have a dimension of m*n, where m can be the dimension of the output feature vector), and add a bias vector (for example, the size of the bias vector can be m). This results in a linear transformation of the target feature vector. The linear transformation of the target feature vector can be achieved using a non-linear activation function, including but not limited to the ReLU function, Sigmoid function, or Tanh function, to introduce non-linear properties and obtain the target coldness data. Through the processing of the fully connected layer, the target feature vector corresponding to the coldness feature vector can be mapped to a low-dimensional representation space, improving the performance of the coldness determination model. The fully connected layer can learn weights and biases to extract meaningful features from the coldness feature vector, enhancing the quality of feature information.

[0043] Among them, the fully connected layer can be a layer structure in a neural network. Each neuron in the fully connected layer is connected to all neurons in the previous layer. The fully connected layer can be used to receive feature maps from the convolutional layer or pooling layer, and to reduce the dimensionality, extract and combine the feature information. The target cooling data can be the final representation of the target feature vector after being processed by the fully connected layer. The target cooling data can be used to characterize the cooling demand for a specified period.

[0044] According to the technical solution provided in this disclosure, by using a trained cooling capacity determination model, at least one unique cooling capacity vector is linearly transformed into a cooling capacity feature vector. A gated residual network can then be used to process the cooling capacity feature vector corresponding to the unique cooling capacity vector to obtain a control vector corresponding to the cooling capacity feature vector. Furthermore, the cooling capacity feature vectors corresponding to each unique cooling capacity vector can be concatenated to obtain a concatenated vector corresponding to the cooling capacity feature vector. By processing this concatenated vector using a gated residual network, a weight vector corresponding to the cooling capacity feature vector can be obtained. This weight vector contains the weight values ​​corresponding to each control vector. Finally, a normalized exponential function can be used to adjust the cooling capacity feature vector... The corresponding weight vectors are exponentially normalized to obtain the variable selection weight values ​​corresponding to each cooling capacity feature vector. According to the preset mapping relationship, the corresponding control vector and variable selection weight values ​​are multiplied and concatenated to obtain the target feature vector corresponding to the cooling capacity feature vector. The target feature vector corresponding to the cooling capacity feature vector can be processed through a fully connected layer to obtain the target cooling capacity data. In this way, the control capability and weight allocation accuracy of the cooling capacity determination model are enhanced by the gated residual network. The processing of the fully connected layer improves the accuracy and reliability of the target cooling capacity data, providing strong technical support for the accurate prediction and control of cooling capacity demand, and improving the efficiency and accuracy of cooling capacity determination.

[0045] In some embodiments, before inputting at least one unique thermal vector of cooling capacity into the trained cooling capacity determination model and performing linear transformation on the at least one unique thermal vector of cooling capacity to obtain the cooling capacity feature vector corresponding to each unique thermal vector of cooling capacity, the method further includes: acquiring at least one environmental monitoring data and at least one cooling plant equipment operation data; performing data cleaning processing on each environmental monitoring data and each cooling plant equipment operation data to obtain at least one cleaned environmental monitoring data and at least one cleaned cooling plant equipment operation data; and performing unique thermal encoding processing on each cleaned environmental monitoring data and each cleaned cooling plant equipment operation data to obtain at least one unique thermal vector of cooling capacity.

[0046] Specifically, at least one set of environmental monitoring data and at least one set of chiller plant equipment operation data can be obtained. The environmental monitoring data may include indoor and outdoor temperature, humidity, pedestrian traffic, wind speed and force, and / or weather, etc., without limitation. The chiller plant equipment operation data may include chiller power, water pump frequency, chilled water supply and return water temperature, chilled water supply and return water pressure, chilled water supply and return water flow rate, and / or cooling tower operation status, etc., without limitation. Data cleaning processing can be performed on the environmental monitoring data and chiller plant equipment operation data. Data cleaning processing can be used to remove noise, fill missing values, and / or correct erroneous data, etc., without limitation. Thus, at least one set of cleaned environmental monitoring data and at least one set of cleaned chiller plant equipment operation data can be obtained. The cleaned data can be processed by one-heat encoding, which can convert categorical variables into a format that can be processed by machine learning models to obtain a chilled capacity one-heat vector.

[0047] Environmental monitoring data can refer to data used to monitor the environmental conditions inside and outside the building, including but not limited to indoor and outdoor temperature, humidity, pedestrian traffic, wind force and speed, and / or weather. Environmental monitoring data can be used to characterize the indoor and outdoor thermal and humidity environment and comfort. Chilling plant equipment operation data can be the operating status and performance data of various cooling equipment in the chilling plant, including but not limited to chiller power, water pump frequency, chilled water supply and return water temperature, chilled water supply and return water pressure, chilled water supply and return water flow rate, and / or cooling tower operating status. Chilling plant equipment operation data can be used to characterize the operating efficiency and performance of the chilling plant equipment. Cleaned environmental monitoring data can be environmental monitoring data after data cleaning processing. Data cleaning can be used to remove noise, missing values, and erroneous data from the original data, and is not limited here. Cleaned chilling plant equipment operation data can be chilling plant equipment operation data after data cleaning processing. Cleaned chilling plant equipment operation data can be used to characterize the actual operating status and performance of the chilling plant equipment.

[0048] For example, indoor temperature data and cooling equipment operating status data can be acquired. Data cleaning processing can be performed on the indoor temperature data and cooling equipment operating status data to obtain cleaned indoor temperature data and cleaned cooling equipment operating status data. Unique thermal encoding processing can be performed on the cleaned indoor temperature data and cleaned cooling equipment operating status data to obtain a cooling capacity unique thermal vector.

[0049] According to the technical solution provided in this disclosure, by acquiring at least one environmental monitoring data and at least one chiller plant equipment operation data, the environmental monitoring data and chiller plant equipment operation data can be cleaned to obtain at least one cleaned environmental monitoring data and at least one cleaned chiller plant equipment operation data. The cleaned data can be processed by unique thermal encoding to obtain a chilling capacity unique thermal vector. By acquiring, cleaning, and uniquely thermal encoding the environmental monitoring data and chiller plant equipment operation data, the accuracy and reliability of the data are improved, the accuracy of the chilling capacity determination model for data determination is enhanced, the operating efficiency of the chiller plant equipment is improved, energy waste is reduced, and operating costs are lowered.

