CN116130721A - Status diagnostic system and method for hydrogen fuel cell - Google Patents
Status diagnostic system and method for hydrogen fuel cell Download PDFInfo
- Publication number
- CN116130721A CN116130721A CN202310376056.6A CN202310376056A CN116130721A CN 116130721 A CN116130721 A CN 116130721A CN 202310376056 A CN202310376056 A CN 202310376056A CN 116130721 A CN116130721 A CN 116130721A
- Authority
- CN
- China
- Prior art keywords
- matrix
- semantic understanding
- operation parameter
- full
- fuel cell
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
- H01M8/00—Fuel cells; Manufacture thereof
- H01M8/04—Auxiliary arrangements, e.g. for control of pressure or for circulation of fluids
- H01M8/04298—Processes for controlling fuel cells or fuel cell systems
- H01M8/04992—Processes for controlling fuel cells or fuel cell systems characterised by the implementation of mathematical or computational algorithms, e.g. feedback control loops, fuzzy logic, neural networks or artificial intelligence
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01M—PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
- H01M8/00—Fuel cells; Manufacture thereof
- H01M8/22—Fuel cells in which the fuel is based on materials comprising carbon or oxygen or hydrogen and other elements; Fuel cells in which the fuel is based on materials comprising only elements other than carbon, oxygen or hydrogen
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
- Y02E60/30—Hydrogen technology
- Y02E60/50—Fuel cells
Landscapes
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Chemical & Material Sciences (AREA)
- Electrochemistry (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Chemical & Material Sciences (AREA)
- Sustainable Energy (AREA)
- Sustainable Development (AREA)
- Manufacturing & Machinery (AREA)
- Automation & Control Theory (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Medical Informatics (AREA)
- Fuzzy Systems (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Fuel Cell (AREA)
Abstract
The application relates to the technical field of intelligent diagnosis of battery states, and particularly discloses a state diagnosis system and method of a hydrogen fuel battery, which adopt an artificial intelligent monitoring and diagnosis technology based on deep learning to excavate high-dimensional implicit associated characteristic information which is contained among various operation parameters of the hydrogen fuel battery in the operation process through a context encoder based on a converter, and enhance the change condition of the characteristic information on each operation parameter in a battery anode and a battery cathode in time sequence by utilizing a spatial attention mechanism so as to carry out classification processing. Thus, the operation state detection of the hydrogen fuel cell is carried out based on the global time sequence related characteristic distribution information of each operation parameter of the hydrogen fuel cell, so that the performance of the cell is optimized, and accidents are avoided.
Description
Technical Field
The present disclosure relates to the field of intelligent diagnosis of battery status, and more particularly, to a status diagnosis system and method for a hydrogen fuel cell.
Background
The hydrogen fuel cell is a power generation device for directly converting chemical energy of hydrogen and oxygen into electric energy, and the basic principle is that the hydrogen and the oxygen are respectively supplied to an anode and a cathode by reverse reaction of electrolytic water, and after the hydrogen is diffused outwards through the anode and reacts with an electrolyte, electrons are released to reach the cathode through an external load.
In the long-term use process of the hydrogen fuel cell, many parameters needing to be monitored, such as hydrogen flow, temperature, humidity, discharge voltage, output current and the like, have important influences on the running state and the cell performance, and if hydrogen leakage or performance degradation occurs, the generated energy of the hydrogen fuel cell is reduced, even certain potential safety hazard is caused, so that a scheme capable of effectively monitoring and diagnosing the state of the hydrogen fuel cell is needed.
Disclosure of Invention
The present application has been made in order to solve the above technical problems. Embodiments of the present application provide a system and a method for diagnosing a state of a hydrogen fuel cell, which employ artificial intelligence monitoring and diagnosing techniques based on deep learning to mine high-dimensional implicit correlation characteristic information included between various operation parameters of the hydrogen fuel cell in an operation process through a context encoder based on a converter, and enhance time-series variation conditions of the various operation parameters in an anode and a cathode of the cell in the characteristic information by using a spatial attention mechanism, so as to perform classification processing. Thus, the operation state detection of the hydrogen fuel cell is carried out based on the global time sequence related characteristic distribution information of each operation parameter of the hydrogen fuel cell, so that the performance of the cell is optimized, and accidents are avoided.
Accordingly, according to one aspect of the present application, there is provided a condition diagnosing system of a hydrogen fuel cell, including:
the operation state monitoring module is used for acquiring operation parameters of a plurality of preset time points of the hydrogen fuel cell to be detected in a preset time period in the operation process, wherein the operation parameters comprise hydrogen flow, temperature, humidity, discharge voltage and output current;
an operation parameter context coding module, configured to pass the operation parameters of the respective predetermined time points through a context encoder based on a converter to obtain a plurality of context operation parameter semantic understanding feature vectors;
the integration module is used for two-dimensionally arranging the context operation parameter semantic understanding feature vectors into a full-time sequence context operation parameter semantic understanding matrix;
the spatial feature extraction module is used for enabling the full-time sequence context operation parameter semantic understanding matrix to obtain a full-time sequence context operation parameter semantic understanding feature matrix through a convolutional neural network model using a spatial attention mechanism;
the distinguishing and optimizing module is used for distinguishing the characteristic values of the semantic understanding characteristic matrix of the full-time sequence context operation parameters so as to obtain the semantic understanding characteristic matrix of the optimized full-time sequence context operation parameters; and
The diagnosis result generation module is used for enabling the semantic understanding feature matrix of the optimized full-time sequence context operation parameters to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the operation state of the hydrogen fuel cell to be detected is normal or not.
In the above-described hydrogen fuel cell condition diagnosis system, the operation parameter context encoding module includes: an encoding unit for inputting an operating parameter at a single predetermined point in time into the converter-based context encoder to obtain the plurality of single-point operating parameter feature vectors; and the cascading unit is used for cascading the plurality of single-point operation parameter feature vectors to obtain the context operation parameter semantic understanding feature vector.
In the above-described hydrogen fuel cell condition diagnosis system, the encoding unit is further configured to: arranging the operation parameters of the single preset time point into an input vector; respectively converting the input vector into a query vector and a key vector through a learning embedding matrix; calculating the product between the query vector and the transpose vector of the key vector to obtain a self-attention correlation matrix; carrying out standardization processing on the self-attention association matrix to obtain a standardized self-attention association matrix; inputting the standardized self-attention association matrix into a Softmax activation function to activate so as to obtain a self-attention feature matrix; and multiplying the self-attention feature matrix with each of the operation parameters at the single predetermined time point as a value vector to obtain the plurality of single-point operation parameter feature vectors.
