Intelligent control system and method for seawater ice making machine

Abstract

An intelligent control system and method for a seawater ice making machine. First, a time series of ice layer thickness and a time series of ambient temperature are arranged into an ice layer thickness time series input vector and an ambient temperature time series input vector, respectively, according to time dimensions; then, a sample covariance matrix of the ice layer thickness time series input vector relative to the ambient temperature time series input vector is calculated, so as to obtain an ice layer thickness-ambient temperature time series collective matrix; then, feature extraction and attention reinforcement are performed on the ice layer thickness-ambient temperature time series collective matrix by means of an ice layer thickness-ambient temperature time series correlation mode feature extractor based on a deep neural network model, so as to obtain a reinforced ice layer thickness-ambient temperature time series correlation feature; finally, whether to automatically start a defrost cycle is determined on the basis of the reinforced ice layer thickness-ambient temperature time series correlation feature.

Description

Intelligent control system and method of seawater ice maker TECHNICAL FIELD

[0001] The present disclosure relates to the field of ice makers, in particular to an intelligent control system and method of seawater ice maker. BACKGROUND

[0002] A seawater ice maker is an ice maker that uses seawater as a refrigerant. Seawater ice maker has the advantages of fast ice making speed, low energy consumption, stable operation, etc., and is widely used in fishery, food processing, medicine and other industries.

[0003] During the operation of the seawater ice maker, the ice layer will continue to thicken. When the thickness of the ice layer reaches a certain degree, defrosting operation is needed to prevent the ice layer from being too thick to affect the ice making efficiency. The traditional defrosting method is timed defrosting, that is, defrosting operation is performed every certain period of time. This method is simple and easy to implement, but it has the problem of excessive defrosting, which can cause water waste, increase energy consumption, and affect the ice making efficiency and service life of the ice maker.

[0004] Therefore, an intelligent control system of seawater ice maker is desired. TECHNICAL PROBLEM

[0005] The present disclosure is made in view of the above problems. One object of the present disclosure is to provide an intelligent control system and method of seawater ice maker. TECHNICAL SOLUTION

[0006] Embodiments of the present disclosure provide an intelligent control system of seawater ice maker, comprising:

[0007] a data acquisition module configured to obtain a time series of ice layer thickness and a time series of ambient temperature;

[0008] a data time sequence arrangement module configured to arrange the time series of ice layer thickness and the time series of ambient temperature into an ice layer thickness time sequence input vector and an ambient temperature time sequence input vector, respectively, according to the time dimension;

[0009] an ice layer thickness-ambient temperature time sequence cooperative correlation module configured to calculate a sample covariance matrix of the ice layer thickness time sequence input vector relative to the ambient temperature time sequence input vector to obtain an ice layer thickness-ambient temperature time sequence cooperative matrix;

[0010] an ice layer thickness-ambient temperature time sequence correlation feature extraction module configured to extract features from the ice layer thickness-ambient temperature time sequence cooperative matrix by an ice layer thickness-ambient temperature time sequence correlation pattern feature extractor based on a deep neural network model to obtain an ice layer thickness-ambient temperature time sequence correlation feature map;

[0011] an attention reinforcement module configured to perform attention reinforcement on the ice layer thickness-environment temperature time series correlation feature map to obtain a reinforced ice layer thickness-environment temperature time series correlation feature; and

[0012] a defrost cycle control module configured to determine whether to automatically start a defrost cycle based on the reinforced ice layer thickness-environment temperature time series correlation feature.

[0013] For example, the intelligent control system of the seawater ice maker according to the embodiments of the present disclosure, wherein the ice layer thickness-environment temperature time series correlation module is configured to: calculate a sample covariance matrix of the ice layer thickness time series input vector with respect to the environment temperature time series input vector to obtain the ice layer thickness-environment temperature time series correlation matrix by using a sample covariance formula as follows: wherein, is the ice layer thickness time series input vector, is the environment temperature time series input vector, is the ice layer thickness-environment temperature time series correlation matrix, denotes a transposition operation. For example, the intelligent control system of the seawater ice maker according to the embodiments of the present disclosure, wherein the deep neural network model is a convolutional neural network model.

[0014] For example, the intelligent control system of the seawater ice maker according to the embodiments of the present disclosure, wherein the attention reinforcement module is configured to: input the ice layer thickness-environment temperature time series correlation feature map into a feature reinforcement device based on an Asm-CBAM attention module to obtain a reinforced ice layer thickness-environment temperature time series correlation feature map as the reinforced ice layer thickness-environment temperature time series correlation feature.

[0015] For example, the intelligent control system of the seawater ice maker according to the embodiments of the present disclosure, wherein the attention reinforcement module is configured to: input the ice layer thickness-environment temperature time series correlation feature map into the feature reinforcement device based on the Asm-CBAM attention module to process the ice layer thickness-environment temperature time series correlation feature map by using a reinforcement formula as follows to obtain the reinforced ice layer thickness-environment temperature time series correlation feature map; wherein the reinforcement formula is: wherein, is the ice layer thickness-environment temperature time series correlation feature map, , and denote a mean-pooling feature map, a max-pooling feature map and a random-pooling feature map obtained by performing global mean-pooling, global max-pooling and global random-pooling on the ice layer thickness-environment temperature time series correlation feature map in a channel dimension, and MLP denotes a multi-layer perceptron with different numbers of channels, for the number of channels, is a Sigmoid activation function, is the reinforced ice layer thickness-environment temperature time series correlation feature map.

