Temperature Control Method and System for High-Temperature Anti-Corrosion Storage Tanks

Abstract

A temperature control method and system for a high-temperature anti-corrosion storage tank are disclosed. First, temperature values at multiple predetermined time points within a predetermined time period are collected by temperature sensors deployed inside the high-temperature anti-corrosion storage tank. Then, the temperature values at the multiple predetermined time points are arranged in a time dimension to form a temperature time-series input vector. Next, local time-series feature extraction is performed on the temperature time-series input vector to obtain a sequence of temperature local time-series feature vectors. Then, time-series global correlation coding is performed on the sequence of temperature local time-series feature vectors to obtain a temperature full-time-series semantic correlation feature. Finally, based on the temperature full-time-series semantic correlation feature, a cooling fan is turned on or a heater is turned on. In this way, the internal temperature of the storage tank can always be kept within a safe range, thereby ensuring the safety, stability, and reliability of the stored substances.

Description

Technical Field

[0001] This application relates to the field of temperature control, and more specifically, to a temperature control method and system for a high-temperature anti-corrosion storage tank. Background Art

[0002] High-temperature anti-corrosion storage tanks are usually used to store substances that need to remain stable in high-temperature environments, such as chemicals, petroleum products, etc., and are widely used in industries such as chemical engineering, petroleum, and refining. In a high-temperature environment, the internal temperature of the storage tank may exceed the safe storage temperature range of the substance, which can cause changes in the properties of the substance, such as intensified chemical reactions, increased evaporation of volatile substances, etc., thus triggering safety risks, such as explosion, leakage, or fire. Moreover, some substances may solidify, crystallize, or become viscous at low temperatures, resulting in a decrease in fluidity, making them difficult to use or process. In addition, the low-temperature environment may also cause damage or embrittlement of the materials of the storage tank, increasing the risk of leakage. Therefore, in order to ensure that the internal temperature of the storage tank is always within the safe range, it is necessary to monitor the temperature inside the high-temperature anti-corrosion storage tank to ensure the safety, stability, and reliability of the stored substances.

[0003] However, the traditional temperature monitoring scheme for high-temperature anti-corrosion storage tanks usually adopts manual operation and simple threshold judgment and adjustment methods. For example, by real-time monitoring the temperature data of the storage tank and judging whether the temperature exceeds the threshold, when the temperature exceeds the preset minimum or maximum temperature range, manually start the corresponding heater or cooling equipment to adjust the temperature value change of the storage tank. This method often relies on manual intervention, and the operation is not intelligent and automated enough to meet complex temperature control requirements.

[0004] In addition, the traditional temperature control scheme only relies on the temperature readings at a single time point for control, which cannot accurately reflect the temperature change trend inside the storage tank, resulting in hysteresis in temperature control, bringing economic losses and safety hazards. Moreover, the traditional method is not flexible enough in responding to temperature changes, unable to detect abnormal temperature conditions in a timely manner and adjust the working state of the heating or cooling equipment, increasing the risk of accidents.

[0005] Therefore, an optimized temperature control scheme for high-temperature anti-corrosion storage tanks is desired. Summary of the Invention

[0006] To solve the above technical problems, the present application is proposed. Embodiments of the present application provide a temperature control method and system for a high-temperature anti-corrosion storage tank. It can collect temperature value data at multiple predetermined time points through temperature sensors deployed inside the high-temperature anti-corrosion storage tank, and introduce data processing and analysis algorithms at the backend to perform time series analysis of the temperature values, so as to monitor and adjust the temperature of the storage tank in real time, so as to keep the internal temperature of the storage tank always within a safe range, thereby ensuring the safety, stability and reliability of the stored substances.

[0007] According to one aspect of the present application, there is provided a temperature control method for a high-temperature anti-corrosion storage tank, which includes:

[0008] Collecting temperature values at multiple predetermined time points within a predetermined time period through temperature sensors deployed inside the high-temperature anti-corrosion storage tank;

[0009] Arranging the temperature values at the multiple predetermined time points in a time dimension to form a temperature time series input vector;

[0010] Performing local time series feature extraction on the temperature time series input vector to obtain a sequence of temperature local time series feature vectors;

[0011] Performing time series global correlation encoding on the sequence of temperature local time series feature vectors to obtain a temperature full time series semantic correlation feature; and

[0012] Based on the temperature full time series semantic correlation feature, determining to turn on the cooling fan or turn on the heater.