[0050] In some embodiments, the gated residual network includes a first fully connected layer, a second fully connected layer, and a gated linear unit. The gated residual network processes the cooling feature vectors corresponding to each cooling feature vector to obtain a control vector corresponding to each cooling feature vector. This process includes: processing the cooling feature vectors corresponding to the cooling feature vectors corresponding to the cooling feature vectors through the first fully connected layer to obtain a first fully connected cooling feature vector; performing nonlinear activation processing on the first fully connected cooling feature vector to obtain an activated feature vector; processing the activated feature vector to obtain a second fully connected cooling feature vector; performing regularization processing on the second fully connected cooling feature vector to obtain a regularized feature vector; processing the regularized feature vector to obtain a gated cooling feature vector; and performing residual connection processing on the gated cooling feature vector and the cooling feature vector corresponding to the cooling feature vector to obtain a control vector.

[0051] Specifically, the cold energy unique heat vector can be processed through the first fully connected layer in the gated residual network. The first fully connected layer can be a densely connected neural network layer, which can be used to learn the linear relationship between the input and output, thereby obtaining the cold energy first fully connected vector corresponding to the cold energy feature vector. The obtained cold energy first fully connected vector can be subjected to nonlinear activation processing. The nonlinear activation function for nonlinear activation processing includes, but is not limited to, the ReLU function, the Sigmoid function, or the Tanh function, etc., which are not limited here, thereby obtaining the activation feature vector corresponding to the cold energy feature vector. The activation feature vector corresponding to the cold energy feature vector can then be processed by the second fully connected layer to obtain the cold energy second fully connected vector corresponding to the cold energy feature vector. The cold energy second fully connected vector can be regularized to obtain the regularized feature vector corresponding to the cold energy feature vector. The regularized feature vector corresponding to the cold energy feature vector can be processed by the gated linear unit to obtain the gated cold energy feature vector corresponding to the cold energy feature vector. The gated cold energy feature vector and the cold energy feature vector corresponding to the cold energy unique heat vector can be residually connected to obtain the control vector corresponding to the cold energy feature vector.

[0052] In this system, the first fully connected layer can be a layer in a neural network, where each input node is connected to an output node, and it can be used to learn the linear relationship between input and output. The second fully connected layer can be a fully connected layer in a gated residual network that is located at a different position from the first fully connected layer. The gated linear unit can be a neural network unit that combines gating mechanism and linear transformation, and can be used to dynamically control the flow of information. The first fully connected vector corresponding to the cold quantity feature vector can be the output vector obtained after processing the cold quantity feature vector through the first fully connected layer. The activation feature vector corresponding to the cold quantity feature vector can be the vector obtained after performing nonlinear activation processing on the first fully connected vector. The second fully connected vector corresponding to the cold quantity feature vector can be the output vector obtained after processing the activation feature vector through the second fully connected layer. The regularized feature vector corresponding to the cold quantity feature vector can be the vector obtained after performing regularization processing on the second fully connected vector. The gated cold quantity feature vector corresponding to the cold quantity feature vector can be the vector obtained after processing the regularized feature vector through the gated linear unit.

[0053] For example, the cooling capacity unique heat vector can be processed by the first fully connected layer in a gated residual network to obtain the first fully connected vector of cooling capacity corresponding to the cooling capacity feature vector. The obtained first fully connected vector of cooling capacity can be nonlinearly activated by the Sigmoid function to obtain the activated feature vector corresponding to the cooling capacity feature vector. The activated feature vector corresponding to the cooling capacity feature vector can then be processed by the second fully connected layer to obtain the second fully connected vector of cooling capacity corresponding to the cooling capacity feature vector. The second fully connected vector of cooling capacity can be weighted and regularized to obtain the regularized feature vector corresponding to the cooling capacity feature vector. The regularized feature vector corresponding to the cooling capacity feature vector can be processed by a gated linear unit to obtain the gated cooling capacity feature vector corresponding to the cooling capacity feature vector. The gated cooling capacity feature vector and the cooling capacity feature vector corresponding to the cooling capacity unique heat vector can be residually connected to obtain the control vector corresponding to the cooling capacity feature vector.

[0054] According to the technical solution provided in this embodiment, the cold quantity unique heat vector is processed by the first fully connected layer in the gated residual network to obtain the cold quantity first fully connected vector corresponding to the cold quantity feature vector. The obtained cold quantity first fully connected vector can be nonlinearly activated to obtain the activated feature vector corresponding to the cold quantity feature vector. The activated feature vector corresponding to the cold quantity feature vector can then be processed by the second fully connected layer to obtain the cold quantity second fully connected vector corresponding to the cold quantity feature vector. The cold quantity second fully connected vector can be regularized to obtain the regularized feature vector corresponding to the cold quantity feature vector. The regularized feature vector corresponding to the cold quantity feature vector can be processed by a gated linear unit to obtain the gated cold quantity feature vector corresponding to the cold quantity feature vector. A residual connection can be performed between the gated cold quantity feature vector and the cold quantity feature vector corresponding to the cold quantity unique heat vector to obtain the control vector corresponding to the cold quantity feature vector. Thus, residual connection avoids information loss during transmission, gated linear unit improves the dynamic capability of information flow, and regularization prevents overfitting, thereby improving the generalization ability of the cold quantity determination model.