[0016] For example, the intelligent control system of the seawater ice maker according to the embodiments of the present disclosure, wherein the defrosting cycle control module is configured to: input the reinforced ice layer thickness-environment temperature time series correlation feature map into the classifier-based defrosting controller to obtain a control instruction, and the control instruction is configured to indicate whether to automatically start the defrosting cycle.

[0017] For example, the intelligent control system of the seawater ice maker according to the embodiments of the present disclosure, further comprising a training module configured to train the ice layer thickness-environment temperature time series correlation pattern feature extractor based on the convolutional neural network model, the feature reinforcement device based on the Asm-CBAM attention module, and the classifier-based defrosting controller.

[0018] For example, the intelligent control system of the seawater ice maker according to the embodiments of the present disclosure, wherein the training module comprises:

[0019] a training data acquisition unit configured to acquire training data, and the training data comprises a time series of training ice layer thickness and a time series of training environment temperature;

[0020] a training data time series arrangement unit configured to arrange the time series of training ice layer thickness and the time series of training environment temperature into a training ice layer thickness time series input vector and a training environment temperature time series input vector, respectively, according to a time dimension;

[0021] a training ice layer thickness-environment temperature time series cooperative correlation unit configured to calculate a sample covariance matrix of the training ice layer thickness time series input vector relative to the training environment temperature time series input vector to obtain a training ice layer thickness-environment temperature time series cooperative matrix;

[0022] a training ice layer thickness-environment temperature time series correlation feature extraction unit configured to extract features of the training ice layer thickness-environment temperature time series cooperative matrix by the ice layer thickness-environment temperature time series correlation pattern feature extractor based on the convolutional neural network model to obtain a training ice layer thickness-environment temperature time series correlation feature map;

[0023] a training attention reinforcement unit configured to input the training ice layer thickness-environment temperature time series correlation feature map into the feature reinforcement device based on the Asm-CBAM attention module to obtain a training reinforced ice layer thickness-environment temperature time series correlation feature map;

[0024] a feature optimization unit configured to optimize the training reinforced ice layer thickness-environment temperature time series correlation feature map to obtain an optimized training reinforced ice layer thickness-environment temperature time series correlation feature map;

[0025] a classification training unit configured to pass the optimized training reinforced ice layer thickness-environment temperature time series correlation feature map through the classifier-based defrosting controller to obtain a classification loss function value; and

[0026] a loss training unit configured to train the convolutional neural network model-based ice layer thickness-environment temperature time series correlation pattern feature extractor, the Asm-CBAM attention module-based feature enhancer, and the classifier-based defrosting controller based on the classification loss function value.

[0027] Embodiments of the present disclosure also provide an intelligent control method of a seawater ice maker, which comprises:

[0028] obtaining a time series of ice layer thickness and a time series of environment temperature;

[0029] arranging the time series of ice layer thickness and the time series of environment temperature into an ice layer thickness time series input vector and an environment temperature time series input vector, respectively, according to the time dimension;

[0030] calculating a sample covariance matrix of the ice layer thickness time series input vector relative to the environment temperature time series input vector to obtain an ice layer thickness-environment temperature time series correlation matrix;

[0031] extracting features from the ice layer thickness-environment temperature time series correlation matrix by an ice layer thickness-environment temperature time series correlation pattern feature extractor based on a deep neural network model to obtain an ice layer thickness-environment temperature time series correlation feature map;

[0032] reinforcing attention on the ice layer thickness-environment temperature time series correlation feature map to obtain reinforced ice layer thickness-environment temperature time series correlation features; and

[0033] determining whether to automatically start a defrosting cycle based on the reinforced ice layer thickness-environment temperature time series correlation features. Advantages

[0034] The intelligent control system of a seawater ice maker according to embodiments of the present disclosure can monitor the ice layer thickness and the environment temperature in real time, and perform time series correlation analysis of the ice layer thickness and the environment temperature by introducing data processing and analysis algorithms in the backend, so as to automatically determine whether to perform defrosting operation according to the ice layer thickness, the environment temperature, and other factors. In this way, the defrosting cycle can be started only when necessary, preventing over-defrosting, thereby saving water resources and reducing energy consumption, and improving the ice making efficiency and service life of the ice maker. BRIEF DESCRIPTION OF DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments of the present disclosure will be briefly introduced below. Obviously, the drawings in the following description are only related to some embodiments of the present disclosure, not a limitation on the present disclosure.

[0036] FIG. 1 shows an application architecture schematic diagram of the intelligent control system of the seawater ice maker in the embodiments of the present disclosure;

[0037] FIG. 2 shows a structural schematic diagram of the intelligent control system of the seawater ice maker in the embodiments of the present disclosure;

[0038] FIG. 3 shows a flowchart of the intelligent control method of the seawater ice maker in the embodiments of the present disclosure;

[0039] FIG. 4 shows an application scenario diagram of the intelligent control system of the seawater ice maker in the embodiments of the present disclosure; and

[0040] FIG. 5 shows a schematic diagram of a storage medium according to an embodiment of the present disclosure. Embodiments of the present invention

[0041] The technical solutions in the embodiments of the present disclosure will be described clearly and completely below with reference to the drawings, and obviously, the described embodiments are only some of the embodiments of the present disclosure, not all the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by those of ordinary skill in the art without creative labor also belong to the scope of protection of the present disclosure.