[0013] According to another aspect of the present application, there is provided a temperature control system for a high-temperature anti-corrosion storage tank, which includes:

[0014] A temperature acquisition module for collecting temperature values at multiple predetermined time points within a predetermined time period through temperature sensors deployed inside the high-temperature anti-corrosion storage tank;

[0015] A vectorization module for arranging the temperature values at the multiple predetermined time points in a time dimension to form a temperature time series input vector;

[0016] A local time series feature extraction module for performing local time series feature extraction on the temperature time series input vector to obtain a sequence of temperature local time series feature vectors;

[0017] A time series global correlation encoding module for performing time series global correlation encoding on the sequence of temperature local time series feature vectors to obtain a temperature full time series semantic correlation feature; and

[0018] A temperature control module for determining to turn on the cooling fan or turn on the heater based on the temperature full time series semantic correlation feature.

[0019] Compared with the prior art, the temperature control method and system for a high-temperature anti-corrosion storage tank provided by the present application first collect temperature values at multiple predetermined time points within a predetermined time period through temperature sensors deployed inside the high-temperature anti-corrosion storage tank. Then, the temperature values at the multiple predetermined time points are arranged in a temperature time series input vector according to the time dimension. Next, local time series feature extraction is performed on the temperature time series input vector to obtain a sequence of temperature local time series feature vectors. Then, time series global correlation coding is performed on the sequence of temperature local time series feature vectors to obtain temperature full-time series semantic correlation features. Finally, based on the temperature full-time series semantic correlation features, a cooling fan is turned on or a heater is turned on. In this way, the internal temperature of the storage tank can always be kept within a safe range, thereby ensuring the safety, stability, and reliability of the stored substances. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] To more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required for the description of the embodiments. The following drawings are not deliberately drawn to scale in actual size, and the focus is on showing the gist of the present application.

[0021] Figure 1 It is a flowchart of the temperature control method for a high-temperature anti-corrosion storage tank according to an embodiment of the present application.

[0022] Figure 2 It is a schematic structural diagram of the temperature control method for a high-temperature anti-corrosion storage tank according to an embodiment of the present application.

[0023] Figure 3 It is a flowchart of sub-step S130 of the temperature control method for a high-temperature anti-corrosion storage tank according to an embodiment of the present application.

[0024] Figure 4 It is a flowchart of sub-step S132 of the temperature control method for a high-temperature anti-corrosion storage tank according to an embodiment of the present application.

[0025] Figure 5 It is a flowchart of sub-step S140 of the temperature control method for a high-temperature anti-corrosion storage tank according to an embodiment of the present application.

[0026] Figure 6 It is a flowchart of sub-step S150 of the temperature control method for a high-temperature anti-corrosion storage tank according to an embodiment of the present application.

[0027] Figure 7 It is a block diagram of the temperature control system for a high-temperature anti-corrosion storage tank according to an embodiment of the present application.

[0028] Figure 8It is an application scenario diagram of the temperature control method for a high-temperature anti-corrosion storage tank according to an embodiment of the present application. Detailed implementation manners

[0029] Next, the technical solutions in the embodiments of the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts also belong to the scope of protection of the present application.

[0030] As shown in the present application and the claims, unless the context clearly indicates otherwise, words such as "a", "an", "one" and / or "the" are not specifically singular and may also include plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of the clearly identified steps and elements, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0031] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or the server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0032] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the operations in front or below do not necessarily need to be executed precisely in sequence. On the contrary, as needed, various steps can be executed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or several steps can be removed from these processes.

[0033] Next, exemplary embodiments of the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the exemplary embodiments described herein.

[0034] In view of the above technical problems, the technical concept of the present application is to collect temperature value data at multiple predetermined time points through temperature sensors deployed in the high-temperature anti-corrosion storage tank, and introduce data processing and analysis algorithms at the backend to perform time-series analysis of the temperature values, so as to monitor and adjust the temperature of the storage tank in real time, so as to keep the internal temperature of the storage tank always within the safe range, thereby ensuring the safety, stability and reliability of the stored substances.

[0035] Figure 1 It is a flowchart of the temperature control method for a high-temperature anti-corrosion storage tank according to an embodiment of the present application. Figure 2The architecture diagram of the temperature control method for a high-temperature anti-corrosion storage tank according to an embodiment of the present application. As Figure 1 and Figure 2 shown, the temperature control method for a high-temperature anti-corrosion storage tank according to an embodiment of the present application includes the steps of: S110, collecting temperature values at multiple predetermined time points within a predetermined time period through a temperature sensor deployed in the high-temperature anti-corrosion storage tank; S120, arranging the temperature values at the multiple predetermined time points into a temperature time series input vector according to the time dimension; S130, performing local time series feature extraction on the temperature time series input vector to obtain a sequence of temperature local time series feature vectors; S140, performing time series global correlation coding on the sequence of temperature local time series feature vectors to obtain a temperature full time series semantic correlation feature; and S150, determining to turn on the cooling fan or the heater based on the temperature full time series semantic correlation feature.

