Intelligent cryogenic storage tank and method

Abstract

The application discloses an intelligent low-temperature storage tank and a method thereof. The pressure value and the overall capacity value inside the low-temperature storage tank are measured in real time by a pressure sensor and a capacity sensor respectively, and a data processing and analysis algorithm is introduced at the rear end to perform time-series interactive correlation analysis on the capacity value and the pressure value, so as to control the exhaust or inflation of the storage tank body, and to adaptively adjust the pressure in the storage tank. In this way, the automatic adjustment of the pressure in the storage tank can be realized, and the automation degree and the safety of the operation of the storage tank are further improved, the manual intervention and the maintenance cost are reduced, and the operation risk is reduced, so that the storage efficiency and the safety are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of low-temperature liquid storage, and more particularly, to an intelligent low-temperature storage tank and a method thereof. BACKGROUND

[0002] A low-temperature storage tank is a device used for storing low-temperature liquid substances, such as liquid natural gas (LNG), liquid oxygen, liquid nitrogen, etc., which need to be kept in a liquid state at extremely low temperatures and need to be stored and transported within a specific pressure range. In a low-temperature storage tank, since the stored substance is usually a liquid gas or liquid, excessive high or low pressure in the tank can affect the safety and stability of the stored substance, so the pressure in the tank needs to be adjusted to keep it within a specific range to ensure safety and stability.

[0003] However, conventional low-temperature storage tanks usually require manual intervention for pressure adjustment. The operator needs to monitor the pressure changes in the tank and perform venting or gassing operations as needed. This method requires manual judgment and operation, which is easily affected by human factors and has certain operation risks and uncertainties. Moreover, due to the need for manual intervention, the pressure adjustment reaction time of conventional low-temperature storage tanks is relatively long. When the pressure in the tank exceeds the safe range, manual operation is needed to adjust it, which may cause delays and untimely responses, increasing the risk of accidents.

[0004] Therefore, an optimized intelligent low-temperature storage tank is desired. SUMMARY

[0005] To solve the above technical problems, the present application is proposed. The embodiments of the present application provide an intelligent low-temperature storage tank and a method thereof, which determines the pressure value and the overall capacity value inside the low-temperature storage tank in real time through a pressure sensor and a capacity sensor, respectively, and introduces data processing and analysis algorithms in the backend to perform time-series interactive correlation analysis of the capacity value and the pressure value, so as to control the venting or gassing of the tank body to adaptively adjust the pressure in the tank. In this way, automatic adjustment of the pressure in the tank can be achieved, further improving the automation degree and safety of tank operation, reducing manual intervention and maintenance costs, and reducing the risk of operation, thereby improving storage efficiency and safety.

[0006] According to an aspect of the present application, an intelligent low-temperature storage tank is provided, comprising:

[0007] a tank body for storing low-temperature substances;

[0008] a pressure sensor arranged inside the tank body, the pressure sensor being configured to detect the pressure value inside the tank body;

[0009] A capacity sensor is provided on the outside of the storage tank body, and is used to detect the capacity value of the storage tank body;

[0010] exhaust system;

[0011] A controller is communicatively connected to the pressure sensor, the capacity sensor and the exhaust device, and is used to control the exhaust device to exhaust or inflate the storage tank body.

[0012] According to another aspect of the present application, a method for using an intelligent cryogenic storage tank is provided, comprising:

[0013] Acquiring capacity values ​​and pressure values ​​collected by the pressure sensor and the capacity sensor at a plurality of predetermined time points within a predetermined time period;

[0014] Arranging the capacity values ​​and pressure values ​​at the plurality of predetermined time points into a capacity value time series input vector and a pressure value time series input vector according to a time dimension;

[0015] performing a time series feature interaction correlation analysis on the capacity value time series input vector and the pressure value time series input vector to obtain a capacity-pressure time series interaction feature;

[0016] Based on the capacity-pressure time series interaction characteristics, it is determined whether to exhaust or inflate the storage tank body.