[0042] The terms used in the specification are those general terms currently widely used in the art in consideration of the functions related to the present disclosure, but the terms can be changed according to the intention of those of ordinary skill in the art, precedents, or new technology in the art. In addition, specific terms can be selected by the applicant, and in this case, the detailed meaning thereof will be described in the detailed description of the present disclosure. Therefore, the terms used in the specification should not be understood as simple names, but based on the meaning of the terms and the overall description of the present disclosure.

[0043] Although the present disclosure makes various references to certain modules in the system according to the embodiments of the present disclosure, however, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0044] Flowcharts in the present disclosure are used to illustrate the operations performed by the system according to embodiments of the present disclosure. It should be understood that the foregoing or following operations are not necessarily performed in sequence. Instead, various steps can be processed in reverse order or simultaneously, as needed. Meanwhile, other operations can also be added to these processes, or one or more steps of operations can be removed from these processes.

[0045] FIG. 1 shows an application architecture diagram of the intelligent control system of the seawater ice maker in embodiments of the present disclosure, including a server 100 and a terminal device 200.

[0046] The terminal device 200 and the server 100 can be connected through the Internet to realize communication between each other. Optionally, the Internet uses standard communication technologies and / or protocols. The Internet is usually the Internet, but can also be any network, including but not limited to a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a mobile, wired or wireless network, a private network or any combination of virtual private networks. In some embodiments, technologies and / or formats including Hyper Text Markup Language (HTML), Extensible Markup Language (XML), etc. are used to represent data exchanged through the network. In addition, all or some links can be encrypted using conventional encryption technologies such as Secure Socket Layer (SSL), Transport Layer Security (TLS), Virtual Private Network (VPN), Internet Protocol Security (IPsec), etc. In other embodiments, custom and / or dedicated data communication technologies can be used instead of or in addition to the above data communication technologies.

[0047] The server 100 can provide various network services for the terminal device 200, wherein the server 100 can be a server, a server cluster composed of several servers or a cloud computing center. Specifically, the server 100 can include a processor 110 (Center Processing Unit, CPU), a memory 120, an input device 130 and an output device 140, etc., the input device 130 can include a keyboard, a mouse, a touch screen, etc., and the output device 140 can include a display device such as a Liquid Crystal Display (LCD), a Cathode Ray Tube (CRT), etc.

[0048] The memory 120 can include a Read Only Memory (ROM) and a Random Access Memory (RAM), and provide the processor 110 with program instructions and data stored in the memory 120. In the embodiments of the present disclosure, the memory 120 can be used to store the program of the intelligent control method of the seawater ice maker in the embodiments of the present disclosure.

[0049] The processor 110 processes the steps of any intelligent control method of the seawater ice maker in the embodiments of the present disclosure by calling the program instructions stored in the memory 120.

[0050] In addition, the application architecture diagram in the embodiments of the present disclosure is used to more clearly illustrate the technical solutions in the embodiments of the present disclosure, and does not constitute a limitation on the technical solutions provided by the embodiments of the present disclosure. Of course, for other application architectures and business applications, the technical solutions provided by the embodiments of the present disclosure are also applicable to similar problems.

[0051] The intelligent control system of the seawater ice maker according to at least one embodiment of the present disclosure is non-limitingly described through several examples or embodiments. As described below, different features in these specific examples or embodiments can be combined with each other without mutual contradiction, so as to obtain new examples or embodiments, and these new examples or embodiments also belong to the protection scope of the present disclosure.

[0052] To solve the above technical problems, the technical concept of the present application is to monitor the ice layer thickness and the environmental temperature in real time, introduce data processing and analysis algorithms in the backend to perform time sequence collaborative analysis on the ice layer thickness and the environmental temperature, and automatically determine whether to perform defrosting operation according to the ice layer thickness, the environmental temperature and other factors. In this way, defrosting cycle can be started only when necessary, over-defrosting can be prevented, water resources can be saved, energy consumption can be reduced, and the ice making efficiency and service life of the ice maker can be improved.

[0053] FIG. 2 shows a structural schematic diagram of the intelligent control system 800 of the seawater ice maker in the embodiments of the present disclosure. The intelligent control system 800 of the seawater ice maker comprises: a data acquisition module 810, configured to acquire a time sequence of ice layer thickness and a time sequence of environmental temperature; a data time sequence arrangement module 820, configured to arrange the time sequence of ice layer thickness and the time sequence of environmental temperature into an ice layer thickness time sequence input vector and an environmental temperature time sequence input vector, respectively, according to the time dimension; an ice layer thickness-environmental temperature time sequence collaborative correlation module 830, configured to calculate a sample covariance matrix of the ice layer thickness time sequence input vector relative to the environmental temperature time sequence input vector to obtain an ice layer thickness-environmental temperature time sequence collaborative matrix; an ice layer thickness-environmental temperature time sequence correlation feature extraction module 840, configured to perform feature extraction on the ice layer thickness-environmental temperature time sequence collaborative matrix by an ice layer thickness-environmental temperature time sequence correlation pattern feature extractor based on a deep neural network model to obtain an ice layer thickness-environmental temperature time sequence correlation feature map; an attention reinforcement module 850, configured to perform attention reinforcement on the ice layer thickness-environmental temperature time sequence correlation feature map to obtain a reinforced ice layer thickness-environmental temperature time sequence correlation feature; and a defrosting cycle control module 860, configured to determine whether to automatically start a defrosting cycle based on the reinforced ice layer thickness-environmental temperature time sequence correlation feature.