[0036] Specifically, in the technical solution of the present application, first, temperature values at multiple predetermined time points collected by a temperature sensor deployed in the high-temperature anti-corrosion storage tank are obtained. Then, considering that the temperature values in the high-temperature anti-corrosion storage tank change continuously over time, and this time series change is relatively weak, it is difficult to analyze the time series change and trend of the temperature through traditional feature extraction methods. Therefore, in order to better capture the time series change feature information of the temperature values in the time dimension, in the technical solution of the present application, after arranging the temperature values at the multiple predetermined time points into a temperature time series input vector to integrate the time series distribution information of the temperature values in the time dimension, the temperature time series input vector is further vector-sliced to obtain a sequence of temperature local time series input vectors, so as to better extract the local time series detail change feature information of the temperature in different local time periods subsequently.

[0037] Then, in order to improve the capture ability of the time series subtle change features of the temperature values in the high-temperature anti-corrosion storage tank within a predetermined time period, in the technical solution of the present application, the sequence of temperature local time series input vectors is further subjected to upsampling processing based on linear interpolation through the upsampling module of the temperature time series feature extractor, so as to obtain a sequence of upsampled temperature local time series input vectors, thereby increasing the time series density and smoothness of the temperature data in the high-temperature anti-corrosion storage tank, and thus facilitating the subsequent better representation of the time series change features of the temperature values. Subsequently, the sequence of upsampled temperature local time series input vectors is passed through the one-dimensional convolutional layer of the temperature time series feature extractor for feature mining, so as to extract the local time series change feature information of the temperature values in each local time segment in the high-temperature anti-corrosion storage tank, thereby obtaining a sequence of temperature local time series feature vectors.

[0038] Furthermore, it is also considered that there is a correlation relationship based on the overall time series among the local time series detail change characteristics of the temperature values in the high-temperature anti-corrosion storage tank at each local time segment. Therefore, in order to capture the long-term dependence and context information in the temperature time series data to better understand the trend and pattern of temperature changes. In the technical solution of this application, the global mean of the transition matrix between every two adjacent temperature local time series feature vectors in the sequence of the temperature local time series feature vectors is further calculated to obtain a temperature time series semantic context feature vector composed of multiple temperature global transition eigenvalues, so as to represent the global time series semantic association feature information between every two adjacent segments in each local time segment regarding the local time series detail change characteristics of the temperature values in the high-temperature anti-corrosion storage tank. It should be understood that by calculating the transition matrix between each temperature local time series feature vector, the relationship and transition probability between adjacent feature vectors can be analyzed. The transition matrix describes the transition situation from one feature vector to another and reflects the state transition and evolution process in the temperature time series data.

[0039] Correspondingly, as Figure 3 shown, performing local time series feature extraction on the temperature time series input vector to obtain a sequence of temperature local time series feature vectors includes: S131, performing vector segmentation on the temperature time series input vector to obtain a sequence of temperature local time series input vectors; and, S132, passing the sequence of temperature local time series input vectors through a temperature time series feature extractor including an upsampling module and a one-dimensional convolutional layer to obtain the sequence of temperature local time series feature vectors. It should be understood that the function of step S131 is to perform vector segmentation on the temperature time series input vector to obtain a sequence of temperature local time series input vectors. This step segments the input temperature time series data according to a certain window size to obtain a series of local time series input vectors. The purpose of doing this is to capture the local time series information of the temperature data so that the model can better understand the change trend of the temperature data. The function of step S132 is to pass the sequence of temperature local time series input vectors through a temperature time series feature extractor including an upsampling module and a one-dimensional convolutional layer to obtain the sequence of temperature local time series feature vectors. This step uses the upsampling module to upsample the input local time series input vectors to expand the length of the vectors, and then, performs a convolution operation on the upsampled vectors through the one-dimensional convolutional layer to extract the time series features of the temperature data. The purpose of doing this is to extract and represent the important time series features in the temperature data so that the subsequent model can better utilize these features for the analysis and prediction of temperature data.

[0040] More specifically, in step S132, as Figure 4As shown, passing the sequence of the temperature local temporal input vectors through a temperature temporal feature extractor including an upsampling module and a one-dimensional convolutional layer to obtain the sequence of the temperature local temporal feature vectors includes: S1321, passing the sequence of the temperature local temporal input vectors through the upsampling module for upsampling processing based on linear interpolation to obtain a sequence of upsampled temperature local temporal input vectors; and, S1322, passing the sequence of the upsampled temperature local temporal input vectors through the one-dimensional convolutional layer to obtain the sequence of the temperature local temporal feature vectors. It should be understood that the role of step S1321 is to pass the sequence of the temperature local temporal input vectors through the upsampling module for upsampling processing based on linear interpolation to obtain a sequence of upsampled temperature local temporal input vectors. Upsampling is a signal processing technique that expands the length of a signal by increasing the sampling rate. In this step, the length of the local temporal input vectors is expanded by linear interpolation, so that each time step of the temperature features obtains more sampling points, thereby improving the ability to express the details of the features. The role of step S1322 is to pass the sequence of the upsampled temperature local temporal input vectors through the one-dimensional convolutional layer to obtain the sequence of the temperature local temporal feature vectors. The one-dimensional convolutional layer is a commonly used neural network layer that can perform convolutional operations on the input vectors by means of a sliding window, thereby extracting local features. In this step, the one-dimensional convolutional layer performs convolutional operations on the upsampled temperature local temporal input vectors to extract the temporal features of the temperature data. The purpose of doing this is to extract and represent the important temporal features in the temperature data so that the subsequent model can better utilize these features for the analysis and prediction of temperature data.