[0017] Compared with the prior art, the present application provides an intelligent cryogenic storage tank and method thereof, which respectively measure the pressure value and overall capacity value inside the cryogenic storage tank in real time through a pressure sensor and a capacity sensor, and introduce data processing and analysis algorithms at the back end to perform time-series interactive correlation analysis of the capacity value and the pressure value, so as to control the exhaust or inflation of the storage tank body to adaptively adjust the pressure inside the storage tank. In this way, automatic adjustment of the pressure inside the storage tank can be achieved, further improving the degree of automation and safety of the tank operation, reducing manual intervention and maintenance costs, and reducing operational risks, thereby improving storage efficiency and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above and other purposes, features, and advantages of the present application will become more apparent through a more detailed description of the embodiments of the present application in conjunction with the accompanying drawings. The accompanying drawings are intended to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0019] Figure 1 is a block diagram of an intelligent cryogenic storage tank according to an embodiment of the present application;

[0020] Figure 2 System architecture diagram of the intelligent low-temperature storage tank according to the embodiment of the present application;

[0021] Figure 3 Block diagram of the controller in the intelligent low-temperature storage tank according to the embodiment of the present application;

[0022] Figure 4 Block diagram of the parameter time sequence characteristic interaction correlation analysis module in the intelligent low-temperature storage tank according to the embodiment of the present application;

[0023] Figure 5 Block diagram of the storage tank body exhaust control module in the intelligent low-temperature storage tank according to the embodiment of the present application;

[0024] Figure 6 Flow chart of the use method of the intelligent low-temperature storage tank according to the embodiment of the present application. DETAILED DESCRIPTION

[0025] Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of embodiments of the present application, and thus are not to be used to limit the present application, and it is to be understood that the present application is not limited to the described example embodiments.

[0026] As shown in the present application and claims, unless the context clearly indicates otherwise, the words "one", "an", "a", and / or "the" do not mean to specify a single number, but also include a plurality. Generally, the terms "comprising" and "including" only indicate including the steps and elements explicitly identified, and these steps and elements do not constitute an exclusive list, and the method or device can also include other steps or elements.

[0027] 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 server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0028] Flow charts 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 preceding 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 be added to these processes, or a step or steps of operation can be removed from these processes.

[0029] Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of embodiments of the present application, and thus are not to be used to limit the present application, and it is to be understood that the present application is not limited to the described example embodiments.

[0030] Traditional cryogenic storage tanks typically require manual intervention for pressure regulation. Operators must monitor pressure changes within the tank and vent or inflate the tank as needed. This approach requires manual judgment and operation, is susceptible to human influence, and carries certain operational risks and uncertainties. Furthermore, due to the need for manual intervention, the pressure regulation response time of traditional cryogenic storage tanks is long. When the pressure within the tank exceeds the safe range, manual intervention is required before adjustment can be made, which can lead to delayed and untimely responses, increasing the risk of accidents. Therefore, an optimized intelligent cryogenic storage tank is desired.

[0031] In the technical solution of this application, an intelligent low-temperature storage tank is proposed. Figure 1 FIG is a block diagram of an intelligent cryogenic storage tank according to an embodiment of the present application. Figure 1 As shown, the intelligent cryogenic storage tank 300 according to an embodiment of the present application includes: a storage tank body 310, for storing cryogenic substances; a pressure sensor 320 arranged inside the storage tank body, and the pressure sensor is used to detect the pressure value inside the storage tank body; a capacity sensor 330 arranged outside the storage tank body, and the capacity sensor is used to detect the capacity value of the storage tank body; an exhaust device 340; and a controller 350, which can be communicatively connected to the pressure sensor, the capacity sensor and the exhaust device, and the controller is used to control the exhaust device to exhaust or inflate the storage tank body.

[0032] In particular, the storage tank 310 is used to store cryogenic substances. The selection of the material and structure of the storage tank should be evaluated and selected based on the properties, storage conditions and requirements of the cryogenic substances.

[0033] In particular, the pressure sensor 320 disposed inside the storage tank body is used to detect the pressure value inside the storage tank body. A pressure sensor is a device for measuring pressure, which converts pressure into an electrical signal output for monitoring, control or recording.

[0034] In particular, the capacity sensor 330 disposed outside the storage tank is used to detect the capacity value of the storage tank. The capacity sensor is a device for measuring capacity, which converts the capacity into an electrical signal output for monitoring, control or recording.

[0035] In particular, the exhaust device 340. An exhaust device is a device or system for discharging or releasing gas. It is typically used to control and manage the flow of gas.

[0036] In particular, the controller 350 is communicatively connected to the pressure sensor, the capacity sensor and the exhaust device, and is used to control the exhaust device to exhaust or inflate the storage tank. In particular, in a specific example of the present application, Figure 2 and Figure 3 As shown, the controller 350 includes: a data acquisition module 351, which is used to obtain the capacity values ​​and pressure values ​​at multiple predetermined time points within a predetermined time period collected by the pressure sensor and the capacity sensor; a data parameter time series arrangement module 352, which is used to arrange the capacity values ​​and pressure values ​​at the multiple predetermined time points into a capacity value time series input vector and a pressure value time series input vector according to the time dimension; a parameter time series feature interaction correlation analysis module 353, which is used to perform time series feature interaction correlation analysis on the capacity value time series input vector and the pressure value time series input vector to obtain a capacity-pressure time series interaction feature; and a storage tank body exhaust control module 354, which is used to determine whether to exhaust or inflate the storage tank body based on the capacity-pressure time series interaction feature.