[0054] It can be understood that the data acquisition module 810, the data time sequence arrangement module 820, the ice layer thickness-environmental temperature time sequence collaborative correlation module 830, the ice layer thickness-environmental temperature time sequence correlation feature extraction module 840, the attention reinforcement module 850, and the defrosting cycle control module 860 collectively constitute an intelligent control system, and through data acquisition, time sequence arrangement, collaborative correlation, feature extraction, attention reinforcement, and other steps, intelligent control of the seawater ice maker and automatic decision of the defrosting cycle are realized.

[0055] Specifically, in the technical solution of the present application, first, a time sequence of ice layer thickness and a time sequence of environmental temperature are acquired. Then, considering that the ice layer thickness and the environmental temperature both have time sequence dynamic change rules in the time dimension, that is, the time sequence data of the ice layer thickness and the time sequence data of the environmental temperature contain the change pattern and the correlation relationship of the data in time. Therefore, in order to analyze and capture the time sequence features of the ice layer thickness and the environmental temperature, in the technical solution of the present application, the time sequence of the ice layer thickness and the time sequence of the environmental temperature need to be arranged into an ice layer thickness time sequence input vector and an environmental temperature time sequence input vector, respectively, according to the time dimension.

[0056] It should be understood that, due to the mutual correlation and influence between the ice layer thickness and the ambient temperature. In order to capture the correlation and synergistic change between the ice layer thickness and the ambient temperature, so as to help understand the correlation between the ice layer thickness and the ambient temperature, and to more accurately determine whether defrosting operation is needed, in the technical solution of the present application, the sample covariance matrix of the ice layer thickness time sequence input vector relative to the ambient temperature time sequence input vector is further calculated to obtain the ice layer thickness-ambient temperature time sequence synergy matrix. It should be understood that the covariance matrix can show the linear relationship between different variables. By calculating the sample covariance matrix between the ice layer thickness and the ambient temperature time sequence input vector, the degree of correlation between them can be revealed, so that different working modes and trends can be better identified, which is beneficial to more accurately determine when to start the defrosting cycle.

[0057] Correspondingly, the ice layer thickness-ambient temperature time sequence synergy correlation module 830 is configured to calculate the sample covariance matrix of the ice layer thickness time sequence input vector relative to the ambient temperature time sequence input vector to obtain the ice layer thickness-ambient temperature time sequence synergy matrix by using the following sample covariance formula: wherein, is the ice layer thickness time sequence input vector, is the ambient temperature time sequence input vector, is the ice layer thickness-ambient temperature time sequence synergy matrix, denotes the transposition operation. For example, the intelligent control system of the seawater ice maker according to an embodiment of the present disclosure, wherein the deep neural network model is a convolutional neural network model.

[0058] It is worth mentioning that the sample covariance matrix is a statistical quantity used to measure the relationship or change trend between two random variables. In statistics, covariance is used to measure the overall error of two variables. The sample covariance matrix is a covariance matrix calculated from sample data. The sample covariance matrix is usually a symmetric matrix, and the elements on the diagonal are the variances of the variables, and the elements on the non-diagonal are the covariances between the variables. The sample covariance matrix can help understand the linear relationship between variables and their change trend.

[0059] Then, the ice layer thickness-ambient temperature time sequence synergy matrix is subjected to feature mining in the ice layer thickness-ambient temperature time sequence correlation pattern feature extractor based on the convolutional neural network model, to extract the time sequence synergy correlation feature information between the ice layer thickness and the ambient temperature, thereby obtaining an ice layer thickness-ambient temperature time sequence correlation feature map.

[0060] Correspondingly, in the ice layer thickness-environmental temperature time sequence correlation feature extraction module 840, the deep neural network model is a convolutional neural network model, that is, the ice layer thickness-environmental temperature time sequence correlation pattern feature extractor based on the deep neural network model is an ice layer thickness-environmental temperature time sequence correlation pattern feature extractor based on a convolutional neural network model. It is worth mentioning that the convolutional neural network (CNN) is a deep learning model, which is particularly suitable for processing data with grid structure. The core idea of the convolutional neural network model is to use convolutional layers (Convolutional Layer) and pooling layers (Pooling Layer) to learn local features and spatial structures in images, so as to realize the learning and recognition of complex patterns. Its main components include: 1. Convolutional layer: by applying a series of filters (convolution kernel) to the input convolution operation, local features are extracted, and the convolution operation can capture the spatial structure information of the input data. 2. Activation function: usually after the convolutional layer, a nonlinear activation function (such as ReLU) is added to introduce nonlinear characteristics and increase the expression ability of the model. 3. Pooling layer: through pooling operation (such as max pooling or average pooling), the spatial dimension of the feature map is reduced, the number of parameters is reduced, and important features are preserved. 4. Fully connected layer: usually at the top of the convolutional neural network model, a fully connected layer is connected to map the features extracted by the convolutional layer to the final output. The convolutional neural network model gradually extracts and combines the features of the input data through multiple convolution and pooling operations, so as to realize the learning and recognition of complex patterns.