In the above-described hydrogen fuel cell condition diagnosis system, the spatial feature extraction module is further configured to: performing depth convolution coding on the full-time sequence context operation parameter semantic understanding matrix by using a convolution coding part of the convolution neural network model to obtain an initial convolution feature map; inputting the initial convolution feature map into a spatial attention portion of the convolution neural network model to obtain a spatial attention map; -passing said spatial attention map through a Softmax activation function to obtain a spatial attention profile; calculating the position-based points of the space attention feature map and the initial convolution feature map to obtain a full-time sequence context operation parameter semantic understanding feature map; and carrying out global average pooling processing along the channel dimension on the full-time sequence context operation parameter semantic understanding feature map to obtain the full-time sequence context operation parameter semantic understanding feature matrix.
In the above-described hydrogen fuel cell condition diagnosis system, the distinction optimizing module is further configured to: performing interactive reinforcement based on distinguishable physical excitation on the full-time sequence context operation parameter semantic understanding feature matrix by using the following formula to obtain the optimized full-time sequence context operation parameter semantic understanding feature matrix; wherein, the formula is:
Wherein the method comprises the steps ofIs the semantic understanding feature matrix of the full-time sequence context operation parameters, and is->And->Is a predetermined superparameter,/->And->Representing the addition and subtraction of the feature matrix by position, the division representing each position of the feature matrix divided by the corresponding value, andrepresenting convolution operations through a single convolution layer, +.>Is the semantic understanding feature matrix of the optimized full-time sequence context operation parameters, and is +.>And->Is an intermediate matrix.
In the above-described hydrogen fuel cell condition diagnosis system, the diagnosis result generation module includes: the expansion unit is used for expanding the semantic understanding feature matrix of the optimized full-time sequence context operation parameter into classification feature vectors according to row vectors or column vectors; the probability unit is used for inputting the classification feature vector into a Softmax classification function of the classifier to obtain a probability value of the classification feature vector belonging to each classification label; and the classification result generation unit is used for determining the classification label corresponding to the maximum probability value as the classification result.
According to another aspect of the present application, there is also provided a condition diagnosing method of a hydrogen fuel cell, including:
acquiring operation parameters of a hydrogen fuel cell to be detected at a plurality of preset time points in a preset time period in the operation process, wherein the operation parameters comprise hydrogen flow, temperature, humidity, discharge voltage and output current;
Passing the operating parameters at each predetermined point in time through a context encoder based on a converter to obtain a plurality of context operating parameter semantic understanding feature vectors;
two-dimensionally arranging the plurality of context operation parameter semantic understanding feature vectors into a full-time sequence context operation parameter semantic understanding matrix;
the full-time sequence context operation parameter semantic understanding matrix is obtained through a convolutional neural network model using a spatial attention mechanism;
performing eigenvalue distinction on the full-time sequence context operation parameter semantic understanding feature matrix to obtain an optimized full-time sequence context operation parameter semantic understanding feature matrix; and
and the semantic understanding feature matrix of the optimized full-time sequence context operation parameters is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the operation state of the hydrogen fuel cell to be detected is normal or not.
In the above-described state diagnosis method of a hydrogen fuel cell, passing the operation parameters at the respective predetermined time points through a context encoder based on a converter to obtain a plurality of context operation parameter semantic understanding feature vectors, comprising: inputting operating parameters at a single predetermined point in time into the converter-based context encoder to obtain the plurality of single-point operating parameter feature vectors; and cascading the plurality of single-point operation parameter feature vectors to obtain the context operation parameter semantic understanding feature vector.
In the above-described state diagnosis method of a hydrogen fuel cell, inputting an operation parameter at a single predetermined point in time to the converter-based context encoder to obtain the plurality of single-point operation parameter feature vectors, comprising: arranging the operation parameters of the single preset time point into an input vector; respectively converting the input vector into a query vector and a key vector through a learning embedding matrix; calculating the product between the query vector and the transpose vector of the key vector to obtain a self-attention correlation matrix; carrying out standardization processing on the self-attention association matrix to obtain a standardized self-attention association matrix; inputting the standardized self-attention association matrix into a Softmax activation function to activate so as to obtain a self-attention feature matrix; and multiplying the self-attention feature matrix with each of the operation parameters at the single predetermined time point as a value vector to obtain the plurality of single-point operation parameter feature vectors.
In the above method for diagnosing a state of a hydrogen fuel cell, the step of obtaining the full-time-series context operation parameter semantic understanding feature matrix by using a convolutional neural network model of a spatial attention mechanism includes: performing depth convolution coding on the full-time sequence context operation parameter semantic understanding matrix by using a convolution coding part of the convolution neural network model to obtain an initial convolution feature map; inputting the initial convolution feature map into a spatial attention portion of the convolution neural network model to obtain a spatial attention map; -passing said spatial attention map through a Softmax activation function to obtain a spatial attention profile; calculating the position-based points of the space attention feature map and the initial convolution feature map to obtain a full-time sequence context operation parameter semantic understanding feature map; and carrying out global average pooling processing along the channel dimension on the full-time sequence context operation parameter semantic understanding feature map to obtain the full-time sequence context operation parameter semantic understanding feature matrix.
In the above method for diagnosing a state of a hydrogen fuel cell, performing feature value differentiation on the full-time-sequence context operation parameter semantic understanding feature matrix to obtain an optimized full-time-sequence context operation parameter semantic understanding feature matrix, including: performing interactive reinforcement based on distinguishable physical excitation on the full-time sequence context operation parameter semantic understanding feature matrix by using the following formula to obtain the optimized full-time sequence context operation parameter semantic understanding feature matrix; wherein, the formula is:
wherein the method comprises the steps ofIs the semantic understanding feature matrix of the full-time sequence context operation parameters, and is->And->Is a predetermined superparameter,/->And->Representing the addition and subtraction of the feature matrix by position, the division representing each position of the feature matrix divided by the corresponding value, andrepresenting convolution operations through a single convolution layer, +.>Is the semantic understanding feature matrix of the optimized full-time sequence context operation parameters, and is +.>And->Is an intermediate matrix.
In the above method for diagnosing the state of the hydrogen fuel cell, the method for determining whether the operation state of the hydrogen fuel cell to be detected is normal includes the steps of: expanding the semantic understanding feature matrix of the optimized full-time sequence context operation parameters into classification feature vectors according to row vectors or column vectors; inputting the classification feature vector into a Softmax classification function of the classifier to obtain a probability value of the classification feature vector belonging to each classification label; and determining the classification label corresponding to the maximum probability value as the classification result.