[0041] More specifically, in step S140, temporal global correlation encoding is performed on the sequence of the temperature local temporal feature vectors to obtain temperature full-temporal semantic correlation features, including: calculating the global mean of the transition matrices between every two adjacent temperature local temporal feature vectors in the sequence of the temperature local temporal feature vectors to obtain a temperature temporal semantic context feature vector composed of multiple temperature global transition eigenvalues as the temperature full-temporal semantic correlation features. It should be understood that the function of this step is to calculate the global mean of the transition matrices between every two adjacent feature vectors in the sequence of temperature local temporal feature vectors to obtain a temperature temporal semantic context feature vector composed of multiple temperature global transition eigenvalues. This feature vector can be regarded as the semantic correlation feature of temperature data in terms of time series. By calculating the global mean of the transition matrices, the average transition features between adjacent feature vectors can be obtained. The transition matrix reflects the relationship and transition pattern between adjacent feature vectors and can capture the semantic context information of temperature data in terms of time series. By calculating the mean of multiple transition features, a comprehensive temperature temporal semantic context feature vector can be obtained, which contains the information of multiple transition features. This temperature temporal semantic context feature vector can be used as part of the temperature full-temporal semantic correlation features to more comprehensively represent the temporal correlation and semantic information of temperature data. Such a feature vector can provide a richer and more accurate feature representation in subsequent analysis and prediction tasks, thereby improving the performance and effect of the model.

[0042] More specifically, as Figure 5As shown, calculating the global mean of the transition matrices between every two adjacent temperature local temporal feature vectors in the sequence of the temperature local temporal feature vectors to obtain a temperature temporal semantic context feature vector composed of multiple temperature global transition eigenvalues as the temperature full-temporal semantic association feature includes: S141, calculating the transition matrices between every two adjacent temperature local temporal feature vectors in the sequence of the temperature local temporal feature vectors to obtain multiple transition matrices; S142, performing a rank permutation distribution soft matching of the feature scales of the respective transition matrices as an imitation mask to obtain multiple optimized transition matrices; and, S143, respectively calculating the global means of the multiple optimized transition matrices to obtain the temperature temporal semantic context feature vector composed of multiple temperature global transition eigenvalues. It should be understood that the function of step S141 is to calculate the transition matrices between every two adjacent feature vectors in the temperature local temporal feature vector sequence to obtain multiple transition matrices, and these transition matrices reflect the relationships and transition patterns between adjacent feature vectors. The function of step S142 is to perform a rank permutation distribution soft matching of the feature scales of each transition matrix as an imitation mask to obtain multiple optimized transition matrices. The imitation mask is a technique for calculating similarity. By comparing the transition matrix with a mask matrix, the similarity distribution of the feature scales can be obtained. By performing a rank permutation distribution soft matching of the transition matrix, the feature scales of the transition matrix can be optimized, so that the transition pattern can more accurately reflect the temporal correlation features of the temperature data. The function of step S143 is to respectively calculate the global means of the multiple optimized transition matrices to obtain the temperature temporal semantic context feature vector composed of multiple temperature global transition eigenvalues. By calculating the global mean, the average eigenvalue of each optimized transition matrix can be obtained, and these eigenvalues reflect the semantic relevance of the temperature data in time series. Combining multiple transition eigenvalues into a feature vector can obtain the temperature temporal semantic context feature vector, which is used to represent the full-temporal semantic association feature of the temperature data. Such a feature vector can provide a richer and more accurate feature representation in subsequent analysis and prediction tasks, thereby improving the performance and effect of the model.

[0043] Subsequently, passing the temperature temporal semantic context feature vector through a classifier to obtain a classification result, where the classification result is used to indicate turning on the cooling fan or turning on the heater. That is, using the temperature temporal change feature in the high-temperature anti-corrosion storage tank based on the global temporal semantic association feature information of long-distance dependence for classification processing, so as to turn on the cooling fan or turn on the heater based on the actual temperature change pattern in the high-temperature anti-corrosion storage tank. In this way, the temperature of the storage tank can be monitored and adjusted in real time to keep the internal temperature of the storage tank always within a safe range.