[0061] Further, in order to improve the expression ability of the network, in the technical solution of the present application, the ice layer thickness-environmental temperature time sequence correlation feature map is further input through a feature enhancer based on an Asm-CBAM attention module to obtain an enhanced ice layer thickness-environmental temperature time sequence correlation feature map. In particular, the Asm-CBAM attention module can add random pooling based on the original CBAM with only max pooling and average pooling. Random pooling randomly selects elements in the feature map according to their probability values, which is different from taking only the maximum value element in max pooling. In this way, the maximum value can be maximized, and it is ensured that not all elements are selected as the maximum value, thereby improving the generalization ability of the network and obtaining more rich deep features between ice layer thickness and environmental temperature. In this way, the correlation and changes between ice layer thickness and environmental temperature can be better captured, the feature representation ability can be improved, noise interference can be reduced, and the performance and intelligent level of the intelligent control system can be improved.

[0062] Correspondingly, the attention reinforcement module 850 is configured to input the ice thickness-environment temperature time correlation feature map into a feature enhancer based on an Asm-CBAM attention module to obtain an enhanced ice thickness-environment temperature time correlation feature map as the enhanced ice thickness-environment temperature time correlation feature.

[0063] It is worth mentioning that the Asm-CBAM attention module is a Channel-wise Attention Module (CBAM) based attention enhancement module used to enhance the representation ability of the feature map. The CBAM module combines channel attention and spatial attention, which can effectively learn the importance of the feature map in the channel dimension and the spatial dimension, and weight the feature map according to the learned attention weights. The CBAM module usually includes two sub-modules: 1. Channel Attention Module: The channel attention module learns the importance of each channel, so that the network can better focus on important channel features. The channel attention module uses global average pooling to obtain global information of each channel, then learns channel attention weights through a fully connected layer and an activation function, and finally applies these weights to each channel. 2. Spatial Attention Module: The spatial attention module is used to learn the importance of different spatial positions in the feature map, so that the network can better focus on important spatial regions. The spatial attention module usually learns spatial weights by using convolution operations, and then applies these weights to the feature map. The Asm-CBAM attention module is improved and enhanced on the basis of CBAM to further improve the representation ability of the feature map. By introducing the Asm-CBAM attention module, the model can better capture important information in the input feature map and improve the performance and generalization ability of the model in the ice thickness-environment temperature time correlation feature extraction task.

[0064] Specifically, the attention reinforcement module 850 is configured to input the ice thickness-environment temperature time correlation feature map into the feature enhancer based on the Asm-CBAM attention module to process it according to the following enhancement formula to obtain the enhanced ice thickness-environment temperature time correlation feature map. wherein, is the ice thickness-environment temperature time correlation feature map, , and respectively represent the mean-pooled feature map, the maximum-pooled feature map and the random-pooled feature map obtained by performing global mean-pooling, global maximum-pooling and global random-pooling on the ice thickness-environment temperature time correlation feature map in the channel dimension, and respectively represent multi-layer perceptrons with different numbers of channels, is the number of channels, is a Sigmoid activation function, is the reinforced ice thickness-environment temperature time correlation feature map.

[0065] Then, the reinforced ice thickness-environment temperature time correlation feature map is input into a classifier-based defrosting controller to obtain a control instruction, which indicates whether to automatically start a defrosting cycle. That is, the time correlation feature information between ice thickness and environment temperature is used for classification processing, so as to automatically determine whether defrosting operation is needed according to factors such as ice thickness and environment temperature, and generate a defrosting control instruction. In this way, the defrosting cycle can be started only when necessary, preventing over-defrosting, thereby saving water resources and reducing energy consumption, improving ice maker ice making efficiency and service life.

[0066] Correspondingly, the defrosting cycle control module 860 is configured to input the reinforced ice thickness-environment temperature time correlation feature map into a classifier-based defrosting controller to obtain a control instruction, which indicates whether to automatically start a defrosting cycle.

[0067] Specifically, the defrosting cycle control module 860 is further configured to expand the reinforced ice thickness-environment temperature time correlation feature map into a reinforced classification feature vector in a row vector or a column vector; use a fully connected layer of the classifier-based defrosting controller to perform fully connected coding on the reinforced classification feature vector to obtain a coded classification feature vector; and input the coded classification feature vector into a Softmax classification function of the classifier-based defrosting controller to obtain the control instruction.

[0068] It should be understood that the role of the classifier is to learn the classification rule and the classifier using the given class and known training data, and then classify (or predict) unknown data. Logistics, SVM, etc. are commonly used to solve binary classification problems. For multi-class classification, logistics or SVM can also be used, but multiple binary classifications are needed to form multi-classification, which is prone to errors and low efficiency. Common multi-classification methods include Softmax classification function.

[0069] Further, in the technical solutions of the present application, the intelligent control system of the seawater ice maker further comprises a training module for training the ice layer thickness-environment temperature time series correlation pattern feature extractor based on a convolutional neural network model, the feature enhancer based on an Asm-CBAM attention module, and the defrosting controller based on a classifier.