According to still another aspect of the present application, there is provided an electronic apparatus including: a processor; and a memory in which computer program instructions are stored, which, when executed by the processor, cause the processor to perform the state diagnostic method of the hydrogen fuel cell as described above.
According to yet another aspect of the present application, there is provided a computer readable medium having stored thereon computer program instructions which, when executed by a processor, cause the processor to perform the method of diagnosing a condition of a hydrogen fuel cell as described above.
Compared with the prior art, the state diagnosis system and method for the hydrogen fuel cell adopt the artificial intelligent monitoring and diagnosis technology based on deep learning, so that the high-dimensional implicit associated characteristic information contained among all operation parameters of the hydrogen fuel cell in the operation process is mined through a context encoder based on a converter, and the time sequence change condition of all operation parameters in the anode and the cathode of the cell in the characteristic information is enhanced by utilizing a spatial attention mechanism, so that classification processing is performed. Thus, the operation state detection of the hydrogen fuel cell is carried out based on the global time sequence related characteristic distribution information of each operation parameter of the hydrogen fuel cell, so that the performance of the cell is optimized, and accidents are avoided.
Drawings
The foregoing and other objects, features and advantages of the present application will become more apparent from the following more particular description of embodiments of the present application, as illustrated in the accompanying drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and not constitute a limitation to the application. In the drawings, like reference numerals generally refer to like parts or steps.
Fig. 1 is an application scenario diagram of a condition diagnosing system of a hydrogen fuel cell according to an embodiment of the present application.
Fig. 2 is a block diagram of a condition diagnosing system of a hydrogen fuel cell according to an embodiment of the present application.
Fig. 3 is a schematic diagram of the architecture of a condition diagnosing system of a hydrogen fuel cell according to an embodiment of the present application.
Fig. 4 is a block diagram of a diagnostic result generation module in a status diagnostic system of a hydrogen fuel cell according to an embodiment of the present application.
Fig. 5 is a flowchart of a state diagnosis method of a hydrogen fuel cell according to an embodiment of the present application.
Fig. 6 is a block diagram of an electronic device according to an embodiment of the present application.
Detailed Description
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application and not all of the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
Summary of the application
As described above, in the long-term use process of the hydrogen fuel cell, there are many parameters to be monitored, such as hydrogen flow, temperature, humidity, discharge voltage, output current, etc., and all parameters of the cell have important effects on the operation state and the cell performance, and if hydrogen leakage or performance degradation occurs, the generated energy of the hydrogen fuel cell is reduced, and even a certain potential safety hazard is caused, so a solution capable of effectively monitoring and diagnosing the state of the hydrogen fuel cell is needed.
Accordingly, in consideration of the fact that in the process of monitoring the state of the hydrogen fuel cell, the conventional scheme only monitors and analyzes each parameter of the hydrogen fuel cell in the operation process to obtain the diagnosis result of the state of the hydrogen fuel cell, but the scheme does not take the association relation among each parameter of the hydrogen fuel cell into consideration, misjudgment or missed judgment can occur, the accuracy of the diagnosis result of the state of the hydrogen fuel cell is reduced, and a certain potential safety hazard is further provided. Therefore, in the technical scheme of the application, the operation state detection of the hydrogen fuel cell is expected to be performed based on the associated characteristic distribution information of each operation parameter of the hydrogen fuel cell in the operation process, so that the performance of the cell is optimized, and accidents are avoided. However, since the correlation characteristic information of each operation parameter exists not only between each operation parameter but also in the time dimension, in this process, there is a difficulty in how to accurately extract the global time-series correlation characteristic of each operation parameter of the hydrogen fuel cell during operation, so as to improve the accuracy of the state diagnosis of the hydrogen fuel cell.
In recent years, deep learning and neural networks have been widely used in the fields of computer vision, natural language processing, text signal processing, and the like. In addition, deep learning and neural networks have also shown levels approaching and even exceeding humans in the fields of image classification, object detection, semantic segmentation, text translation, and the like.
The development of deep learning and neural networks provides new solutions and schemes for mining global time sequence associated characteristic information of each operation parameter of the hydrogen fuel cell in the operation process. Those of ordinary skill in the art will appreciate that deep learning based deep neural network models may be adapted with appropriate training strategies, such as by gradient descent back-propagation algorithms, to adjust parameters of the deep neural network model to enable modeling of complex nonlinear correlations between things, which is obviously suitable for modeling and mining global timing-related characteristic information of various operating parameters of the hydrogen fuel cell during operation.
Specifically, in the technical scheme of the application, first, operation parameters of a to-be-detected hydrogen fuel cell at a plurality of preset time points in a preset time period in an operation process are obtained, wherein the operation parameters comprise hydrogen flow, temperature, humidity, discharge voltage and output current. Then, considering that the hydrogen fuel cell to be detected has a correlation between each operation parameter in the operation process, in order to fully extract the correlation feature distribution information between each operation parameter, in the technical scheme of the application, the operation parameters at each preset time point are encoded in a context encoder based on a converter, and the global context correlation feature distribution information about each operation parameter at each preset time point is extracted, so that a plurality of context operation parameter semantic understanding feature vectors are obtained. That is, based on the transform concept, the converter is used to capture the long-distance context-dependent characteristic, and the global context-based semantic coding is performed on each operation parameter at each predetermined time point to obtain a context semantic association feature representation, i.e. the context operation parameter semantic understanding feature vector, with the overall semantic association of each operation parameter at each predetermined time point as a context background. It should be appreciated that, in the technical solution of the present application, the respective operation parameters at the respective predetermined time points may be captured by the context encoder based on the converter, which is based on the global context semantic association feature representation.
Then, considering that the operation parameters of the hydrogen fuel cell to be detected in the operation process have not only the semantic relevance of the context among the parameters, but also the relevant characteristic information of the dynamics in the time dimension, that is, the operation parameters have the dynamic change rule in the time dimension. Therefore, in order to fully and accurately mine global time sequence related characteristics of each operation parameter of the hydrogen fuel cell in the operation process, after the context operation parameter semantic understanding characteristic vectors are two-dimensionally arranged into a full time sequence context operation parameter semantic understanding matrix, a convolution neural network model with excellent performance in terms of implicit characteristic extraction is used for carrying out characteristic mining of the full time sequence context operation parameter semantic understanding matrix.