[0044] Correspondingly, in step S150, based on the full-time-series semantic association features of the temperature, determining to turn on the cooling fan or the heater includes: passing the temperature time-series semantic context feature vector through a classifier to obtain a classification result, where the classification result is used to represent turning on the cooling fan or the heater.

[0045] More specifically, as Figure 6 shown, passing the temperature time-series semantic context feature vector through a classifier to obtain a classification result, where the classification result is used to represent turning on the cooling fan or the heater, includes: S151, using the fully connected layer of the classifier to perform fully connected encoding on the temperature time-series semantic context feature vector to obtain an encoded classification feature vector; and, S152, inputting the encoded classification feature vector into the Softmax classification function of the classifier to obtain the classification result.

[0046] That is, in the technical solution of the present disclosure, the labels of the classifier include turning on the cooling fan (the first label) and turning on the heater (the second label). Among them, the classifier determines which classification label the temperature time-series semantic context feature vector belongs to through the softmax function. It should be noted that the first label p1 and the second label p2 here do not include artificially set concepts. In fact, during the training process, the computer model does not have the concept of "turning on the cooling fan or the heater". It only has two classification labels and outputs the probabilities of the output features under these two classification labels, that is, the sum of p1 and p2 is one. Therefore, the classification result of turning on the cooling fan or the heater is actually transformed into a binary classification probability distribution that conforms to natural laws through classification labels. Essentially, it uses the physical meaning of the natural probability distribution of the labels, rather than the language text meaning of "turning on the cooling fan or the heater".

[0047] It should be understood that the role of the classifier is to use the given categories and known training data to learn classification rules and classifiers, and then classify (or predict) unknown data. Logistic regression, SVM, etc. are commonly used to solve binary classification problems. For multi-class classification problems, logistic regression or SVM can also be used. However, it requires multiple binary classifications to form multi-class classification, which is prone to errors and inefficient. Commonly used multi-class classification methods include the Softmax classification function.

[0048] It is worth mentioning that fully connected encoding refers to linearly transforming and non-linearly activating an input vector through a fully connected layer to obtain an encoded feature vector. The fully connected layer is a common layer in neural networks, where each input feature is fully connected to the output features, that is, each input feature is connected to all output features of this layer. In fully connected encoding, the input feature vector undergoes a linear transformation by a weight matrix and a bias vector, and then a non-linear mapping through an activation function, finally obtaining the encoded feature vector. This process can be regarded as a process of feature extraction and representation for the input feature vector. By learning the weight and bias parameters suitable for the task, the input features can be mapped into a higher-dimensional feature space. Fully connected encoding can help the model learn the complex relationships and abstract representations between input features, thereby extracting more discriminative features. It can be used to convert the input feature vector into an encoded feature vector, providing a more informative and expressive feature representation for subsequent classification, regression, or other tasks. In step S151, fully connected encoding is used to encode the temperature time-series semantic context feature vector to obtain an encoded classification feature vector. Such an encoded feature vector can better represent the time-series semantic association features of temperature data and be passed as input to the classifier for classification, representing the classification results of turning on the cooling fan or turning on the heater.

[0049] Specifically, in the technical solution of this application, the sequence of the temperature local time-series input vectors is passed through a temperature time-series feature extractor including an upsampling module and a one-dimensional convolutional layer to obtain a sequence of temperature local time-series feature vectors. Each temperature local time-series feature vector in the sequence of temperature local time-series feature vectors can express the time-series association features of temperature values in the local time domain. Thus, when calculating the transition matrix between every two adjacent temperature local time-series feature vectors in the sequence of temperature local time-series feature vectors, the transition matrix expresses the time-domain transition features between local time domains. Therefore, if the time-series association features of temperature values in each local time domain are regarded as foreground object features, while extracting the inter-domain transition features between local time domains, background distribution noise will also be introduced. Moreover, when performing high-rank distribution representation between vectors and matrices during the calculation of the transition matrix, due to the temporal high-dimensional feature's temporal space heterogeneous distribution between two adjacent temperature local time-series feature vectors, it will cause a temporal probability density mapping error of the transition matrix relative to the sequence of temperature local time-series feature vectors, affecting the accuracy of the classification results obtained by the classifier for the temperature time-series semantic context feature vector composed of multiple state transition eigenvalues obtained from the global mean of the calculated transition matrix.

[0050] Based on this, the applicant of this application performs rank arrangement distribution soft matching with feature scale as an imitation mask for each of the transition matrices, for example, denoted as M.