[0070] The training module comprises a training data acquisition unit for acquiring training data, the training data comprising a time series of training ice layer thickness and a time series of training environment temperature; a training data time series arrangement unit for arranging the time series of training ice layer thickness and the time series of training environment temperature into a training ice layer thickness time series input vector and a training environment temperature time series input vector, respectively, according to the time dimension; a training ice layer thickness-environment temperature time series cooperative correlation unit for calculating a sample covariance matrix of the training ice layer thickness time series input vector relative to the training environment temperature time series input vector to obtain a training ice layer thickness-environment temperature time series cooperative matrix; a training ice layer thickness-environment temperature time series correlation feature extraction unit for extracting features from the training ice layer thickness-environment temperature time series cooperative matrix by the ice layer thickness-environment temperature time series correlation pattern feature extractor based on a convolutional neural network model to obtain a training ice layer thickness-environment temperature time series correlation feature map; a training attention enhancement unit for inputting the training ice layer thickness-environment temperature time series correlation feature map through the feature enhancer based on an Asm-CBAM attention module to obtain a training enhanced ice layer thickness-environment temperature time series correlation feature map; a feature optimization unit for optimizing the training enhanced ice layer thickness-environment temperature time series correlation feature map to obtain an optimized training enhanced ice layer thickness-environment temperature time series correlation feature map; a classification training unit for inputting the optimized training enhanced ice layer thickness-environment temperature time series correlation feature map through the defrosting controller based on a classifier to obtain a classification loss function value; and a loss training unit for training the ice layer thickness-environment temperature time series correlation pattern feature extractor based on a convolutional neural network model, the feature enhancer based on an Asm-CBAM attention module, and the defrosting controller based on a classifier based on the classification loss function value.

[0071] In the technical solution of the present application, each feature matrix of the training ice layer thickness-environment temperature time sequence correlation feature map represents the local high-order correlation feature of the full time sequence covariance correlation of the training ice layer thickness and the training environment temperature, and the channel distribution of the convolutional neural network model is followed between each feature matrix. Thus, after the training ice layer thickness-environment temperature time sequence correlation feature map is input through the feature enhancer based on the Asm-CBAM attention module, due to the multiple channel attention mechanism of Asm-CBAM, the overall feature distribution of some feature matrices of the training enhanced ice layer thickness-environment temperature time sequence correlation feature map is enhanced, and significant channel distribution differences are also introduced, so that the distribution integrity of the training enhanced ice layer thickness-environment temperature time sequence correlation feature map in the unit of feature matrix feature distribution decreases, thereby affecting the class probability convergence effect of the training enhanced ice layer thickness-environment temperature time sequence correlation feature map through the classifier, that is, affecting the speed of classification training and the accuracy of the classification result. Based on this, the present application optimizes the training enhanced ice layer thickness-environment temperature time sequence correlation feature map.

[0072] Specifically, the feature optimization unit comprises: firstly unfolding the training enhanced ice layer thickness-environment temperature time sequence correlation feature map into a training enhanced ice layer thickness-environment temperature time sequence correlation feature vector, calculating the self-correlation matrix of the training enhanced ice layer thickness-environment temperature time sequence correlation feature vector and its transpose, and calculating the inner product of the first and the second row vectors of the self-correlation matrix as the matrix value of the first position of the weight matrix, so as to obtain the weight matrix, then performing matrix multiplication between the weight matrix and the self-correlation matrix, and further performing matrix-vector multiplication between the weight matrix and the training enhanced ice layer thickness-environment temperature time sequence correlation feature vector to obtain a correction vector, and finally performing dot multiplication between the correction vector and the training enhanced ice layer thickness-environment temperature time sequence correlation feature vector to obtain an optimized training enhanced ice layer thickness-environment temperature time sequence correlation feature vector, and restoring the optimized training enhanced ice layer thickness-environment temperature time sequence correlation feature vector to an optimized training enhanced ice layer thickness-environment temperature time sequence correlation feature map.

[0073] Thus, based on the self-association dimension of the training reinforced ice layer thickness-environment temperature time series correlation feature vector as the object to be modulated, the correlation expansion is performed based on the spatial sub-dimension complexity of the high-dimensional feature space of the feature distribution, thereby introducing the decomposable correlation dimension offset of the training reinforced ice layer thickness-environment temperature time series correlation feature vector into the heterogeneous correlation embedding space, so as to enhance the correlation self-consistency relationship of the decomposable dimension set represented by the heterogeneous correlation embedding space through joint fine-tuning, to improve the predetermined category compatibility of the feature vector of the training reinforced ice layer thickness-environment temperature time series correlation feature vector in the correlation target classification domain, thereby improving the accuracy of the classification result obtained by the training reinforced ice layer thickness-environment temperature time series correlation feature vector during model inference. In this way, it can be automatically judged whether defrosting operation is needed according to the ice layer thickness and environmental temperature and other factors, and a defrosting control instruction is generated. In this way, defrosting cycle can be started only when necessary, preventing over-defrosting, thereby saving water resources and reducing energy consumption, improving the ice making efficiency and service life of the ice maker.

[0074] Based on the above embodiment, referring to FIG. 3, a flowchart of an intelligent control method of a seawater ice maker according to an embodiment of the present disclosure is shown. For example, the intelligent control method of the seawater ice maker can be executed by a server, which can be the server 100 shown in FIG. 1. As shown in FIG. 3, the intelligent control method of the seawater ice maker according to an embodiment of the present disclosure includes the following steps: S510, obtaining a time series of ice layer thickness and a time series of environmental temperature; S520, arranging the time series of ice layer thickness and the time series of environmental temperature into an ice layer thickness time series input vector and an environmental temperature time series input vector, respectively, according to the time dimension; S530, calculating a sample covariance matrix of the ice layer thickness time series input vector relative to the environmental temperature time series input vector to obtain an ice layer thickness-environment temperature time series correlation matrix; S540, performing feature extraction on the ice layer thickness-environment temperature time series correlation matrix by an ice layer thickness-environment temperature time series correlation pattern feature extractor based on a deep neural network model to obtain an ice layer thickness-environment temperature time series correlation feature map; S550, performing attention reinforcement on the ice layer thickness-environment temperature time series correlation feature map to obtain a reinforced ice layer thickness-environment temperature time series correlation feature; and S560, determining whether to automatically start a defrosting cycle based on the reinforced ice layer thickness-environment temperature time series correlation feature.