In particular, in consideration of the fact that since the hydrogen fuel cell supplies hydrogen and oxygen to the anode and the cathode, respectively, in which the hydrogen diffuses outwardly through the anode and reacts with the electrolyte, electrons are discharged to the cathode through an external load, if leakage of hydrogen gas or degradation of performance occurs, the amount of generated electricity of the hydrogen fuel cell is reduced, variations in spatial positions with respect to the respective operation parameters in the anode and the cathode should be more focused in the actual process of detecting the operation state of the hydrogen fuel cell. In view of the ability of the attention mechanism to select the focus position, a more resolved representation of the feature is produced, and the feature after addition to the attention module will change adaptively as the network deepens. Therefore, in the technical scheme of the application, the full-time sequence context operation parameter semantic understanding matrix is processed in a convolutional neural network model by using a spatial attention mechanism to extract global time sequence dynamic associated feature information which is focused on the spatial positions of the anode and the cathode of the hydrogen fuel cell and is based on the operation parameters in the time dimension, so that the full-time sequence context operation parameter semantic understanding feature matrix is obtained. It should be noted that, the image features extracted by the spatial attention reflect the weights of the differences of the spatial dimension features, so as to inhibit or strengthen the features of different spatial positions, thereby extracting the time sequence dynamic associated feature information of the operation parameters which are spatially focused on the anode and the cathode of the hydrogen fuel cell.
Further, the semantic understanding feature matrix of the full-time sequence context operation parameter is used as a classification feature matrix to be subjected to classification processing in a classifier to obtain a classification result, and the classification result is used for indicating whether the operation state of the hydrogen fuel cell to be detected is normal or not. That is, in the technical solution of the present application, the label of the classifier includes that the operation state of the hydrogen fuel cell to be detected is normal, and the operation state of the hydrogen fuel cell to be detected is abnormal, where the classifier determines, through a soft maximum function, which classification label the classification feature matrix belongs to, so as to accurately detect and diagnose the operation state of the hydrogen fuel cell to be detected, thereby optimizing the performance of the battery and avoiding occurrence of accidents.
In particular, in the technical solution of the present application, for the full-time sequence context operation parameter semantic understanding feature matrix obtained by using the convolutional neural network model of the spatial attention mechanism, since the spatial attention mechanism strengthens the extraction of the associated features on the predetermined spatial position of the full-time sequence context operation parameter semantic understanding feature matrix, the feature values of some positions of the full-time sequence context operation parameter semantic understanding feature matrix have more significant importance with respect to the feature values of other positions, so if the feature values of the full-time sequence context operation parameter semantic understanding feature matrix can be effectively distinguished in the classification task, the training speed of the model and the accuracy of the classification result can be obviously improved.
Thus, the applicant of the present application semantically understands feature matrices for the full-time contextual operating parameters, e.g., noted asInteractive augmentation based on distinguishable physical stimulus is performed, expressed as:
wherein the method comprises the steps ofIs an optimized full time sequence contextOperational parameter semantic understanding feature matrix,>and->Is a predetermined superparameter,/->And->Representing the addition and subtraction of the feature matrix by position, division representing each position of the feature matrix divided by the corresponding value, and +.>Representing a convolution operation through a single convolution layer.
Here, the discriminative physical stimulus-based interaction enhancement is used to promote interactions between feature space and solution space of classification problems during back propagation through gradient descent, which extracts and mimics viable features (actionable feature) in a physical stimulus-like manner, whereby a general purpose low-dimensional guided physical stimulus approach is used to obtain a physical representation of viable features with gradient discriminativity, thereby enhancing the full-time context operational parameter semantic understanding feature matrix during trainingActive part in order to promote the semantic understanding feature matrix of the optimized full time sequence context operation parameters +. >The training speed under the classification task and the accuracy of the classification result of the semantic understanding characteristics of the full-time sequence context operation parameters after training. Therefore, the operation state of the hydrogen fuel cell can be accurately monitored and diagnosed in real time, so that the performance of the cell is optimized, and accidents are avoided.
Fig. 1 is an application scenario diagram of a condition diagnosing system of a hydrogen fuel cell according to an embodiment of the present application. As shown in fig. 1, in this application scenario, first, operation parameters of a hydrogen fuel cell (e.g., C as illustrated in fig. 1) to be detected at a plurality of predetermined time points during a predetermined period of time in operation are acquired, the operation parameters including a hydrogen flow rate, a temperature, a humidity, a discharge voltage, and an output current, the operation parameters being acquired by respective sensors disposed in an operation state monitoring module (e.g., 110 as illustrated in fig. 1) of a state diagnosis system (e.g., 100 as illustrated in fig. 1) of the hydrogen fuel cell, wherein the respective sensors include a flow sensor (e.g., se1 as illustrated in fig. 1), a temperature sensor (e.g., se2 as illustrated in fig. 1), a humidity sensor (e.g., se3 as illustrated in fig. 1), a voltmeter (e.g., se4 as illustrated in fig. 1), and a ammeter (e.g., se5 as illustrated in fig. 1). Further, the operation parameters of the hydrogen fuel cell to be detected at a plurality of predetermined time points in a predetermined period of time during operation are input to a server (e.g., S as illustrated in fig. 1) in which a state diagnosis algorithm of the hydrogen fuel cell is deployed, wherein the server is capable of processing the operation parameters of the hydrogen fuel cell to be detected at a plurality of predetermined time points in a predetermined period of time during operation based on the state diagnosis algorithm of the hydrogen fuel cell to obtain a classification result for indicating whether the operation state of the hydrogen fuel cell to be detected is normal.
Having described the basic principles of the present application, various non-limiting embodiments of the present application will now be described in detail with reference to the accompanying drawings.
Exemplary System
Fig. 2 is a block diagram of a condition diagnosing system of a hydrogen fuel cell according to an embodiment of the present application. As shown in fig. 2, a condition diagnosis system 100 of a hydrogen fuel cell according to an embodiment of the present application includes: an operation state monitoring module 110, configured to obtain operation parameters of the hydrogen fuel cell to be detected at a plurality of predetermined time points within a predetermined time period in an operation process, where the operation parameters include a hydrogen flow rate, a temperature, a humidity, a discharge voltage, and an output current; an operation parameter context coding module 120, configured to pass the operation parameters at the respective predetermined time points through a context encoder based on a converter to obtain a plurality of context operation parameter semantic understanding feature vectors; an integration module 130, configured to two-dimensionally arrange the plurality of context operation parameter semantic understanding feature vectors into a full-time sequence context operation parameter semantic understanding matrix; the spatial feature extraction module 140 is configured to obtain the full-time sequence context operation parameter semantic understanding feature matrix by using a convolutional neural network model of a spatial attention mechanism; the distinguishing and optimizing module 150 is configured to distinguish the feature values of the semantic understanding feature matrix of the full-time sequence context operation parameter to obtain an optimized semantic understanding feature matrix of the full-time sequence context operation parameter; and a diagnostic result generating module 160, configured to pass the optimized full-time sequence context operation parameter semantic understanding feature matrix through a classifier to obtain a classification result, where the classification result is used to indicate whether the operation state of the hydrogen fuel cell to be detected is normal.