[0055] In some embodiments, before inputting at least one uniquely heated cooling vector into a trained cooling capacity determination model and performing a linear transformation on the at least one uniquely heated cooling vector to obtain cooling capacity feature vectors corresponding to each uniquely heated cooling vector, the method further includes: acquiring multiple training sample data sets and labels corresponding to each training sample data set, wherein each training sample data set contains multiple environmental monitoring training data and multiple chiller plant equipment operation training data, and the labels corresponding to the training sample data sets are used to characterize the actual cooling capacity of the training sample data sets; performing data cleaning on each training sample data set to obtain at least one cleaned environmental monitoring training data and at least one cleaned chiller plant equipment operation training data; performing uniquely heated encoding on each cleaned environmental monitoring training data and each cleaned chiller plant equipment operation training data to obtain at least one uniquely heated cooling training vector; performing a linear transformation on the at least one uniquely heated cooling training vector to obtain cooling capacity feature training vectors corresponding to each uniquely heated cooling training vector; and using a gated residual network to process the cooling capacity features corresponding to each uniquely heated cooling training vector. The training vectors are processed to obtain the control vectors corresponding to each cold quantity feature training vector; the cold quantity feature training vectors corresponding to each cold quantity unique heat training vector are concatenated to obtain the concatenated vector; the concatenated vector corresponding to the cold quantity feature training vector is processed through a gated residual network to obtain the weight vector corresponding to the cold quantity feature training vector; the weight vector corresponding to the cold quantity feature training vector is exponentially normalized to obtain the variable selection weight value corresponding to each cold quantity feature training vector; according to a preset mapping relationship, the control vector and the variable selection weight value corresponding to each cold quantity feature training vector are concatenated to obtain the target feature vector corresponding to the cold quantity feature vector; the target feature vector corresponding to the cold quantity feature training vector is processed through a fully connected layer to obtain the training cold quantity data; the loss of the cold quantity determination model is determined based on the root mean square error loss function, the training cold quantity data, and the labels corresponding to the training sample data set; the parameters in the cold quantity determination model are updated according to the loss of the cold quantity determination model through iterative loops.

[0056] Specifically, multiple training sample datasets can be acquired, each containing multiple environmental monitoring data sets and multiple chiller plant equipment operation data sets, each labeled with a representation of the actual cooling capacity. Data cleaning can be performed on these training sample datasets to remove abnormal or erroneous data. The cleaned environmental monitoring training data and the cleaned chiller plant equipment operation training data can be converted into chiller capacity one-hot training vectors through one-hot encoding. Further linear transformation can be performed to obtain the chiller capacity feature training vectors corresponding to these one-hot training vectors. These chiller capacity feature training vectors can then be processed using a gated residual network to obtain the corresponding control vectors. Simultaneously, the chiller capacity feature training vectors can be further processed... The training vectors are concatenated to obtain the concatenated vector corresponding to the cold volume feature training vector. Then, a gated residual network is used to process the concatenated vector corresponding to the cold volume feature training vector to obtain the weight vector corresponding to the cold volume feature training vector. The weight vector corresponding to the cold volume feature training vector can then be exponentially normalized to obtain the variable selection weight value, which can be concatenated with the control vector to form the target feature vector. The target feature vector can be processed through a fully connected layer to obtain the training cold volume data. The difference between the training cold volume data and the real cold volume label can be compared according to the root mean square error loss function to determine the loss of the cold volume determination model, and the model parameters are optimized through iterative loops.

[0057] The training sample dataset can be a dataset containing multiple environmental monitoring training data and cooling station equipment operation training data, which can be used in the training process of the cooling quantity determination model. The label corresponding to the training sample dataset can be the actual cooling quantity value of each training sample dataset, which can be used for model performance evaluation in the supervised learning process. The actual cooling quantity corresponding to the training sample dataset can be the actual measured or recorded cooling quantity value, which can be used as the target value obtained by the cooling quantity determination model. The root mean square error loss function can be a loss function used for regression problems, which can be used to measure the average difference between the predicted value and the actual value of the cooling quantity determination model. The loss of the cooling quantity determination model can be a measure of the difference between the predicted cooling quantity value and the actual cooling quantity value of the cooling quantity determination model during the training process, which can be used to guide the parameter update and optimization of the cooling quantity determination model.

[0058] Furthermore, it should be noted that parameters can be updated using adaptive optimization algorithms. For example, the Adam algorithm can set the initial learning rate to 0.001 and dynamically decay it. Regularization can be achieved through strategies such as random deactivation, L1 norm regularization, and weight decay regularization to prevent overfitting, which are not limited here.

[0059] According to the technical solution provided in this disclosure, by acquiring multiple training sample datasets, each dataset containing multiple environmental monitoring data and multiple chiller equipment operation data, and attaching labels representing the actual cooling capacity, the training sample datasets can be cleaned to remove abnormal or erroneous data. The cleaned environmental monitoring training data and the cleaned chiller equipment operation training data can be converted into cooling capacity one-hot training vectors through one-hot encoding, and can be further linearly transformed to obtain cooling capacity feature training vectors corresponding to the cooling capacity one-hot training vectors. The cooling capacity feature training vectors can be processed through a gated residual network to obtain the control vectors corresponding to the cooling capacity feature training vectors. Simultaneously, the cooling capacity feature training vectors can be concatenated to obtain the concatenated vectors corresponding to the cooling capacity feature training vectors. Then, the control vectors corresponding to the cooling capacity feature training vectors can be processed through a gated residual network. The concatenated vectors are processed to obtain the weight vectors corresponding to the cold quantity feature training vectors. Then, the weight vectors corresponding to the cold quantity feature training vectors can be exponentially normalized to obtain variable selection weight values, which can be concatenated with the control vector to form the target feature vector. The target feature vector can be processed through a fully connected layer to obtain training cold quantity data. The difference between the training cold quantity data and the actual cold quantity labels can be compared using the root mean square error loss function to determine the loss of the cold quantity determination model. The model parameters are optimized through iterative iteration. By combining environmental monitoring training data and cold station equipment operation training data with a gated residual network, the accuracy of cold quantity determination is improved. Data cleaning and one-hot encoding improve data quality. The global information capture capability and feature selection mechanism of the gated residual network enhance the generalization ability and prediction accuracy of the cold quantity determination model.

[0060] In some embodiments, the method further includes: acquiring at least one real-time environmental monitoring data and at least one real-time chiller plant equipment operation data; performing data cleaning processing on the at least one real-time environmental monitoring data and at least one real-time chiller plant equipment operation data to obtain at least one cleaned real-time environmental monitoring data and at least one cleaned real-time chiller plant equipment operation data; performing unique thermal encoding processing on each cleaned real-time environmental monitoring data and each cleaned real-time chiller plant equipment operation data to obtain at least one real-time cooling capacity unique thermal vector; and processing the at least one real-time cooling capacity unique thermal vector according to a preset sliding window and a preset cooling capacity determination time period through a cooling capacity determination model to obtain real-time cooling capacity data.