[0075] Here, those skilled in the art can understand that the specific operations of each step in the above intelligent control method of the seawater ice maker have been described in detail above with reference to the description of the intelligent control system 800 of the seawater ice maker of FIG. 2, and therefore the repeated description thereof will be omitted.

[0076] FIG. 4 is an application scenario diagram of the intelligent control system of the seawater ice maker according to an embodiment of the present disclosure. As shown in FIG. 4, in this application scenario, first, a time sequence of ice layer thickness (for example, D1 shown in FIG. 4) and a time sequence of ambient temperature (for example, D2 shown in FIG. 4) are obtained, and then the time sequence of ice layer thickness and the time sequence of ambient temperature are input into a server (for example, S shown in FIG. 4) in which an intelligent control algorithm of the seawater ice maker is deployed, wherein the server can use the intelligent control algorithm of the seawater ice maker to process the time sequence of ice layer thickness and the time sequence of ambient temperature to obtain a result indicating whether to automatically start a defrosting cycle.

[0077] Based on the above embodiments, another exemplary embodiment of an electronic device is also provided in the embodiments of the present disclosure. In some possible embodiments, the electronic device in the embodiments of the present disclosure can include a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the intelligent control method of the seawater ice maker in the above embodiments when executing the program.

[0078] For example, taking the server 100 in FIG. 1 of the present disclosure as an example, the processor in the electronic device is the processor 110 in the server 100, and the memory in the electronic device is the memory 120 in the server 100.

[0079] Embodiments of the present disclosure also provide a computer-readable storage medium. FIG. 5 shows a schematic diagram of a computer-readable storage medium 1000 according to an embodiment of the present disclosure. As shown in FIG. 5, the computer-readable storage medium 1000 stores computer executable instructions 1001. When the computer executable instructions 1001 are executed by a processor, the intelligent control method of the seawater ice maker according to an embodiment of the present disclosure described with reference to the above drawings can be performed. The computer-readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory, etc. The non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, etc.

[0080] Embodiments of the present disclosure also provide a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the intelligent control method of the seawater ice maker according to an embodiment of the present disclosure.

[0081] Those skilled in the art will appreciate that the disclosure disclosed herein is susceptible to variations and modifications other than those specifically described. It is to be understood that the application includes all such variations and modifications. For example, the various devices or components described above can be implemented by hardware, or by software, firmware, or a combination of some or all of the above.

[0082] In addition, while the present disclosure has been made in reference to certain elements in the system according to embodiments of the present disclosure, any number of different elements can be used and run on the client and / or server. The elements are illustrative only, and different aspects of the system and method can use different elements.

[0083] Those of ordinary skill in the art will understand that all or portions of the methods described herein can be embodied in a program instruction for use by or to instruct a processing system or other programmable apparatus to perform a series of operations steps. Such program instructions can be stored in any storage media, such as readonly memory, magnetic disk or optical disk. Alternatively, some or all of the steps of the above-described embodiments can be implemented using one or more integrated circuits. Accordingly, the various modules / units in the above-described embodiments can be implemented in the form of hardware or in the form of software function modules. The present disclosure is not limited to any particular form of combination of hardware and software.

[0084] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.

[0085] The foregoing is a summary of the present disclosure, and is not to be considered as limiting its scope. While the present disclosure has been described with respect to certain illustrative embodiments thereof, those skilled in the art will note that many modifications can be made to the described embodiments without departing from the novel teachings and advantages of the present disclosure. It is, therefore, intended that all such modifications come within the scope of the present disclosure. It is to be understood that the above description is intended to be illustrative, and not restrictive. Many other embodiments will be apparent to those of skill in the art upon reviewing the above description. The scope of the present disclosure should, therefore, be determined not with reference to the above description, but instead with reference to the appended claims, along with their full scope of equivalents.

Claims

1. An intelligent control system for a seawater ice maker, characterized in that, The method comprises the following steps: a data acquisition module is used to obtain a time sequence of ice layer thickness and a time sequence of ambient temperature; a data time sequence arrangement module is used to arrange the time sequence of ice layer thickness and the time sequence of ambient temperature into an ice layer thickness time sequence input vector and an ambient temperature time sequence input vector respectively according to the time dimension; an ice layer thickness-ambient temperature time sequence cooperative correlation module is used to calculate a sample covariance matrix of the ice layer thickness time sequence input vector relative to the ambient temperature time sequence input vector to obtain an ice layer thickness-ambient temperature time sequence cooperative matrix; an ice layer thickness-ambient temperature time sequence correlation feature extraction module is used to extract features from the ice layer thickness-ambient temperature time sequence cooperative matrix by an ice layer thickness-ambient temperature time sequence correlation pattern feature extractor based on a deep neural network model to obtain an ice layer thickness-ambient temperature time sequence correlation feature map; an attention reinforcement module is used to perform attention reinforcement on the ice layer thickness-ambient temperature time sequence correlation feature map to obtain reinforced ice layer thickness-ambient temperature time sequence correlation features; and a defrosting cycle control module is used to determine whether to automatically start a defrosting cycle based on the reinforced ice layer thickness-ambient temperature time sequence correlation features. The deep neural network model is a convolutional neural network model.