Fig. 3 is a schematic diagram of the architecture of a condition diagnosing system of a hydrogen fuel cell according to an embodiment of the present application. As shown in fig. 3, first, operation parameters including hydrogen flow rate, temperature, humidity, discharge voltage, and output current at a plurality of predetermined time points in a predetermined period of time during operation of a hydrogen fuel cell to be detected are acquired; then, the operation parameters of each preset time point pass through a context encoder based on a converter to obtain a plurality of context operation parameter semantic understanding feature vectors; secondly, two-dimensionally arranging the context operation parameter semantic understanding feature vectors into a full-time sequence context operation parameter semantic understanding matrix; then, the full-time sequence context operation parameter semantic understanding matrix is obtained through a convolution neural network model using a spatial attention mechanism; then, distinguishing characteristic values of the semantic understanding characteristic matrix of the full-time sequence context operation parameters to obtain an optimized semantic understanding characteristic matrix of the full-time sequence context operation parameters; and finally, the semantic understanding feature matrix of the optimized full-time sequence context operation parameters is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the operation state of the hydrogen fuel cell to be detected is normal or not.
As described above, in the long-term use process of the hydrogen fuel cell, there are many parameters to be monitored, such as hydrogen flow, temperature, humidity, discharge voltage, output current, etc., and all parameters of the cell have important effects on the operation state and the cell performance, and if hydrogen leakage or performance degradation occurs, the generated energy of the hydrogen fuel cell is reduced, and even a certain potential safety hazard is caused, so a solution capable of effectively monitoring and diagnosing the state of the hydrogen fuel cell is needed.
Accordingly, in consideration of the fact that in the process of monitoring the state of the hydrogen fuel cell, the conventional scheme only monitors and analyzes each parameter of the hydrogen fuel cell in the operation process to obtain the diagnosis result of the state of the hydrogen fuel cell, but the scheme does not take the association relation among each parameter of the hydrogen fuel cell into consideration, misjudgment or missed judgment can occur, the accuracy of the diagnosis result of the state of the hydrogen fuel cell is reduced, and a certain potential safety hazard is further provided. Therefore, in the technical scheme of the application, the operation state detection of the hydrogen fuel cell is expected to be performed based on the associated characteristic distribution information of each operation parameter of the hydrogen fuel cell in the operation process, so that the performance of the cell is optimized, and accidents are avoided. However, since the correlation characteristic information of each operation parameter exists not only between each operation parameter but also in the time dimension, in this process, there is a difficulty in how to accurately extract the global time-series correlation characteristic of each operation parameter of the hydrogen fuel cell during operation, so as to improve the accuracy of the state diagnosis of the hydrogen fuel cell.
In recent years, deep learning and neural networks have been widely used in the fields of computer vision, natural language processing, text signal processing, and the like. In addition, deep learning and neural networks have also shown levels approaching and even exceeding humans in the fields of image classification, object detection, semantic segmentation, text translation, and the like.
The development of deep learning and neural networks provides new solutions and schemes for mining global time sequence associated characteristic information of each operation parameter of the hydrogen fuel cell in the operation process. Those of ordinary skill in the art will appreciate that deep learning based deep neural network models may be adapted with appropriate training strategies, such as by gradient descent back-propagation algorithms, to adjust parameters of the deep neural network model to enable modeling of complex nonlinear correlations between things, which is obviously suitable for modeling and mining global timing-related characteristic information of various operating parameters of the hydrogen fuel cell during operation.
In the above-mentioned state diagnosis system 100 for a hydrogen fuel cell, the operation state monitoring module 110 is configured to obtain operation parameters of a plurality of predetermined time points of the hydrogen fuel cell to be detected during a predetermined period of time in the operation process, where the operation parameters include a hydrogen flow rate, a temperature, a humidity, a discharge voltage, and an output current. And in the process of acquiring the operation parameters, respectively acquiring corresponding operation parameter data by using a flow sensor, a temperature sensor, a humidity sensor, a voltmeter and an ammeter.
In the above-described hydrogen fuel cell condition diagnosing system 100, the operation parameter context encoding module 120 is configured to pass the operation parameters at the respective predetermined time points through a context encoder based on a converter to obtain a plurality of context operation parameter semantic understanding feature vectors. In order to fully extract the correlation characteristic distribution information among the operation parameters, in the technical scheme of the application, the operation parameters at each preset time point are encoded in a context encoder based on a converter, and the global context correlation characteristic distribution information about the operation parameters at each preset time point is extracted, so that a plurality of context operation parameter semantic understanding characteristic vectors are obtained. That is, based on the transform concept, the converter is used to capture the long-distance context-dependent characteristic, and the global context-based semantic coding is performed on each operation parameter at each predetermined time point to obtain a context semantic association feature representation, i.e. the context operation parameter semantic understanding feature vector, with the overall semantic association of each operation parameter at each predetermined time point as a context background. It should be appreciated that, in the technical solution of the present application, the respective operation parameters at the respective predetermined time points may be captured by the context encoder based on the converter, which is based on the global context semantic association feature representation.
Specifically, in the embodiment of the present application, the encoding process of the operation parameter context encoding module 120 includes: firstly, inputting operation parameters of a single preset time point into the context encoder based on the converter through an encoding unit to obtain a plurality of single-point operation parameter feature vectors; and then, cascading the plurality of single-point operation parameter feature vectors through a cascading unit to obtain the context operation parameter semantic understanding feature vector.
More specifically, in the embodiment of the present application, the encoding process of the encoding unit is: firstly, arranging the operation parameters of the single preset time point into an input vector; then, the input vector is respectively converted into a query vector and a key vector through a learning embedding matrix; then, calculating the product between the query vector and the transpose vector of the key vector to obtain a self-attention correlation matrix; then, carrying out standardization processing on the self-attention association matrix to obtain a standardized self-attention association matrix; subsequently, the standardized self-attention association matrix is input into a Softmax activation function to be activated so as to obtain a self-attention feature matrix; and finally, multiplying the self-attention characteristic matrix by each operation parameter in the operation parameters at the single preset time point as a value vector to obtain a plurality of single-point operation parameter characteristic vectors.