[0037] Specifically, the data acquisition module 351 is configured to acquire capacity values ​​and pressure values ​​at a plurality of predetermined time points within a predetermined time period acquired by the pressure sensor and the capacity sensor.

[0038] Accordingly, in one possible implementation, the following steps can be used to obtain the capacity and pressure values ​​collected by the pressure sensor and the capacity sensor at multiple predetermined time points within a predetermined time period. For example, the following steps may be performed: determining a predetermined time period for obtaining capacity and pressure values; determining multiple predetermined time points within the predetermined time period as needed; ensuring that the data acquisition system or controller is properly configured and capable of receiving and recording data from the capacity and pressure sensors. This may involve setting parameters such as sampling frequency and data storage location; reading the capacity value from the capacity sensor at each predetermined time point; and using appropriate methods based on the sensor's interface and communication protocol to obtain the capacity value. This may involve sending a query command and receiving a response from the sensor; reading the pressure value from the pressure sensor at each predetermined time point. Similarly, using appropriate methods based on the sensor's interface and communication protocol to obtain the pressure value; and recording the capacity and pressure values ​​at each time point. The data may be stored in the data acquisition system or controller, or exported to an external storage device or database for long-term storage and analysis; and performing analysis and application on the obtained capacity and pressure values. This may include generating trend graphs, calculating averages or maximum / minimum values, and comparing against set thresholds.

[0039] Specifically, the data parameter time series arrangement module 352 is used to arrange the capacity values ​​and pressure values ​​of the multiple predetermined time points into a capacity value time series input vector and a pressure value time series input vector according to the time dimension. Considering that the capacity value and pressure value of the low-temperature storage tank are constantly changing in the time dimension, they have a time series dynamic change regularity in the time dimension, that is, the capacity values ​​and pressure values ​​of the multiple predetermined time points each have a time series correlation relationship. Therefore, in the technical solution of the present application, in order to capture and depict the time series dynamic change characteristics of the capacity value and pressure value of the low-temperature storage tank in the time dimension, it is necessary to arrange the capacity values ​​and pressure values ​​of the multiple predetermined time points into a capacity value time series input vector and a pressure value time series input vector according to the time dimension, so as to respectively integrate the time series distribution information of the capacity value and pressure value in the time dimension.

[0040] Specifically, the parameter time series feature interaction correlation analysis module 353 is used to perform time series feature interaction correlation analysis on the capacity value time series input vector and the pressure value time series input vector to obtain the capacity-pressure time series interaction feature. In particular, in a specific example of the present application, Figure 4 As shown, the parameter time series feature interaction correlation analysis module 353 includes: an upsampling unit 3531, which is used to pass the capacity value time series input vector and the pressure value time series input vector through the upsampling module respectively to obtain an upsampled capacity value time series input vector and an upsampled pressure value time series input vector; a parameter time series feature extraction unit 3532, which is used to perform time series feature extraction on the upsampled capacity value time series input vector and the upsampled pressure value time series input vector respectively through a time series feature extractor based on a deep neural network model to obtain a capacity value time series feature vector and a pressure value time series feature vector; a parameter time series feature interaction unit 3533, which is used to perform time series feature interaction correlation encoding on the capacity value time series feature vector and the pressure value time series feature vector to obtain the capacity-pressure time series interaction feature.

[0041] More specifically, the upsampling unit 3531 is configured to pass the capacity value time series input vector and the pressure value time series input vector, respectively, through an upsampling module to obtain an upsampled capacity value time series input vector and an upsampled pressure value time series input vector. In order to increase sensitivity to changes in pressure within the storage tank body and thereby more fully capture and depict the time series characteristic information of the pressure changes within the storage tank body, the technical solution of the present application further passes the capacity value time series input vector and the pressure value time series input vector, respectively, through an upsampling module to obtain an upsampled capacity value time series input vector and an upsampled pressure value time series input vector. This increases the density and smoothness of the pressure and capacity data, thereby facilitating a better subsequent representation of the time series characteristics of both data, thereby enabling more accurate pressure control.