[0061] Specifically, at least one real-time environmental monitoring data and at least one real-time cooling station equipment operation data can be acquired. Then, the real-time environmental monitoring data and the real-time cooling station equipment operation data can be cleaned to remove noise, outliers, and non-standard format data. The cleaned data can be processed by one-hot encoding to obtain at least one real-time cooling capacity one-hot vector. Based on a preset sliding window and cooling capacity determination time period, the real-time cooling capacity one-hot vector can be processed by a trained cooling capacity determination model to obtain real-time cooling capacity data.

[0062] Among them, real-time environmental monitoring data can be data on environmental conditions obtained in real time through an environmental monitoring system, including but not limited to real-time indoor and outdoor temperature, humidity, traffic flow, wind force and speed, weather, etc. Real-time chiller equipment operation data can be data generated in real time during the operation of chiller equipment, including but not limited to real-time acquired chiller power, water pump frequency, chilled water supply and return water temperature, chilled water supply and return water pressure, chilled water supply and return water flow, cooling tower operation status, etc. Real-time cooling capacity unique heat vector can be a vector obtained by performing unique heat encoding processing on the cleaned real-time environmental monitoring data and the cleaned real-time chiller equipment operation data, which can be used to characterize the real-time cooling capacity status. The preset sliding window can be a time range, which can be used to select data within a time period from the real-time data for analysis. The size of the sliding window can be set according to actual needs, including but not limited to 1 hour, 2 hours, or 5 hours, etc. The preset cooling capacity determination time period can be a specific time period used to determine the cooling capacity, including but not limited to the next 4 hours, 8 hours, or 24 hours, etc. Real-time cooling capacity data can be data that can be used to characterize the cooling capacity at the current moment by processing the real-time cooling capacity unique heat vector through the cooling capacity determination model.

[0063] For example, real-time environmental monitoring data and real-time chiller equipment operation data can be cleaned and processed, and the cleaned data can be processed by unique thermal encoding to obtain a real-time chilling capacity unique thermal vector. The preset sliding window can be in 1-hour increments, and the preset chilling capacity determination period can be the chilling capacity demand for the next 24 hours. Through this chilling capacity determination model, real-time chilling capacity data can be obtained.

[0064] According to the technical solution provided in this disclosure, by acquiring at least one real-time environmental monitoring data and at least one real-time cooling plant equipment operation data, data cleaning processing can be performed on the real-time environmental monitoring data and the real-time cooling plant equipment operation data. The cleaned data can be processed by unique heat encoding to obtain at least one real-time cooling capacity unique heat vector. Based on a preset sliding window and cooling capacity determination time period, the real-time cooling capacity unique heat vector can be processed by a trained cooling capacity determination model to obtain real-time cooling capacity data. In this way, by acquiring environmental monitoring data and cooling plant equipment operation data in real time, combined with data cleaning and unique heat encoding processing, and determining cooling capacity through a preset sliding window and a preset cooling capacity determination time period, dynamic, accurate, and real-time monitoring and prediction of cooling capacity is ensured, reducing energy consumption and improving operating efficiency.

[0065] Based on the foregoing embodiments, the method further includes: storing real-time cooling data in a preset database; filtering the real-time cooling data stored in the preset database based on a preset data accumulation time to obtain real-time cooling data corresponding to the preset data accumulation time; and updating the parameters in the cooling capacity determination model based on the real-time cooling data corresponding to the preset data accumulation time.

[0066] Specifically, real-time cooling data can be stored in a preset database. This database can be used to store and manage cooling-related data. The real-time cooling data stored in the database can then be filtered based on a preset data accumulation time. This filtering process may include removing outliers, calculating statistics such as average, maximum, and minimum values, and / or selecting real-time cooling data that fits within the preset data accumulation time. This yields real-time cooling data within the preset data accumulation time. The cooling data determination model can then be trained based on the real-time cooling data corresponding to the preset data accumulation time, thereby updating the parameters in the cooling data determination model.

[0067] The preset database can be a pre-defined data storage structure used to store and manage real-time cooling data. The preset data accumulation time can be a pre-defined time range used to filter real-time cooling data for a specific time period in the database. The preset data accumulation time can be set according to actual needs, including but not limited to 7 days, 1 month, 1 year, etc. The real-time cooling data corresponding to the preset data accumulation time can be the real-time cooling data filtered from the preset database within the preset data accumulation time. The real-time cooling data corresponding to the preset data accumulation time can be used to train the cooling determination model, thereby updating the parameters of the cooling determination model.

[0068] For example, the preset data accumulation time can be 7 days. Real-time cooling data obtained in the last 7 days can be filtered from the preset database. The real-time cooling data obtained in the last 7 days can be used as the training sample data set to train the cooling determination model, and the parameters in the cooling determination model can be updated according to the obtained loss.

[0069] In addition, it should be noted that the performance of the cold volume determination model trained on real-time cold volume data corresponding to a preset data accumulation time can be verified through the validation set. If the model performs better than the current online model on the validation set, the new model can be updated and deployed through hot update.

[0070] According to the technical solution provided in this disclosure, by storing real-time cooling data in a preset database, the real-time cooling data stored in the preset database can be filtered based on a preset data accumulation time to obtain real-time cooling data within the preset data accumulation time. The cooling data determination model can be trained based on the real-time cooling data corresponding to the preset data accumulation time, thereby updating the parameters in the cooling data determination model. In this way, by storing real-time cooling data in a preset database, centralized management of cooling data and preservation of historical data are achieved. Filtering real-time cooling data based on a preset data accumulation time improves the effectiveness of updating the parameters of the cooling data determination model. Regularly updating the parameters of the cooling data determination model improves the accuracy and adaptability of the cooling data determination model.