[0051] Accordingly, in a specific example, performing soft matching of rank arrangement distribution with the feature scale of each of the transfer matrices as an imitation mask to obtain a plurality of optimized transfer matrices, including: performing soft matching of rank arrangement distribution with the feature scale of each of the transfer matrices as an imitation mask according to the following optimization formula to obtain the plurality of optimized transfer matrices; wherein, the optimization formula is:

[0052]

[0053] wherein, M is each of the transfer matrices, and m i,j is the eigenvalue at the (i, j) position of each of the transfer matrices, S is the scale of each of the transfer matrices, that is, width multiplied by height, represents the square of the Frobenius norm of each of the transfer matrices, ‖M‖ 2 represents the two-norm of each of the transfer matrices, that is, the spectral norm λ max is the maximum eigenvalue of M T M, and α is a weighted hyperparameter, exp(·) represents the exponential operation of a numerical value, and the exponential operation of the numerical value represents calculating the value of the natural exponential function with the numerical value as the power, and m i,j is the eigenvalue at the (i, j) position of each of the optimized transfer matrices.

[0054] Here, when mapping the high-dimensional features to be classified and regressed into the probability density space, the soft matching of rank arrangement distribution with the feature scale as an imitation mask can use the feature scale as an imitation mask for mapping to focus on the foreground object features and ignore the background distribution noise, and perform soft matching of the pyramid rank arrangement distribution through different norms of the transfer matrix M, thereby effectively capturing the correlation between the central region and the tail region of the probability density distribution, avoiding the probability density mapping deviation caused by the temporal and spatial heterogeneous distribution of the high-dimensional features of the transfer matrix M, and thus improving the accuracy of the classification result obtained by the classifier for the temperature time-series semantic context feature vector composed of multiple state transition eigenvalues calculated from the global mean of the transfer matrix. In this way, it is possible to adjust the temperature data of the storage tank based on the real-time temperature change in the high-temperature anti-corrosion storage tank to keep the internal temperature of the storage tank always within the safe range, thereby ensuring the safety, stability, and reliability of the stored substances.

[0055] In summary, the temperature control method for a high-temperature anti-corrosion storage tank according to the embodiments of the present application is elucidated, which can keep the internal temperature of the storage tank always within the safe range, thereby ensuring the safety, stability, and reliability of the stored substances.

[0056] Figure 7 is a block diagram of a temperature control system 100 for a high-temperature anti-corrosion storage tank according to an embodiment of the present application. AsFigure 7 As shown in Figure 7 , the temperature control system 100 for a high-temperature anti-corrosion storage tank according to an embodiment of the present application includes: a temperature acquisition module 110, configured to acquire temperature values at a plurality of predetermined time points within a predetermined time period through a temperature sensor deployed in the high-temperature anti-corrosion storage tank; a vectorization module 120, configured to arrange the temperature values at the plurality of predetermined time points into a temperature time series input vector according to the time dimension; a local time series feature extraction module 130, configured to perform local time series feature extraction on the temperature time series input vector to obtain a sequence of temperature local time series feature vectors; a time series global correlation encoding module 140, configured to perform time series global correlation encoding on the sequence of temperature local time series feature vectors to obtain a temperature full time series semantic correlation feature; and a temperature control module 150, configured to determine to turn on a cooling fan or a heater based on the temperature full time series semantic correlation feature.

[0057] In one example, in the temperature control system 100 for a high-temperature anti-corrosion storage tank described above, the local time series feature extraction module 130 includes: a vector segmentation unit, configured to perform vector segmentation on the temperature time series input vector to obtain a sequence of temperature local time series input vectors; and a temperature time series feature extraction unit, configured to pass the sequence of temperature local time series input vectors through a temperature time series feature extractor including an upsampling module and a one-dimensional convolutional layer to obtain the sequence of temperature local time series feature vectors.

[0058] Here, those skilled in the art can understand that the specific functions and operations of each module in the temperature control system 100 for a high-temperature anti-corrosion storage tank described above have been introduced in detail in the description of the temperature control method for a high-temperature anti-corrosion storage tank with reference to Figures 1 to 6 above, and therefore, the repeated description thereof will be omitted.

[0059] As described above, the temperature control system 100 for a high-temperature anti-corrosion storage tank according to an embodiment of the present application can be implemented in various wireless terminals, such as a server having a temperature control algorithm for a high-temperature anti-corrosion storage tank. In one example, the temperature control system 100 for a high-temperature anti-corrosion storage tank according to an embodiment of the present application can be integrated into a wireless terminal as a software module and / or a hardware module. For example, the temperature control system 100 for a high-temperature anti-corrosion storage tank can be a software module in the operating system of the wireless terminal, or can be an application program developed for the wireless terminal; of course, the temperature control system 100 for a high-temperature anti-corrosion storage tank can also be one of the numerous hardware modules of the wireless terminal.

[0060] Alternatively, in another example, the temperature control system 100 for the high-temperature anti-corrosion storage tank and the wireless terminal can also be separate devices, and the temperature control system 100 for the high-temperature anti-corrosion storage tank can be connected to the wireless terminal through a wired and / or wireless network and transmit interactive information in accordance with a predefined data format.