2. The intelligent control system of the seawater ice maker according to claim 1, wherein, The ice thickness-environment temperature time series synergic correlation module is configured to calculate a sample covariance matrix of the ice thickness time series input vector relative to the environment temperature time series input vector by using a sample covariance formula to obtain the ice thickness-environment temperature time series synergic matrix, wherein the sample covariance formula is: wherein, is the ice thickness time series input vector, is the environment temperature time series input vector, is the ice thickness-environment temperature time series synergic matrix, denotes a transposition operation.

3. The intelligent control system of the seawater ice maker according to claim 2, wherein, The attention reinforcement module is used to input the ice layer thickness-ambient temperature time sequence correlation feature map into a feature enhancer based on an Asm-CBAM attention module to obtain a reinforced ice layer thickness-ambient temperature time sequence correlation feature map as the reinforced ice layer thickness-ambient temperature time sequence correlation features.

4. The intelligent control system of the seawater ice maker according to claim 3, wherein, The defrosting cycle control module is used to input the reinforced ice layer thickness-ambient temperature time sequence correlation feature map into a defrosting controller based on a classifier to obtain a control instruction, which is used to indicate whether to automatically start a defrosting cycle.

5. The intelligent control system of the seawater ice maker according to claim 4, wherein, The attention reinforcement module is configured to input the ice layer thickness-environment temperature time sequence correlation feature map into the feature reinforcement device based on the Asm-CBAM attention module, and process the ice layer thickness-environment temperature time sequence correlation feature map through a reinforcement formula to obtain a reinforced ice layer thickness-environment temperature time sequence correlation feature map. wherein, is the ice layer thickness-environment temperature time sequence correlation feature map, , and represent a mean-pooled feature map, a maximum-pooled feature map and a random-pooled feature map obtained by performing global mean-pooling, global maximum-pooling and global random-pooling on the ice layer thickness-environment temperature time sequence correlation feature map in a channel dimension, and represent multi-layer perceptrons with different numbers of channels, is the number of channels, is a Sigmoid activation function, is the reinforced ice layer thickness-environment temperature time sequence correlation feature map.

6. The intelligent control system of the seawater ice maker according to claim 5, wherein, The method further comprises a training module used to train an ice layer thickness-ambient temperature time sequence correlation pattern feature extractor based on a convolutional neural network model, a feature enhancer based on an Asm-CBAM attention module, and a defrosting controller based on a classifier.

7. The intelligent control system of the seawater ice maker according to claim 6, wherein, The training module comprises:

8. The intelligent control system of the seawater ice maker according to claim 7, wherein, a training data acquisition unit used to obtain training data, wherein the training data comprises a training time sequence of ice layer thickness and a training time sequence of ambient temperature; a training data time sequence arrangement unit used to arrange the training time sequence of ice layer thickness and the training time sequence of ambient temperature into a training ice layer thickness time sequence input vector and a training ambient temperature time sequence input vector respectively according to the time dimension; a training ice layer thickness-ambient temperature time sequence cooperative correlation unit used to calculate a sample covariance matrix of the training ice layer thickness time sequence input vector relative to the training ambient temperature time sequence input vector to obtain a training ice layer thickness-ambient temperature time sequence cooperative matrix; a training ice layer thickness-ambient temperature time sequence correlation feature extraction unit used to extract features from the training ice layer thickness-ambient temperature time sequence cooperative matrix by the ice layer thickness-ambient temperature time sequence correlation pattern feature extractor based on the convolutional neural network model to obtain a training ice layer thickness-ambient temperature time sequence correlation feature map; and a training ice layer thickness-ambient temperature time sequence correlation pattern feature extractor based on a convolutional neural network model. The training attention reinforcement unit is configured to input the training ice thickness-environment temperature time series correlation feature map into the feature reinforcement device based on the Asm-CBAM attention module to obtain a training reinforced ice thickness-environment temperature time series correlation feature map. The feature optimization unit is configured to optimize the training reinforced ice thickness-environment temperature time series correlation feature map to obtain an optimized training reinforced ice thickness-environment temperature time series correlation feature map. The classification training unit is configured to input the optimized training reinforced ice thickness-environment temperature time series correlation feature map into the defrosting controller based on the classifier to obtain a classification loss function value. The loss training unit is configured to train the ice thickness-environment temperature time series correlation pattern feature extractor based on the convolutional neural network model, the feature reinforcement device based on the Asm-CBAM attention module, and the defrosting controller based on the classifier based on the classification loss function value.

9. An intelligent control method of a seawater ice maker, characterized by, The method comprises: obtaining a time series of ice thickness and a time series of environment temperature; arranging the time series of ice thickness and the time series of environment temperature into an ice thickness time series input vector and an environment temperature time series input vector respectively according to the time dimension; calculating a sample covariance matrix of the ice thickness time series input vector relative to the environment temperature time series input vector to obtain an ice thickness-environment temperature time series correlation matrix; extracting features of the ice thickness-environment temperature time series correlation matrix by an ice thickness-environment temperature time series correlation pattern feature extractor based on a deep neural network model to obtain an ice thickness-environment temperature time series correlation feature map; performing attention reinforcement on the ice thickness-environment temperature time series correlation feature map to obtain reinforced ice thickness-environment temperature time series correlation features; and based on the reinforced ice thickness-environment temperature time series correlation features, determining whether to automatically start a defrosting cycle. ​

Images