[0071] In some embodiments, environmental monitoring data includes indoor temperature and humidity data; chiller station equipment operation data includes cooling capacity data; cleaned environmental monitoring data includes indoor temperature and humidity analysis data and indoor wet-bulb temperature data; cleaned chiller station equipment operation data includes non-negative cooling capacity data; data cleaning processing is performed on each environmental monitoring data and each chiller station equipment operation data to obtain at least one cooling capacity-related cleaning data, including: truncating the environmental monitoring cooling capacity data to obtain non-negative cooling capacity data; performing exploratory data analysis processing on the indoor temperature and humidity data to obtain indoor temperature and humidity analysis data; and determining the indoor wet-bulb temperature data based on the indoor temperature and humidity data.

[0072] Specifically, data cleaning can be performed on environmental monitoring data and chiller plant equipment operation data. For indoor temperature and humidity data in environmental monitoring data, exploratory data analysis can be performed, including but not limited to data validity checks, outlier handling, and / or missing value filling. Indoor temperature and humidity analysis data can be obtained. Based on the temperature and humidity information in the indoor temperature and humidity data, the indoor wet-bulb temperature data can be determined using the wet-bulb temperature calculation formula or lookup table method. For cooling capacity data in chiller plant equipment operation data, negative value truncation can be performed. By setting negative cooling capacity data to zero or a suitable non-negative value, non-negative cooling capacity data can be obtained.

[0073] Among them, indoor temperature and humidity data can be the temperature and humidity information in the indoor environment, which can be obtained through devices such as temperature and humidity sensors; cooling capacity data can be the cooling capacity data generated by the refrigeration station equipment during operation, which can be used to characterize the cooling capacity and efficiency of the equipment; indoor temperature and humidity analysis data can be the indoor temperature and humidity data obtained after exploratory data analysis and processing; indoor wet-bulb temperature data can be the wet-bulb temperature information calculated based on indoor temperature and humidity data, which can be used to assess the comfort of the indoor environment; and non-negative cooling capacity data can be the cooling capacity data after being truncated to negative values, which can be used to ensure that all cooling capacity values ​​are non-negative and conform to physical laws.

[0074] For example, negative values ​​can be truncated to zero in the acquired cooling data; exploratory data analysis and statistical feature extraction can be performed on indoor temperature and humidity data; data can be filtered by filtering abnormal sensors; and wet-bulb temperature can be calculated as additional feature data based on the acquired indoor temperature and humidity data.

[0075] Furthermore, it should be noted that additional date and time feature data can be constructed based on the obtained time data, including but not limited to holiday time, work-rest schedule, number of days until the holiday, and / or business hours; the cooling data can also be accumulated to obtain the hourly cumulative cooling, which can be used to ensure the stability of the feature data; the skewed feature data can also be processed, and the processing methods include but are not limited to transformation, normalization, and / or truncation.

[0076] According to the technical solution provided in this disclosure, by performing data cleaning processing on environmental monitoring data and chiller plant equipment operation data, exploratory data analysis processing can be performed on indoor temperature and humidity data in the environmental monitoring data to obtain indoor temperature and humidity analysis data. Indoor wet-bulb temperature data can be determined based on the temperature and humidity information in the indoor temperature and humidity data. Negative value truncation processing can be performed on the cooling capacity data in the chiller plant equipment operation data. In this way, through data cleaning processing, the accuracy of cooling capacity cleaning data is improved, the operating efficiency of the chiller plant is improved, and energy consumption is reduced.

[0077] Figure 2 This is a schematic diagram of the structure of a gated residual network provided in an embodiment of this disclosure, as shown below. Figure 2 As shown, the structure of this gated residual network includes:

[0078] The first fully connected layer 201 can be a layer in a neural network, where each input node is connected to the output node. It is used to learn the linear relationship between the input and the output, and thus obtain the first fully connected vector of coldness corresponding to the coldness feature vector.

[0079] The second fully connected layer 202 can be a fully connected layer in the gated residual network that is located at a different position from the first fully connected layer.

[0080] The gated linear unit 203 can be a neural network unit that combines gating mechanism and linear transformation, and can be used to dynamically control the flow of information.

[0081] According to the technical solution provided in this embodiment, by inputting the unique heat vector of cold energy into the first fully connected layer 201 of the gated residual network for processing, a first fully connected vector corresponding to the cold energy feature vector is obtained. Then, nonlinear activation processing can be performed on the first fully connected vector corresponding to the cold energy feature vector. This can be achieved by multiplying the first fully connected vector corresponding to the input cold energy feature vector with the result of Tanh(log(1+Sigmoid(x))), where x can be the input cold energy feature vector. This yields the activation feature vector corresponding to the cold energy feature vector. The activation feature vector corresponding to the cold energy feature vector can then be processed by the second fully connected layer 202 to obtain the second fully connected vector corresponding to the cold energy feature vector. Weight decay regularization processing can then be performed on the second fully connected vector to obtain... The regularized feature vector corresponding to the cold quantity feature vector can be processed by a gated linear unit 203 to obtain a gated cold quantity feature vector. The gated linear unit 203 can be composed of a linear transformation and a Sigmoid function product. It can perform residual connection on the gated cold quantity feature vector and the cold quantity feature vector corresponding to the cold quantity unique heat vector to obtain the control vector corresponding to the cold quantity feature vector. The residual connection can be processed by a gated residual network, which can be obtained by ReLU function and linear transformation. In this way, the loss of information in the process of information transmission is avoided by residual connection, the dynamic capability of information flow is improved by gated linear unit, and overfitting is prevented by regularization, thereby improving the generalization ability of the cold quantity determination model.

[0082] All of the above-mentioned optional technical solutions can be combined in any way to form optional embodiments of this disclosure, and will not be described in detail here.

[0083] The following are embodiments of the apparatus disclosed herein, which can be used to execute embodiments of the method disclosed herein. For details not disclosed in the apparatus embodiments of this disclosure, please refer to the embodiments of the method disclosed herein.

[0084] Figure 3 This is a schematic diagram of a cooling capacity determination device provided in an embodiment of this disclosure. Figure 3 As shown, the device for determining the cooling capacity includes:

[0085] The first processing module 301 is used to input at least one cold quantity unique heat vector into the trained cold quantity determination model, and perform linear transformation processing on the at least one cold quantity unique heat vector to obtain the cold quantity feature vector corresponding to each cold quantity unique heat vector.