In the above-mentioned hydrogen fuel cell condition diagnosis system 100, the integrating module 130 and the spatial feature extracting module 140 are configured to two-dimensionally arrange the plurality of context operation parameter semantic understanding feature vectors into a full-time-sequence context operation parameter semantic understanding matrix, and obtain the full-time-sequence context operation parameter semantic understanding feature matrix by using a convolutional neural network model of a spatial attention mechanism. Considering that the operation parameters of the hydrogen fuel cell to be detected in the operation process have semantic relations of contexts among the parameters, the parameters also have dynamic association characteristic information in the time dimension, that is, the operation parameters have dynamic change rules in the time dimension. Therefore, in order to fully and accurately mine global time sequence related characteristics of each operation parameter of the hydrogen fuel cell in the operation process, after the context operation parameter semantic understanding characteristic vectors are two-dimensionally arranged into a full time sequence context operation parameter semantic understanding matrix, a convolution neural network model with excellent performance in terms of implicit characteristic extraction is used for carrying out characteristic mining of the full time sequence context operation parameter semantic understanding matrix.
In particular, in consideration of the fact that since the hydrogen fuel cell supplies hydrogen and oxygen to the anode and the cathode, respectively, in which the hydrogen diffuses outwardly through the anode and reacts with the electrolyte, electrons are discharged to the cathode through an external load, if leakage of hydrogen gas or degradation of performance occurs, the amount of generated electricity of the hydrogen fuel cell is reduced, variations in spatial positions with respect to the respective operation parameters in the anode and the cathode should be more focused in the actual process of detecting the operation state of the hydrogen fuel cell. In view of the ability of the attention mechanism to select the focus position, a more resolved representation of the feature is produced, and the feature after addition to the attention module will change adaptively as the network deepens. Therefore, in the technical scheme of the application, the full-time sequence context operation parameter semantic understanding matrix is processed in a convolutional neural network model by using a spatial attention mechanism to extract global time sequence dynamic associated feature information which is focused on the spatial positions of the anode and the cathode of the hydrogen fuel cell and is based on the operation parameters in the time dimension, so that the full-time sequence context operation parameter semantic understanding feature matrix is obtained. It should be noted that, here, the image features extracted by the spatial attention reflect the weights of the differences of the spatial dimension features, so as to suppress or strengthen the features of different spatial positions, thereby extracting the time sequence dynamic associated feature information of the operation parameters which are spatially focused on the anode and the cathode of the hydrogen fuel cell.
Specifically, in the embodiment of the present application, the encoding process of the spatial feature extraction module 140 includes: firstly, performing depth convolution coding on the full-time sequence context operation parameter semantic understanding matrix by using a convolution coding part of the convolution neural network model to obtain an initial convolution characteristic diagram; then, inputting the initial convolution feature map into a spatial attention portion of the convolution neural network model to obtain a spatial attention map; then, the spatial attention is sought to be activated by Softmax to obtain a spatial attention profile; then, the space attention feature map and the initial convolution feature map are calculated and multiplied by position points to obtain a full-time sequence context operation parameter semantic understanding feature map; and finally, carrying out global average pooling processing along the channel dimension on the full-time sequence context operation parameter semantic understanding feature map to obtain the full-time sequence context operation parameter semantic understanding feature matrix.
In the above-mentioned hydrogen fuel cell condition diagnosis system 100, the distinguishing and optimizing module 150 is configured to distinguish the feature values of the full-time-sequence context operation parameter semantic understanding feature matrix to obtain an optimized full-time-sequence context operation parameter semantic understanding feature matrix. In particular, in the technical solution of the present application, for the full-time sequence context operation parameter semantic understanding feature matrix obtained by using the convolutional neural network model of the spatial attention mechanism, since the spatial attention mechanism strengthens the extraction of the associated features on the predetermined spatial position of the full-time sequence context operation parameter semantic understanding feature matrix, the feature values of some positions of the full-time sequence context operation parameter semantic understanding feature matrix have more significant importance with respect to the feature values of other positions, so if the feature values of the full-time sequence context operation parameter semantic understanding feature matrix can be effectively distinguished in the classification task, the training speed of the model and the accuracy of the classification result can be obviously improved.
Thus, the applicant of the present application semantically understands feature matrices for the full-time contextual operating parameters, e.g., noted asInteractive augmentation based on distinguishable physical stimulus is performed, expressed as: />
Wherein the method comprises the steps ofIs the semantic understanding feature matrix of the full-time sequence context operation parameters, and is->And->Is a predetermined superparameter,/->And->Representing the addition and subtraction of the feature matrix by position, the division representing each position of the feature matrix divided by the corresponding value, andrepresenting convolution operations through a single convolution layer, +.>Is the optimized full-time sequence contextOperational parameter semantic understanding feature matrix,>and->Is an intermediate matrix.
Here, the discriminative physical stimulus-based interaction enhancement is used to promote interactions between feature space and solution space of classification problems during back propagation through gradient descent, which extracts and mimics viable features (actionable feature) in a physical stimulus-like manner, whereby a general purpose low-dimensional guided physical stimulus approach is used to obtain a physical representation of viable features with gradient discriminativity, thereby enhancing the full-time context operational parameter semantic understanding feature matrix during training Active part in order to promote the semantic understanding feature matrix of the optimized full time sequence context operation parameters +.>The training speed under the classification task and the accuracy of the classification result of the semantic understanding characteristics of the full-time sequence context operation parameters after training. Therefore, the operation state of the hydrogen fuel cell can be accurately monitored and diagnosed in real time, so that the performance of the cell is optimized, and accidents are avoided.
In the above-mentioned state diagnosis system 100 for a hydrogen fuel cell, the diagnosis result generating module 160 is configured to pass the optimized full-time-sequence context operation parameter semantic understanding feature matrix through a classifier to obtain a classification result, where the classification result is used to indicate whether the operation state of the hydrogen fuel cell to be detected is normal. That is, in the technical solution of the present application, the label of the classifier includes that the operation state of the hydrogen fuel cell to be detected is normal, and the operation state of the hydrogen fuel cell to be detected is abnormal, where the classifier determines, through a soft maximum function, which classification label the classification feature matrix belongs to, so as to accurately detect and diagnose the operation state of the hydrogen fuel cell to be detected, thereby optimizing the performance of the battery and avoiding occurrence of accidents.
Fig. 4 is a block diagram of a diagnostic result generation module in a status diagnostic system of a hydrogen fuel cell according to an embodiment of the present application. As shown in fig. 4, the diagnostic result generation module 160 includes: a developing unit 161, configured to develop the optimized full-time sequence context operation parameter semantic understanding feature matrix into a classification feature vector according to a row vector or a column vector; a probability unit 162, configured to input the classification feature vector into a Softmax classification function of the classifier to obtain a probability value of the classification feature vector belonging to each classification label; and a classification result generation unit 163 for determining a classification label corresponding to the maximum probability value as the classification result.