[0042] It is worth noting that upsampling is a signal processing and data processing technique used to increase the sampling rate or resolution of a signal or data. During upsampling, the sampling points of the original signal or data are inserted into new sampling points to increase the detail and accuracy of the signal or data.

[0043] Accordingly, in a possible implementation, the capacity value time series input vector and the pressure value time series input vector can be respectively passed through an upsampling module to obtain an upsampled capacity value time series input vector and an upsampled pressure value time series input vector through the following steps, for example: determining how many times the sampling rate of the input vector is to be increased. The upsampling multiple can be selected according to specific needs, such as 2 times, 4 times, etc.; inserting blank values ​​or zero values ​​between each sampling point in the capacity value time series input vector to increase the number of sampling points. The number of inserted blank values ​​is determined by the upsampling multiple; using an interpolation algorithm to estimate the inserted blank values ​​to obtain the upsampled capacity value time series input vector. Commonly used interpolation algorithms include linear interpolation, spline interpolation, and convolution interpolation; inserting blank values ​​or zero values ​​between each sampling point in the pressure value time series input vector to increase the number of sampling points. The number of inserted blank values ​​is determined by the upsampling multiple; using an interpolation algorithm to estimate the inserted blank values ​​to obtain the upsampled pressure value time series input vector. Common interpolation algorithms include linear interpolation, spline interpolation, and convolution interpolation. At this time, the upsampled capacity value time series input vector and the upsampled pressure value time series input vector respectively contain data of the original input vector after upsampling.

[0044] More specifically, the parameter time series feature extraction unit 3532 is configured to perform time series feature extraction on the upsampled capacity value time series input vector and the upsampled pressure value time series input vector, respectively, through a time series feature extractor based on a deep neural network model, to obtain a capacity value time series feature vector and a pressure value time series feature vector. That is, in a specific example of the present application, the upsampled capacity value time series input vector and the upsampled pressure value time series input vector are subjected to feature mining in a time series feature extractor based on a one-dimensional convolutional layer, respectively, to extract the time series dynamic correlation feature information of the pressure value and capacity value of the cryogenic storage tank in the time dimension, respectively, thereby obtaining a capacity value time series feature vector and a pressure value time series feature vector. Specifically, each layer of the one-dimensional convolutional layer-based temporal feature extractor performs the following on the input data in the forward pass of the layer: convolution processing on the input data to obtain a convolution feature map; pooling based on the feature matrix on the convolution feature map to obtain a pooled feature map; and nonlinear activation on the pooled feature map to obtain an activation feature map; wherein the output of the last layer of the one-dimensional convolutional layer-based temporal feature extractor is the capacity value temporal feature vector and the pressure value temporal feature vector, and the input of the first layer of the one-dimensional convolutional layer-based temporal feature extractor is the upsampled capacity value temporal input vector and the upsampled pressure value temporal input vector.

[0045] It is worth noting that the one-dimensional convolution layer is a type of layer commonly used in deep learning to process one-dimensional sequence data, such as time series data or text data. The one-dimensional convolution layer can extract local patterns and features in the input data and share weights at different positions, thereby achieving feature extraction and representation learning of the input data. The operation of the one-dimensional convolution layer is as follows: Input data: The input of the one-dimensional convolution layer is a one-dimensional data sequence, usually represented as a vector; Convolution kernel: The one-dimensional convolution layer contains one or more convolution kernels (also called filters or feature detectors). Each convolution kernel is a small one-dimensional weight vector used to perform a sliding window convolution operation on the input data; Convolution operation: The convolution kernel performs a sliding window convolution operation on the input data. For each position, the convolution kernel is element-wise multiplied with a part of the input data, and the results are then added to obtain an element of the convolution output; Feature map: The result of the convolution operation is a new one-dimensional sequence called a feature map. The length of the feature map can be determined by the convolution kernel settings and the length of the input data. Activation function: Typically, after the convolution operation, an activation function is applied to each element of the feature map to introduce nonlinear features and enhance the network's representational capabilities. Optional pooling operation: After the one-dimensional convolution layer, a pooling operation (such as max pooling or average pooling) can be applied to reduce the length of the feature map and reduce the data dimension, thereby reducing the number of parameters and computational effort. One-dimensional convolutional layers are widely used in deep learning for various tasks such as speech recognition, natural language processing, and time series prediction. They can automatically learn local patterns and features in the input data and build more complex models by stacking multiple convolutional layers.