[0086] The second processing module 302 is used to process the cooling feature vectors corresponding to each cooling feature vector through a gated residual network to obtain the control vectors corresponding to each cooling feature vector.

[0087] The third processing module 303 is used to concatenate the cold energy feature vectors corresponding to each cold energy unique heat vector to obtain the concatenated vectors corresponding to the cold energy feature vectors.

[0088] The fourth processing module 304 is used to process the concatenated vector corresponding to the cooling capacity feature vector through a gated residual network to obtain the weight vector corresponding to the cooling capacity feature vector.

[0089] The fifth processing module 305 is used to perform exponential normalization on the weight vector corresponding to the cooling capacity feature vector to obtain the variable selection weight value corresponding to each cooling capacity feature vector.

[0090] The determining module 306 is used to select weight values ​​for the control vectors corresponding to each cooling capacity feature vector and the variables corresponding to each cooling capacity feature vector according to the preset mapping relationship, and then concatenate them to obtain the target feature vector corresponding to the cooling capacity feature vector.

[0091] The sixth processing module 307 is used to process the target feature vector corresponding to the cooling feature vector through a fully connected layer to obtain the target cooling data.

[0092] According to the technical solution provided in this disclosure, by using a trained cooling capacity determination model, at least one unique cooling capacity vector is linearly transformed into a cooling capacity feature vector. A gated residual network can then be used to process the cooling capacity feature vector corresponding to the unique cooling capacity vector to obtain a control vector corresponding to the cooling capacity feature vector. Furthermore, the cooling capacity feature vectors corresponding to each unique cooling capacity vector can be concatenated to obtain a concatenated vector corresponding to the cooling capacity feature vector. By processing this concatenated vector using a gated residual network, a weight vector corresponding to the cooling capacity feature vector can be obtained. This weight vector contains the weight values ​​corresponding to each control vector. Finally, a normalized exponential function can be used to adjust the cooling capacity feature vector... The corresponding weight vectors are exponentially normalized to obtain the variable selection weight values ​​corresponding to each cooling capacity feature vector. According to the preset mapping relationship, the corresponding control vector and variable selection weight values ​​are multiplied and concatenated to obtain the target feature vector corresponding to the cooling capacity feature vector. The target feature vector corresponding to the cooling capacity feature vector can be processed through a fully connected layer to obtain the target cooling capacity data. In this way, the control capability and weight allocation accuracy of the cooling capacity determination model are enhanced by the gated residual network. The processing of the fully connected layer improves the accuracy and reliability of the target cooling capacity data, providing strong technical support for the accurate prediction and control of cooling capacity demand, and improving the efficiency and accuracy of cooling capacity determination.

[0093] In some embodiments, the above-described cooling capacity determination device is further configured to: acquire at least one environmental monitoring data and at least one cooling station equipment operation data; perform data cleaning processing on each environmental monitoring data and each cooling station equipment operation data to obtain at least one cleaned environmental monitoring data and at least one cleaned cooling station equipment operation data; and perform unique thermal encoding processing on each cleaned environmental monitoring data and each cleaned cooling station equipment operation data to obtain at least one cooling capacity unique thermal vector.

[0094] In some embodiments, the second processing module 302 is specifically configured to: process the cold energy feature vector corresponding to the cold energy unique heat vector through a first fully connected layer to obtain a first fully connected vector of cold energy corresponding to the cold energy feature vector; perform nonlinear activation processing on the first fully connected vector of cold energy corresponding to the cold energy feature vector to obtain an activated feature vector of cold energy corresponding to the cold energy feature vector; process the activated feature vector of cold energy corresponding to the cold energy feature vector through a second fully connected layer to obtain a second fully connected vector of cold energy corresponding to the cold energy feature vector; perform regularization processing on the second fully connected vector of cold energy corresponding to the cold energy feature vector to obtain a regularized feature vector of cold energy corresponding to the cold energy feature vector; process the regularized feature vector of cold energy corresponding to the cold energy feature vector through a gated linear unit to obtain a gated cold energy feature vector of cold energy corresponding to the cold energy feature vector; and perform residual connection processing on the gated cold energy feature vector of cold energy corresponding to the cold energy feature vector and the cold energy feature vector corresponding to the cold energy unique heat vector to obtain a control vector of cold energy corresponding to the cold energy feature vector.

[0095] In some embodiments, the above-mentioned cooling capacity determination device is further configured to: acquire multiple training sample data sets and labels corresponding to each training sample data set, wherein each training sample data set includes multiple environmental monitoring training data and multiple chiller plant equipment operation training data, and the labels corresponding to the training sample data sets are used to characterize the actual cooling capacity corresponding to the training sample data sets; perform data cleaning processing on each training sample data set to obtain at least one cleaned environmental monitoring training data and at least one cleaned chiller plant equipment operation training data; perform one-hot encoding processing on each cleaned environmental monitoring training data and each cleaned chiller plant equipment operation training data to obtain at least one cooling capacity one-hot training vector; perform linear transformation processing on the at least one cooling capacity one-hot training vector to obtain cooling capacity feature training vectors corresponding to each cooling capacity one-hot training vector; and process the cooling capacity feature training vectors corresponding to each cooling capacity one-hot training vector through a gated residual network to obtain control vectors corresponding to each cooling capacity feature training vector. The cold quantity feature training vectors corresponding to each unique cold quantity training vector are concatenated to obtain the concatenated vector. A gated residual network is then used to process this concatenated vector to obtain the weight vector. The weight vector is then exponentially normalized to obtain the variable selection weights for each cold quantity feature training vector. Based on a predefined mapping relationship, the control vector and the variable selection weights for each cold quantity feature training vector are concatenated to obtain the target feature vector. A fully connected layer is then used to process the target feature vector to obtain the training cold quantity data. The loss of the cold quantity determination model is determined based on the root mean square error loss function, the training cold quantity data, and the labels corresponding to the training sample dataset. Finally, the parameters in the cold quantity determination model are updated iteratively based on the loss of the cold quantity determination model.