In summary, the state diagnosis system 100 of the hydrogen fuel cell according to the embodiment of the present application is illustrated, which adopts the artificial intelligence monitoring and diagnosis technology based on deep learning to mine the high-dimensional implicit correlation characteristic information included between each operation parameter of the hydrogen fuel cell in the operation process through the context encoder based on the converter, and uses the spatial attention mechanism to enhance the time sequence change condition of each operation parameter in the anode and the cathode of the cell in the characteristic information, so as to perform the classification processing. Thus, the operation state detection of the hydrogen fuel cell is carried out based on the global time sequence related characteristic distribution information of each operation parameter of the hydrogen fuel cell, so that the performance of the cell is optimized, and accidents are avoided.
As described above, the state diagnostic system 100 of a hydrogen fuel cell according to the embodiment of the present application can be implemented in various terminal devices, such as a server or the like for state diagnosis of a hydrogen fuel cell. In one example, the condition diagnosing system 100 of the hydrogen fuel cell according to the embodiment of the present application may be integrated into the terminal device as one software module and/or hardware module. For example, the condition diagnosing system 100 of the hydrogen fuel cell may be a software module in the operating system of the terminal apparatus, or may be an application developed for the terminal apparatus; of course, the condition diagnosing system 100 of the hydrogen fuel cell may also be one of a plurality of hardware modules of the terminal apparatus.
Alternatively, in another example, the hydrogen fuel cell condition diagnosis system 100 and the terminal device may be separate devices, and the hydrogen fuel cell condition diagnosis system 100 may be connected to the terminal device through a wired and/or wireless network and transmit interactive information in a contracted data format.
Exemplary method
Fig. 5 is a flowchart of a state diagnosis method of a hydrogen fuel cell according to an embodiment of the present application. As shown in fig. 5, the state diagnosis method of the hydrogen fuel cell according to the embodiment of the present application includes: s110, acquiring operation parameters of a to-be-detected hydrogen fuel cell at a plurality of preset time points in a preset time period in the operation process, wherein the operation parameters comprise hydrogen flow, temperature, humidity, discharge voltage and output current; s120, enabling the operation parameters of each preset time point to pass through a context encoder based on a converter to obtain a plurality of context operation parameter semantic understanding feature vectors; s130, two-dimensionally arranging the context operation parameter semantic understanding feature vectors into a full-time sequence context operation parameter semantic understanding matrix; s140, the full-time sequence context operation parameter semantic understanding matrix is obtained through a convolution neural network model using a spatial attention mechanism; s150, distinguishing characteristic values of the semantic understanding characteristic matrix of the full-time sequence context operation parameters to obtain an optimized semantic understanding characteristic matrix of the full-time sequence context operation parameters; and S160, enabling the optimized full-time sequence context operation parameter semantic understanding feature matrix to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the operation state of the hydrogen fuel cell to be detected is normal or not.
Here, it will be understood by those skilled in the art that the respective steps and operations in the above-described state diagnosis method of the hydrogen fuel cell have been described in detail in the above description of the state diagnosis system 100 of the hydrogen fuel cell with reference to fig. 1 to 4, and thus, repetitive descriptions thereof will be omitted.
Exemplary electronic device
Next, an electronic device according to an embodiment of the present application is described with reference to fig. 6. Fig. 6 is a block diagram of an electronic device according to an embodiment of the present application. As shown in fig. 6, the electronic device 10 includes one or more processors 11 and a memory 12.
The processor 11 may be a Central Processing Unit (CPU) or other form of processing unit having data processing and/or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.
In one example, the electronic device 10 may further include: an input device 13 and an output device 14, which are interconnected by a bus system and/or other forms of connection mechanisms (not shown).
The input means 13 may comprise, for example, a keyboard, a mouse, etc.
The output device 14 may output various information including the classification result and the like to the outside. The output means 14 may include, for example, a display, speakers, a printer, and a communication network and remote output devices connected thereto, etc.
Of course, only some of the components of the electronic device 10 that are relevant to the present application are shown in fig. 6 for simplicity, components such as buses, input/output interfaces, etc. are omitted. In addition, the electronic device 10 may include any other suitable components depending on the particular application.
Exemplary computer program product and computer readable storage Medium
In addition to the methods and apparatus described above, embodiments of the present application may also be a computer program product comprising computer program instructions which, when executed by a processor, cause the processor to perform steps in the functions of the state diagnostic method of a hydrogen fuel cell according to various embodiments of the present application described in the "exemplary methods" section of the present specification.
The computer program product may write program code for performing the operations of embodiments of the present application in any combination of one or more programming languages, including an object oriented programming language such as Java, C++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device, partly on a remote computing device, or entirely on the remote computing device or server.
Furthermore, embodiments of the present application may also be a computer-readable storage medium, on which computer program instructions are stored, which, when being executed by a processor, cause the processor to perform steps in the functions in the status diagnosing method of a hydrogen fuel cell according to various embodiments of the present application described in the above-described "exemplary method" section of the present specification.
The computer readable storage medium may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may include, for example, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium would include the following: an electrical connection having one or more wires, a portable disk, a hard disk, random Access Memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
The basic principles of the present application have been described above in connection with specific embodiments, however, it should be noted that the advantages, benefits, effects, etc. mentioned in the present application are merely examples and not limiting, and these advantages, benefits, effects, etc. are not to be considered as necessarily possessed by the various embodiments of the present application. Furthermore, the specific details disclosed herein are for purposes of illustration and understanding only, and are not intended to be limiting, as the application is not intended to be limited to the details disclosed herein as such.
The block diagrams of the devices, apparatuses, devices, systems referred to in this application are only illustrative examples and are not intended to require or imply that the connections, arrangements, configurations must be made in the manner shown in the block diagrams. As will be appreciated by one of skill in the art, the devices, apparatuses, devices, systems may be connected, arranged, configured in any manner. Words such as "including," "comprising," "having," and the like are words of openness and mean "including but not limited to," and are used interchangeably therewith. The terms "or" and "as used herein refer to and are used interchangeably with the term" and/or "unless the context clearly indicates otherwise. The term "such as" as used herein refers to, and is used interchangeably with, the phrase "such as, but not limited to.
It is also noted that in the apparatus, devices and methods of the present application, the components or steps may be disassembled and/or assembled. Such decomposition and/or recombination should be considered as equivalent to the present application.