[0046] More specifically, the parameter time series feature interaction unit 3533 is used to perform time series feature interaction correlation encoding on the capacity value time series feature vector and the pressure value time series feature vector to obtain the capacity-pressure time series interaction feature. In the technical solution of the present application, a feature interaction layer is used to perform feature interaction based on the attention mechanism on the capacity value time series feature vector and the pressure value time series feature vector to obtain the capacity-pressure time series interaction feature vector, so as to capture the association and mutual influence between the capacity time series change feature and the pressure time series change feature. It should be understandable that since the goal of the traditional attention mechanism is to learn an attention weight matrix, larger weights are assigned to important features and smaller weights are assigned to secondary features, thereby selecting information that is more critical to the current task goal. This approach focuses more on weighting the importance of each feature and ignores the dependency between features. The feature interaction layer can capture the correlation and mutual influence between the capacity time series change features and the pressure time series change features through feature interaction based on the attention mechanism. It can learn the dependencies between different features related to the capacity time series change and pressure time series change of the cryogenic storage tank, and interact and integrate the features based on these dependencies to obtain the capacity-pressure time series interaction feature vector, which is conducive to more accurate subsequent adaptive regulation of pressure.

[0047] Accordingly, in one possible implementation, a feature interaction layer can be used to perform an attention-based feature interaction on the capacity value time series feature vector and the pressure value time series feature vector to obtain a capacity-pressure time series interaction feature vector as the capacity-pressure time series interaction feature. For example, the following steps may be used: the capacity value time series feature vector and the pressure value time series feature vector are used as input; the feature interaction layer is used to interact with the capacity value time series feature vector and the pressure value time series feature vector to capture the relationship and dependency between them; in the feature interaction layer, an attention mechanism is used to calculate the attention weight between the capacity value time series feature vector and the pressure value time series feature vector. The attention weight represents the importance weight between the capacity value and the pressure value at each time step; the attention weight is used to perform a weighted addition of the capacity value time series feature vector and the pressure value time series feature vector to obtain the capacity-pressure time series interaction feature vector. The higher the attention weight, the greater the contribution of the capacity value and pressure value at that time step to the feature interaction; the capacity-pressure time series interaction feature vector serves as the output of the feature interaction layer and is used in subsequent tasks or models.

[0048] It is worth mentioning that in other specific examples of the present application, the capacity value time series input vector and the pressure value time series input vector can also be analyzed for time series feature interaction correlation in other ways to obtain capacity-pressure time series interaction features, for example: create an empty capacity-pressure time series interaction feature vector for storing the interaction features between capacity and pressure. The length of this vector should be equal to the number of predetermined time points; traverse the predetermined time points in chronological order: for each predetermined time point: obtain the capacity value and pressure value at the current time point; according to the needs, different interaction feature calculation methods can be selected, for example: direct splicing: directly splicing the capacity value and the pressure value into a feature vector; product interaction: multiplying the capacity value and the pressure value to obtain a new feature; difference interaction: subtracting the pressure value from the capacity value to obtain a new feature; other interaction methods: according to specific needs, other interaction methods can be tried, such as sum, average, etc.; add the calculated interaction features to the capacity-pressure time series interaction feature vector; at this time, the elements in the capacity-pressure time series interaction feature vector contain the capacity-pressure interaction features at each predetermined time point in chronological order.

[0049] Specifically, the storage tank body exhaust control module 354 is configured to determine to exhaust or inflate the storage tank body based on the capacity-pressure time series interaction features. In particular, in one specific example of the present application, as shown in Figure 5 Specifically, the storage tank body exhaust control module 354 is configured to determine to exhaust or inflate the storage tank body based on the capacity-pressure time series interaction features. In particular, in one specific example of the present application, as shown in

[0050] More specifically, the feature gain unit 3541 is used to perform distribution gain based on the probability density feature imitation paradigm on the capacity-pressure time series interaction feature vector to obtain the gained capacity-pressure time series interaction feature vector. In particular, in the technical solution of the present application, when the upsampled capacity value time series input vector and the upsampled pressure value time series input vector are respectively passed through a time series feature extractor based on a one-dimensional convolutional layer to obtain a capacity value time series feature vector and a pressure value time series feature vector, the capacity value time series feature vector and the pressure value time series feature vector respectively express the local time series correlation features of the capacity value and the pressure value. In this way, when the feature interaction layer is used to perform feature interaction based on the attention mechanism on the capacity value time series feature vector and the pressure value time series feature vector, the dependency relationship features between the capacity value time series feature vector and the pressure value time series feature vector can be extracted. However, relative to the local temporal correlation features expressed by the capacity value time series feature vector and the pressure value time series feature vector as foreground object features, when performing dependency feature extraction based on attention-based feature interaction, background distribution noise related to the distribution interference of the local temporal correlation features expressed by the capacity value time series feature vector and the pressure value time series feature vector will also be introduced. In addition, the capacity-pressure time series interaction feature vector also has a hierarchical feature expression in the time domain space and interaction space of the capacity value time series feature vector and the pressure value time series feature vector. Therefore, it is expected to enhance its expression effect based on the distribution characteristics of the capacity-pressure time series interaction feature vector. Therefore, the applicant of this application performs a distribution gain based on the probability density feature imitation paradigm on the capacity-pressure time series interaction feature vector, which is specifically expressed as:

[0051]

[0052] Where V is the volume-pressure time series interaction characteristic vector, v i is the eigenvalue of the i-th position of the volume-pressure time series interaction feature vector, N is the length of the volume-pressure time series interaction feature vector, represents the square of the bi-norm of the volume-pressure time series interaction feature vector, α is a weighted hyperparameter, exp(·) represents the exponential operation, and v' iis the eigenvalue of the i-th position of the volume-pressure time series interaction feature vector after the gain. Here, based on the feature imitation paradigm of the standard Cauchy distribution for the natural Gaussian distribution on the probability density, the distribution gain based on the probability density feature imitation paradigm can use the feature scale as an imitation mask to distinguish the foreground object features and the background distribution noise in the high-dimensional feature space, thereby performing a distribution soft matching of the semantic cognition of the feature space mapping of the high-dimensional space based on the spatial hierarchical semantics of the high-dimensional features to obtain the unconstrained distribution gain of the high-dimensional feature distribution, thereby improving the expression effect of the volume-pressure time series interaction feature vector based on the feature distribution characteristics, and thus improving the expression effect of the transfer matrix of the pressure value time series feature vector relative to the volume-pressure time series interaction feature vector, that is, improving the accuracy of the classification result obtained by the classifier of the transfer matrix as the pressure transfer correlation feature matrix, thereby improving the control accuracy and intelligence of the exhaust or inflation of the intelligent cryogenic storage tank. In this way, the tank body can be automatically vented or inflated based on the temporal changes between the capacity and pressure of the low-temperature tank to achieve adaptive adjustment of the pressure inside the tank, thereby improving the automation and safety of tank operation, reducing manual intervention and maintenance costs, and lowering operational risks, thereby improving storage efficiency and safety.

[0053] More specifically, the pressure transfer correlation encoding unit 3542 is configured to calculate a transfer matrix of the pressure value time series feature vector relative to the post-gain capacity-pressure time series interaction feature vector as a pressure transfer correlation feature matrix. In other words, to further improve the accuracy of pressure control, the technical solution of the present application further calculates a transfer matrix of the pressure value time series feature vector relative to the capacity-pressure time series interaction feature vector as a pressure transfer correlation feature matrix. This allows the pressure time series variation characteristics to be mapped into the high-dimensional space of the pressure and capacity time series interaction correlation characteristics of the cryogenic storage tank, thereby more accurately characterizing the pressure time series variation characteristics and degree within the tank.

[0054] More specifically, the tank pressure control unit 3543 is configured to pass the pressure transfer correlation feature matrix through a classifier to obtain a classification result, wherein the classification result is used to indicate whether the tank body is to be vented or inflated. That is, classification processing is performed using information about the temporal variation characteristics of the pressure within the cryogenic tank based on the temporal interaction correlation characteristics of the pressure and capacity within the cryogenic tank, thereby controlling the venting or inflation of the tank body to adaptively adjust the pressure value changes within the tank. In this way, automatic regulation of the pressure within the tank can be achieved, further improving the automation and safety of tank operations and reducing manual intervention and maintenance costs. Specifically, the pressure transfer correlation feature matrix is ​​expanded into a classification feature vector based on row vectors or column vectors; the classification feature vector is fully connected encoded using multiple fully connected layers of the classifier to obtain an encoded classification feature vector; and the encoded classification feature vector is passed through the Softmax classification function of the classifier to obtain the classification result.

[0055] A classifier is a machine learning model or algorithm that is used to classify input data into different categories or labels. Classifiers are part of supervised learning and perform classification tasks by learning a mapping from input data to output categories.

[0056] A fully connected layer is a common layer type in neural networks. In a fully connected layer, each neuron is connected to all neurons in the previous layer, and each connection has a weight. This means that each neuron in a fully connected layer receives input from all neurons in the previous layer, performs a weighted sum of these inputs using the weights, and then passes the result to the next layer.