[0061] Figure 8 FIG. is an application scenario diagram of the temperature control method for a high-temperature anti-corrosion storage tank according to an embodiment of the present application. As Figure 8 shown, in this application scenario, first, temperature sensors (e.g., Figure 8 L shown therein) deployed in a high-temperature anti-corrosion storage tank (e.g., Figure 8 N shown therein) collect temperature values at multiple predetermined time points within a predetermined time period (e.g., Figure 8 D shown therein), and then, the temperature values at the multiple predetermined time points are input into a server (e.g., Figure 8 S shown therein) where a temperature control algorithm for the high-temperature anti-corrosion storage tank is deployed. Among them, the server can use the temperature control algorithm for the high-temperature anti-corrosion storage tank to process the temperature values at the multiple predetermined time points to obtain a classification result indicating to turn on the cooling fan or turn on the heater.

[0062] In addition, those skilled in the art can understand that various aspects of the present application can be illustrated and described by several patentable types or situations, including any new and useful processes, machines, products, or combinations of substances, or any new and useful improvements thereto. Accordingly, various aspects of the present application can be executed entirely by hardware, entirely by software (including firmware, resident software, microcode, etc.), or by a combination of hardware and software. The above hardware or software can all be referred to as "data blocks", "modules", "engines", "units", "components", or "systems". In addition, various aspects of the present application may be embodied as a computer product located in one or more computer-readable media, and the product includes computer-readable program codes.

[0063] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art to which the present invention belongs. It should also be understood that terms such as those defined in a general dictionary should be interpreted as having a meaning consistent with their meaning in the context of the relevant art, and should not be interpreted in an idealized or overly formal sense unless explicitly defined as such herein.

[0064] The foregoing is a description of the present invention and should not be construed as limiting thereof. Although several exemplary embodiments of the present invention have been described, those skilled in the art will readily appreciate that many modifications can be made to the exemplary embodiments without departing from the novel teachings and advantages of the present invention. Accordingly, all such modifications are intended to be included within the scope of the present invention as defined by the claims. It should be understood that the foregoing is a description of the present invention and should not be considered limited to the specific embodiments disclosed, and modifications to the disclosed embodiments as well as other embodiments are intended to be included within the scope of the appended claims. The present invention is defined by the claims and their equivalents.

Claims

1. A temperature control method for a high-temperature anti-corrosion storage tank, characterized in that, it includes: collecting temperature values at multiple predetermined time points within a predetermined time period through a temperature sensor deployed in the high-temperature anti-corrosion storage tank; arranging the temperature values at the multiple predetermined time points in a time dimension to form a temperature time series input vector; performing local time series feature extraction on the temperature time series input vector to obtain a sequence of temperature local time series feature vectors; performing time series global correlation coding on the sequence of temperature local time series feature vectors to obtain a temperature full time series semantic correlation feature; and based on the temperature full time series semantic correlation feature, determining to turn on a cooling fan or a heater; performing time series global correlation coding on the sequence of temperature local time series feature vectors to obtain a temperature full time series semantic correlation feature, including: calculating a transition matrix between every two adjacent temperature local time series feature vectors in the sequence of temperature local time series feature vectors to obtain a plurality of transition matrices; performing a rank arrangement distribution soft matching of the feature scale as an imitation mask on each of the transition matrices to obtain a plurality of optimized transition matrices, wherein, by comparing the transition matrix with a mask matrix, a similarity distribution of the feature scale is obtained, and by performing a rank arrangement distribution soft matching of the imitation mask on the transition matrix, the feature scale of the transition matrix is optimized. When mapping the high-dimensional features to be classified and regressed into the probability density space, the feature scale is used as the imitation mask for mapping to focus on the foreground object features and ignore the background distribution noise, and through the pyramid rank arrangement distribution soft matching of different norms of the transition matrix, the correlation between the central region and the tail region of the probability density distribution is captured, avoiding the probability density mapping deviation caused by the temporal space heterogeneous distribution of the high-dimensional features of the transition matrix; and respectively calculating the global means of the plurality of optimized transition matrices to obtain the temperature time series semantic context feature vector composed of a plurality of temperature global transition eigenvalues as the temperature full time series semantic correlation feature.

2. The temperature control method for a high-temperature anti-corrosion storage tank according to claim 1, characterized in that, performing local time series feature extraction on the temperature time series input vector to obtain a sequence of temperature local time series feature vectors, including: performing vector segmentation on the temperature time series input vector to obtain a sequence of temperature local time series input vectors; and passing the sequence of temperature local time series input vectors through a temperature time series feature extractor including an upsampling module and a one-dimensional convolutional layer to obtain the sequence of temperature local time series feature vectors.