[0096] In some embodiments, the above-mentioned cooling capacity determination device is further configured to: acquire at least one real-time environmental monitoring data and at least one real-time cooling station equipment operation data; perform data cleaning processing on the at least one real-time environmental monitoring data and at least one real-time cooling station equipment operation data to obtain at least one cleaned real-time environmental monitoring data and at least one cleaned real-time cooling station equipment operation data; perform unique thermal encoding processing on each cleaned real-time environmental monitoring data and each cleaned real-time cooling station equipment operation data to obtain at least one real-time cooling capacity unique thermal vector; and process the at least one real-time cooling capacity unique thermal vector according to a preset sliding window and a preset cooling capacity determination time period through a cooling capacity determination model to obtain real-time cooling capacity data.

[0097] In some embodiments, the above-described cooling capacity determination device is further configured to: store real-time cooling capacity data in a preset database; filter the real-time cooling capacity data stored in the preset database based on a preset data accumulation time to obtain real-time cooling capacity data corresponding to the preset data accumulation time; and update the parameters in the cooling capacity determination model based on the real-time cooling capacity data corresponding to the preset data accumulation time.

[0098] In some embodiments, data cleaning processing is performed on various environmental monitoring data and various refrigeration station equipment operation data to obtain at least one cooling capacity-related clean data. Specifically, negative value truncation processing is performed on the environmental monitoring cooling capacity data to obtain non-negative cooling capacity data; exploratory data analysis processing is performed on the indoor temperature and humidity data to obtain indoor temperature and humidity analysis data; and indoor wet-bulb temperature data is determined based on the indoor temperature and humidity data.

[0099] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this disclosure.

[0100] Figure 4 This is a schematic diagram of the electronic device 4 provided in an embodiment of this disclosure. Figure 4 As shown, the electronic device 4 of this embodiment includes: a processor 401, a memory 402, and a computer program 403 stored in the memory 402 and executable on the processor 401. When the processor 401 executes the computer program 403, it implements the steps in the various method embodiments described above. Alternatively, when the processor 401 executes the computer program 403, it implements the functions of each module / unit in the various device embodiments described above.

[0101] Electronic device 4 can be a desktop computer, laptop, handheld computer, cloud server, or other electronic device. Electronic device 4 may include, but is not limited to, processor 401 and memory 402. Those skilled in the art will understand that... Figure 4 This is merely an example of electronic device 4 and does not constitute a limitation on electronic device 4. It may include more or fewer components than shown, or different components.

[0102] The processor 401 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0103] The memory 402 can be an internal storage unit of the electronic device 4, such as a hard disk or RAM of the electronic device 4. The memory 402 can also be an external storage device of the electronic device 4, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, Flash Card, etc., equipped on the electronic device 4. The memory 402 can also include both internal and external storage units of the electronic device 4. The memory 402 is used to store computer programs and other programs and data required by the electronic device.

[0104] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0105] If integrated modules / units are implemented as software functional units and sold or used as independent products, they can be stored in a readable storage medium (e.g., a computer-readable storage medium). Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program may include computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. A computer-readable storage medium may include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0106] The above embodiments are only used to illustrate the technical solutions of this disclosure, and are not intended to limit it. Although this disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this disclosure, and should all be included within the protection scope of this disclosure.

Claims

1. A method for determining cooling capacity, characterized in that, include: At least one cold quantity unique heat vector is input into the trained cold quantity determination model, and a linear transformation is performed on the at least one cold quantity unique heat vector to obtain the cold quantity feature vector corresponding to each cold quantity unique heat vector. By processing the cooling feature vectors corresponding to each of the cooling feature vectors using a gated residual network, the control vectors corresponding to each of the cooling feature vectors are obtained. The cold energy feature vectors corresponding to each of the cold energy unique heat vectors are concatenated to obtain the concatenated vectors corresponding to the cold energy feature vectors. The concatenated vector corresponding to the cooling capacity feature vector is processed by the gated residual network to obtain the weight vector corresponding to the cooling capacity feature vector. The weight vectors corresponding to the cooling capacity feature vectors are subjected to exponential normalization to obtain the variable selection weight values ​​corresponding to each cooling capacity feature vector. According to the preset mapping relationship, the control vectors corresponding to each of the cooling capacity feature vectors and the variables corresponding to each of the cooling capacity feature vectors are selected and weighted and concatenated to obtain the target feature vectors corresponding to the cooling capacity feature vectors. The target feature vector corresponding to the cold energy feature vector is processed by a fully connected layer to obtain the target cold energy data; Before inputting at least one unique cold quantity vector into the trained cold quantity determination model and performing a linear transformation on the at least one unique cold quantity vector to obtain the cold quantity feature vector corresponding to each unique cold quantity vector, the method further includes: Obtain at least one environmental monitoring data point and at least one cooling plant equipment operation data point; The environmental monitoring data and the operating data of the cooling station equipment are cleaned to obtain at least one cleaned environmental monitoring data and at least one cleaned cooling station equipment operating data. The environmental monitoring data and the operating data of the cooling station equipment after each cleaning are processed by unique thermal encoding to obtain at least one unique thermal vector of cooling capacity.

2. The method for determining cooling capacity according to claim 1, characterized in that, The gated residual network includes a first fully connected layer, a second fully connected layer, and a gated linear unit; The step of processing the cooling feature vectors corresponding to each of the cooling unique vectors through a gated residual network to obtain the control vectors corresponding to each of the cooling feature vectors includes: The first fully connected layer processes the cold energy feature vector corresponding to the cold energy unique heat vector to obtain the cold energy first fully connected vector corresponding to the cold energy feature vector. The first fully connected vector of cold energy corresponding to the cold energy feature vector is subjected to nonlinear activation processing to obtain the activated feature vector corresponding to the cold energy feature vector. The activation feature vector corresponding to the cold energy feature vector is processed by the second fully connected layer to obtain the second fully connected vector of cold energy corresponding to the cold energy feature vector; The second fully connected vector of cold energy corresponding to the cold energy feature vector is regularized to obtain the regularized feature vector corresponding to the cold energy feature vector. The gated linear unit processes the regularized feature vector corresponding to the cooling capacity feature vector to obtain the gated cooling capacity feature vector corresponding to the cooling capacity feature vector. The gated cold energy feature vector corresponding to the cold energy feature vector and the cold energy feature vector corresponding to the cold energy unique heat vector are subjected to residual connection processing to obtain the control vector corresponding to the cold energy feature vector.