[0057] The Softmax classification function is a commonly used activation function for multi-classification problems. It converts each element of the input vector into a probability value between 0 and 1, where the sum of these probabilities equals 1. The Softmax function is often used in the output layer of neural networks and is particularly well-suited for multi-classification problems because it maps the network output into a probability distribution for each category. During training, the output of the Softmax function is used to calculate the loss function and update the network parameters through the backpropagation algorithm. It is worth noting that the output of the Softmax function does not change the relative size of the elements; it simply normalizes them. Therefore, the Softmax function does not change the characteristics of the input vector; it simply converts it into a probability distribution.

[0058] It's worth noting that in other specific examples of the present application, other methods can be used to determine whether to vent or inflate the tank based on the capacity-pressure time series interaction feature. For example, the capacity-pressure time series interaction feature can be used as input; a venting or inflation threshold can be set based on specific issues and needs. The threshold can be a fixed, pre-set value or dynamically adjusted based on historical data and experience; the pressure portion of the capacity-pressure time series interaction feature is compared with the venting threshold. If the pressure is below the venting threshold, a venting operation is determined to be necessary; the pressure portion of the capacity-pressure time series interaction feature is compared with the inflation threshold. If the pressure is above the inflation threshold, a inflation operation is determined to be necessary; and based on the venting or inflation determination, a venting or inflation operation is determined to be performed on the tank. For example, based on the capacity-pressure time series interaction feature, a threshold can be set to determine whether the tank needs to be vented or inflated. The venting or inflation determination can be based on a set pressure threshold: if the pressure is below the venting threshold, venting is required; if the pressure is above the inflation threshold, inflation is required. In this way, the storage tank can be operated in a timely manner according to the changes in the capacity-pressure time series interaction characteristics to maintain the appropriate pressure level.

[0059] As described above, the intelligent cryogenic storage tank 300 according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with an intelligent cryogenic storage algorithm. In one possible implementation, the intelligent cryogenic storage tank 300 according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or hardware module. For example, the intelligent cryogenic storage tank 300 can be a software module in the operating system of the wireless terminal, or it can be an application developed for the wireless terminal; of course, the intelligent cryogenic storage tank 300 can also be one of the many hardware modules of the wireless terminal.

[0060] Alternatively, in another example, the smart cryogenic storage tank 300 and the wireless terminal may be separate devices, and the smart cryogenic storage tank 300 may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0061] Furthermore, a method for using an intelligent low-temperature storage tank is also provided.

[0062] Figure 6 Flowchart of the method for using the intelligent cryogenic storage tank according to the embodiment of the present application. Figure 6As shown, the method for using the intelligent cryogenic storage tank according to the embodiment of the present application includes the following steps: S1, obtaining the capacity value and pressure value at multiple predetermined time points within a predetermined time period collected by the pressure sensor and the capacity sensor; S2, arranging the capacity value and pressure value at the multiple predetermined time points into a capacity value time series input vector and a pressure value time series input vector according to the time dimension; S3, performing a time series feature interaction correlation analysis on the capacity value time series input vector and the pressure value time series input vector to obtain a capacity-pressure time series interaction feature; S4, determining whether to vent or inflate the storage tank body based on the capacity-pressure time series interaction feature.

[0063] In summary, the method of using the intelligent cryogenic storage tank according to the embodiment of the present application is explained, which measures the pressure value and the overall capacity value inside the cryogenic storage tank in real time through a pressure sensor and a capacity sensor respectively, and introduces data processing and analysis algorithms at the back end to perform time-series interactive correlation analysis of the capacity value and the pressure value, so as to control the exhaust or inflation of the tank body to adaptively adjust the pressure inside the tank. In this way, the automatic adjustment of the pressure inside the tank can be achieved, which further improves the degree of automation and safety of the tank operation, reduces manual intervention and maintenance costs, and reduces operational risks, thereby improving storage efficiency and safety.