3. The temperature control method for a high-temperature anti-corrosion storage tank according to claim 2, characterized in that, passing the sequence of temperature local time series input vectors through a temperature time series feature extractor including an upsampling module and a one-dimensional convolutional layer to obtain the sequence of temperature local time series feature vectors, including: performing upsampling processing based on linear interpolation on the sequence of temperature local time series input vectors through the upsampling module to obtain a sequence of upsampled temperature local time series input vectors; and Pass the sequence of the upsampled temperature local temporal input vectors through the one-dimensional convolutional layer to obtain a sequence of the temperature local temporal feature vectors.

4. The temperature control method for a high-temperature anti-corrosion storage tank according to claim 3, wherein, performing feature-scale soft matching of rank permutation distribution with an imitation mask on each of the transition matrices to obtain a plurality of optimized transition matrices, including: performing feature-scale soft matching of rank permutation distribution with an imitation mask on each of the transition matrices according to the following optimization formula to obtain the plurality of optimized transition matrices; wherein, the optimization formula is: where M is each of the transition matrices, and m i,j is the eigenvalue at the (i, j)-th position of each of the transition matrices, S is the scale of each of the transition matrices, represents the square of the Frobenius norm of each of the transition matrices, ||M|| 2 represents the 2-norm of each of the transition matrices, and α is a weighted hyperparameter, exp(·) represents the exponential operation of a numerical value, and the exponential operation of the numerical value represents calculating the value of the natural exponential function with the numerical value as the power, and m′ i,j is the eigenvalue at the (i, j)-th position of each of the optimized transition matrices.

5. The temperature control method for a high-temperature anti-corrosion storage tank according to claim 4, wherein, based on the temperature full-temporal semantic association feature, determining to turn on the cooling fan or turn on the heater, including: passing the temperature temporal semantic context feature vector through a classifier to obtain a classification result, where the classification result is used to represent turning on the cooling fan or turning on the heater.

6. The temperature control method for a high-temperature anti-corrosion storage tank according to claim 5, wherein, passing the temperature temporal semantic context feature vector through a classifier to obtain a classification result, where the classification result is used to represent turning on the cooling fan or turning on the heater, including: performing fully connected encoding on the temperature temporal semantic context feature vector using the fully connected layer of the classifier to obtain an encoded classification feature vector; and inputting the encoded classification feature vector into the Softmax classification function of the classifier to obtain the classification result.

7. A temperature control system for a high-temperature anti-corrosion storage tank, wherein, comprising: a temperature acquisition module, configured to acquire temperature values at a plurality of predetermined time points within a predetermined time period through a temperature sensor deployed inside the high-temperature anti-corrosion storage tank; a vectorization module, configured to arrange the temperature values at the plurality of predetermined time points in a time dimension to form a temperature temporal input vector; a local temporal feature extraction module, configured to perform local temporal feature extraction on the temperature temporal input vector to obtain a sequence of temperature local temporal feature vectors; a temporal global association encoding module, configured to perform temporal global association encoding on the sequence of the temperature local temporal feature vectors to obtain a temperature full-temporal semantic association feature; and a temperature control module, configured to determine to turn on the cooling fan or turn on the heater based on the temperature full-temporal semantic association feature; the temporal global association encoding module is configured to calculate transition matrices between every two adjacent temperature local temporal feature vectors in the sequence of the temperature local temporal feature vectors to obtain a plurality of transition matrices; Perform eigen-scale rank arrangement distribution soft matching on each of the transfer matrices as an imitation mask to obtain multiple optimized transfer matrices. Among them, by comparing the transfer matrix with a mask matrix, the similarity distribution of the eigen-scale is obtained. By performing eigen-scale rank arrangement distribution soft matching on the transfer matrix, the eigen-scale of the transfer matrix is optimized. When mapping the high-dimensional features to be classified and regressed into the probability density space, the eigen-scale is used as an imitation mask for mapping to focus on the foreground object features and ignore the background distribution noise. And through the distribution soft matching of the pyramid rank arrangement distribution performed by different norms of the transfer matrix, the correlation between the central region and the tail region of the probability density distribution is captured, avoiding the probability density mapping deviation caused by the temporal-spatial heterogeneous distribution of the high-dimensional features of the transfer matrix; And Calculate the global mean of multiple optimized transfer matrices respectively to obtain the temperature temporal semantic context feature vector composed of multiple temperature global transfer eigenvalues as the temperature full-temporal semantic association feature.

8. The temperature control system for a high-temperature anti-corrosion storage tank according to claim 7, characterized in that the local temporal feature extraction module includes: a vector segmentation unit for segmenting the temperature temporal input vector to obtain a sequence of temperature local temporal input vectors; and a temperature temporal feature extraction unit for passing the sequence of temperature local temporal input vectors through a temperature temporal feature extractor including an upsampling module and a one-dimensional convolutional layer to obtain a sequence of temperature local temporal feature vectors.

Images