3. The method for determining cooling capacity according to claim 1, characterized in that, Before inputting at least one unique cold quantity vector into the trained cold quantity determination model and performing a linear transformation on the at least one unique cold quantity vector to obtain the cold quantity feature vector corresponding to each unique cold quantity vector, the method further includes: Multiple training sample data sets and labels corresponding to each training sample data set are obtained. Each training sample data set contains multiple environmental monitoring training data and multiple cooling station equipment operation training data. The labels corresponding to the training sample data sets are used to characterize the actual cooling capacity corresponding to the training sample data sets. Data cleaning processing is performed on each of the training sample datasets to obtain at least one cleaned environmental monitoring training data and at least one cleaned cold station equipment operation training data. The environmental monitoring training data and the operating training data of the cold station equipment after cleaning are processed by unique thermal encoding to obtain at least one unique thermal training vector of cooling capacity. Perform a linear transformation on at least one of the cold-energy-only-hot training vectors to obtain the cold-energy feature training vectors corresponding to each of the cold-energy-only-hot training vectors. The gated residual network is used to process the cold quantity feature training vectors corresponding to each of the cold quantity unique heat training vectors to obtain the control vectors corresponding to each of the cold quantity feature training vectors. The cold quantity feature training vectors corresponding to each of the cold quantity unique heat training vectors are concatenated to obtain the concatenated vectors corresponding to the cold quantity feature training vectors. The concatenated vector corresponding to the cold quantity feature training vector is processed by the gated residual network to obtain the weight vector corresponding to the cold quantity feature training vector. The weight vectors corresponding to the cold energy feature training vectors are subjected to exponential normalization to obtain the variable selection weight values ​​corresponding to each cold energy feature training vector. According to the preset mapping relationship, the control vectors corresponding to each of the cold energy feature training vectors and the variables corresponding to each of the cold energy feature vectors are selected and weighted and concatenated to obtain the target feature vector corresponding to the cold energy feature vector. The target feature vector corresponding to the cold volume feature training vector is processed by the fully connected layer to obtain training cold volume data; The loss of the cold quantity determination model is determined based on the root mean square error loss function, the training cold quantity data, and the labels corresponding to the training sample data set. The parameters in the cooling capacity determination model are updated based on the loss of the cooling capacity determination model through iterative loops.

4. The method for determining cooling capacity according to claim 1, characterized in that, The method further includes: Acquire at least one real-time environmental monitoring data point and at least one real-time cooling plant equipment operation data point; Data cleaning processing is performed on at least one of the real-time environmental monitoring data and at least one of the real-time cooling plant equipment operation data to obtain at least one cleaned real-time environmental monitoring data and at least one cleaned real-time cooling plant equipment operation data. The real-time environmental monitoring data and the real-time cooling station equipment operation data after each cleaning are processed by unique thermal encoding to obtain at least one real-time cooling capacity unique thermal vector. Based on a preset sliding window and a preset cooling capacity determination time period, at least one of the real-time cooling capacity unique heat vectors is processed by the cooling capacity determination model to obtain real-time cooling capacity data.

5. The method for determining cooling capacity according to claim 4, characterized in that, The method further includes: The real-time cooling capacity data is stored in a preset database; Based on a preset data accumulation time, the real-time cooling data stored in the preset database is filtered to obtain the real-time cooling data corresponding to the preset data accumulation time. The parameters in the cooling capacity determination model are updated based on the real-time cooling capacity data corresponding to the preset data accumulation time.

6. The method for determining cooling capacity according to claim 1, characterized in that, The environmental monitoring data includes indoor temperature and humidity data; the chiller station equipment operation data includes cooling capacity data; the environmental monitoring data after cleaning includes indoor temperature and humidity analysis data and indoor wet-bulb temperature data; the chiller station equipment operation data after cleaning includes non-negative cooling capacity data; The process of cleaning the environmental monitoring data and the operating data of the chiller equipment to obtain at least one set of cooling capacity-related cleaned data includes: The negative values ​​of the cooling energy data are truncated to obtain the non-negative cooling energy data. The indoor temperature and humidity data are subjected to exploratory data analysis to obtain the indoor temperature and humidity analysis data; Based on the indoor temperature and humidity data, determine the indoor wet-bulb temperature data.

7. A device for determining cooling capacity, characterized in that, include: The first processing module is used to input at least one cold quantity unique heat vector into the trained cold quantity determination model, and perform linear transformation processing on the at least one cold quantity unique heat vector to obtain the cold quantity feature vector corresponding to each cold quantity unique heat vector. The second processing module is used to process the cooling feature vectors corresponding to each of the cooling unique vectors through a gated residual network to obtain the control vectors corresponding to each of the cooling feature vectors. The third processing module is used to concatenate the cold energy feature vectors corresponding to each of the cold energy unique heat vectors to obtain the concatenated vectors corresponding to the cold energy feature vectors. The fourth processing module is used to process the concatenated vector corresponding to the cooling capacity feature vector through the gated residual network to obtain the weight vector corresponding to the cooling capacity feature vector. The fifth processing module is used to perform exponential normalization on the weight vectors corresponding to the cooling capacity feature vectors to obtain the variable selection weight values ​​corresponding to each cooling capacity feature vector. The determining module is used to concatenate the control vectors corresponding to each of the cooling capacity feature vectors and the variables corresponding to each of the cooling capacity feature vectors according to a preset mapping relationship, so as to obtain the target feature vectors corresponding to the cooling capacity feature vectors. The sixth processing module is used to process the target feature vector corresponding to the cooling feature vector through a fully connected layer to obtain the target cooling data; The cooling capacity determination device is further configured to: acquire at least one environmental monitoring data and at least one cooling station equipment operation data; perform data cleaning processing on each of the environmental monitoring data and each of the cooling station equipment operation data to obtain at least one cleaned environmental monitoring data and at least one cleaned cooling station equipment operation data; and perform unique thermal encoding processing on each of the cleaned environmental monitoring data and each of the cleaned cooling station equipment operation data to obtain at least one unique thermal vector of cooling capacity.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 6.

9. A readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 6.

Images