[0064] While various embodiments of the present disclosure have been described above, the above descriptions are illustrative, non-exhaustive, and not intended to be limiting of the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or improvements to existing technologies, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. An intelligent cryogenic storage tank, characterized in that: include: Storage tank body, used to store cryogenic substances; A pressure sensor is provided inside the storage tank body, and the pressure sensor is used to detect the pressure value inside the storage tank body; A capacity sensor is provided on the outside of the storage tank body, and is used to detect the capacity value of the storage tank body; exhaust system; a controller, the controller being communicatively connected to the pressure sensor, the capacity sensor, and the exhaust device, the controller being configured to control the exhaust device to exhaust or inflate the storage tank; The controller includes: a data acquisition module, configured to acquire capacity values ​​and pressure values ​​collected by the pressure sensor and the capacity sensor at a plurality of predetermined time points within a predetermined time period; a data parameter time series arrangement module, configured to arrange the capacity values ​​and pressure values ​​at the plurality of predetermined time points into a capacity value time series input vector and a pressure value time series input vector according to a time dimension; a parameter time series feature interaction correlation analysis module, configured to perform time series feature interaction correlation analysis on the capacity value time series input vector and the pressure value time series input vector to obtain a capacity-pressure time series interaction feature; a storage tank exhaust control module, configured to determine whether to exhaust or inflate the storage tank based on the capacity-pressure time series interaction characteristics; The capacity-pressure time series interaction feature is a capacity-pressure time series interaction feature vector. The storage tank exhaust control module includes: a feature gain unit, configured to perform distribution gain on the volume-pressure time series interaction feature vector based on a probability density feature mimicking paradigm to obtain a gained volume-pressure time series interaction feature vector; a pressure transfer correlation coding unit, configured to calculate a transfer matrix of the pressure value time series eigenvector relative to the post-gain capacity-pressure time series interaction eigenvector as a pressure transfer correlation characteristic matrix; a tank pressure control unit, configured to pass the pressure transfer correlation feature matrix through a classifier to obtain a classification result, wherein the classification result is used to indicate whether the tank is to be vented or inflated; The feature gain unit is configured to perform distribution gain based on a probability density feature mimicking paradigm on the volume-pressure time series interaction feature vector using the following optimization formula to obtain the gained volume-pressure time series interaction feature vector; Wherein, the optimization formula is: in, is the volume-pressure time series interaction feature vector, is the volume-pressure time series interaction feature vector The eigenvalues ​​at the positions, is the length of the volume-pressure time series interaction feature vector, represents the square of the binormal of the volume-pressure time series interaction eigenvector, and is a weighted hyperparameter, represents exponential operation, is the first characteristic vector of the volume-pressure time series interaction after the gain The eigenvalues ​​at each position.

2. The intelligent cryogenic storage tank according to claim 1, characterized in that: The parameter time series feature interaction correlation analysis module includes: an upsampling unit, configured to pass the capacity value time series input vector and the pressure value time series input vector through an upsampling module to obtain an upsampled capacity value time series input vector and an upsampled pressure value time series input vector respectively; a parameter time series feature extraction unit, configured to perform time series feature extraction on the upsampled capacity value time series input vector and the upsampled pressure value time series input vector respectively through a time series feature extractor based on a deep neural network model to obtain a capacity value time series feature vector and a pressure value time series feature vector; A parameter time series feature interaction unit is used to perform time series feature interaction correlation coding on the capacity value time series feature vector and the pressure value time series feature vector to obtain the capacity-pressure time series interaction feature.

3. The intelligent cryogenic storage tank according to claim 2, characterized in that: The temporal feature extractor based on the deep neural network model is a temporal feature extractor based on a one-dimensional convolutional layer.

4. The intelligent cryogenic storage tank according to claim 3, characterized in that: The parameter time series feature interaction unit is used to: use a feature interaction layer to perform feature interaction based on an attention mechanism on the capacity value time series feature vector and the pressure value time series feature vector to obtain a capacity-pressure time series interaction feature vector.

5. The intelligent cryogenic storage tank according to claim 4, characterized in that: The storage tank pressure control unit includes: an expansion subunit, configured to expand the pressure transfer correlation feature matrix into a classification feature vector based on a row vector or a column vector; a fully connected encoding subunit, configured to perform fully connected encoding on the classification feature vector using multiple fully connected layers of the classifier to obtain an encoded classification feature vector; and The classification result generating subunit is used to pass the encoded classification feature vector through the Softmax classification function of the classifier to obtain the classification result.

6. A method for using an intelligent cryogenic storage tank, performed by the intelligent cryogenic storage tank according to any one of claims 1 to 5, characterized in that: include: Acquiring capacity values ​​and pressure values ​​collected by the pressure sensor and the capacity sensor at a plurality of predetermined time points within a predetermined time period; Arranging the capacity values ​​and pressure values ​​at the plurality of predetermined time points into a capacity value time series input vector and a pressure value time series input vector according to a time dimension; performing a time series feature interaction correlation analysis on the capacity value time series input vector and the pressure value time series input vector to obtain a capacity-pressure time series interaction feature; Based on the capacity-pressure time series interaction characteristics, it is determined whether to exhaust or inflate the storage tank body.

Images