The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
The foregoing description has been presented for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of the application to the form disclosed herein. Although a number of example aspects and embodiments have been discussed above, a person of ordinary skill in the art will recognize certain variations, modifications, alterations, additions, and subcombinations thereof.
Claims (10)
1. A condition diagnosing system of a hydrogen fuel cell, comprising:
the operation state monitoring module is used for acquiring operation parameters of a plurality of preset time points of the hydrogen fuel cell to be detected in a preset time period in the operation process, wherein the operation parameters comprise hydrogen flow, temperature, humidity, discharge voltage and output current;
An operation parameter context coding module, configured to pass the operation parameters of the respective predetermined time points through a context encoder based on a converter to obtain a plurality of context operation parameter semantic understanding feature vectors;
the integration module is used for two-dimensionally arranging the context operation parameter semantic understanding feature vectors into a full-time sequence context operation parameter semantic understanding matrix;
the spatial feature extraction module is used for enabling the full-time sequence context operation parameter semantic understanding matrix to obtain a full-time sequence context operation parameter semantic understanding feature matrix through a convolutional neural network model using a spatial attention mechanism;
the distinguishing and optimizing module is used for distinguishing the characteristic values of the semantic understanding characteristic matrix of the full-time sequence context operation parameters so as to obtain the semantic understanding characteristic matrix of the optimized full-time sequence context operation parameters; and
the diagnosis result generation module is used for enabling the semantic understanding feature matrix of the optimized full-time sequence context operation parameters to pass through a classifier to obtain a classification result, and the classification result is used for indicating whether the operation state of the hydrogen fuel cell to be detected is normal or not.
2. The hydrogen fuel cell condition diagnostic system of claim 1, wherein the operating parameter context encoding module comprises:
An encoding unit for inputting an operating parameter at a single predetermined point in time into the converter-based context encoder to obtain the plurality of single-point operating parameter feature vectors; and
and the cascading unit is used for cascading the plurality of single-point operation parameter feature vectors to obtain the context operation parameter semantic understanding feature vector.
3. The condition diagnosing system of a hydrogen fuel cell according to claim 2, wherein the encoding unit is further configured to:
arranging the operation parameters of the single preset time point into an input vector;
respectively converting the input vector into a query vector and a key vector through a learning embedding matrix;
calculating the product between the query vector and the transpose vector of the key vector to obtain a self-attention correlation matrix;
carrying out standardization processing on the self-attention association matrix to obtain a standardized self-attention association matrix;
inputting the standardized self-attention association matrix into a Softmax activation function to activate so as to obtain a self-attention feature matrix; and
and multiplying the self-attention characteristic matrix with each operation parameter in the operation parameters at the single preset time point as a value vector to obtain a plurality of single-point operation parameter characteristic vectors.
4. The condition diagnosing system of a hydrogen fuel cell according to claim 3, wherein the spatial feature extraction module is further configured to:
performing depth convolution coding on the full-time sequence context operation parameter semantic understanding matrix by using a convolution coding part of the convolution neural network model to obtain an initial convolution feature map;
inputting the initial convolution feature map into a spatial attention portion of the convolution neural network model to obtain a spatial attention map;
-passing said spatial attention map through a Softmax activation function to obtain a spatial attention profile;
calculating the position-based points of the space attention feature map and the initial convolution feature map to obtain a full-time sequence context operation parameter semantic understanding feature map; and
and carrying out global averaging treatment along the channel dimension on the full-time sequence context operation parameter semantic understanding feature map to obtain the full-time sequence context operation parameter semantic understanding feature matrix.
5. The hydrogen fuel cell condition diagnosing system according to claim 4, wherein the distinction optimizing module is further configured to:
performing interactive reinforcement based on distinguishable physical excitation on the full-time sequence context operation parameter semantic understanding feature matrix by using the following formula to obtain the optimized full-time sequence context operation parameter semantic understanding feature matrix;
Wherein, the formula is:
wherein the method comprises the steps ofIs the semantic understanding feature matrix of the full-time sequence context operation parameters, and is->And->Is a predetermined superparameter,/->And->Representing the addition and subtraction of the feature matrix by position, division representing each position of the feature matrix divided by the corresponding value, and +.>Representing convolution operations through a single convolution layer, +.>Is the semantic understanding feature matrix of the optimized full-time sequence context operation parameters, and is +.>And->Is an intermediate matrix.
6. The state diagnostic system of a hydrogen fuel cell according to claim 5, wherein the diagnostic result generation module includes:
the expansion unit is used for expanding the semantic understanding feature matrix of the optimized full-time sequence context operation parameter into classification feature vectors according to row vectors or column vectors;
the probability unit is used for inputting the classification feature vector into a Softmax classification function of the classifier to obtain a probability value of the classification feature vector belonging to each classification label; and
and the classification result generation unit is used for determining the classification label corresponding to the maximum probability value as the classification result.
7. A condition diagnosing method of a hydrogen fuel cell, characterized by comprising:
Acquiring operation parameters of a hydrogen fuel cell to be detected at a plurality of preset time points in a preset time period in the operation process, wherein the operation parameters comprise hydrogen flow, temperature, humidity, discharge voltage and output current;
passing the operating parameters at each predetermined point in time through a context encoder based on a converter to obtain a plurality of context operating parameter semantic understanding feature vectors;
two-dimensionally arranging the plurality of context operation parameter semantic understanding feature vectors into a full-time sequence context operation parameter semantic understanding matrix;
the full-time sequence context operation parameter semantic understanding matrix is obtained through a convolutional neural network model using a spatial attention mechanism;
performing eigenvalue distinction on the full-time sequence context operation parameter semantic understanding feature matrix to obtain an optimized full-time sequence context operation parameter semantic understanding feature matrix; and
and the semantic understanding feature matrix of the optimized full-time sequence context operation parameters is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the operation state of the hydrogen fuel cell to be detected is normal or not.
8. The method according to claim 7, wherein passing the operation parameters at the respective predetermined time points through a context encoder based on a converter to obtain a plurality of context operation parameter semantic understanding feature vectors, comprises:
Inputting operating parameters at a single predetermined point in time into the converter-based context encoder to obtain the plurality of single-point operating parameter feature vectors; and
and cascading the plurality of single-point operation parameter feature vectors to obtain the context operation parameter semantic understanding feature vector.
9. The method of diagnosing a condition of a hydrogen fuel cell according to claim 8, wherein inputting the operating parameters at a single predetermined point in time into the converter-based context encoder to obtain the plurality of single-point operating parameter feature vectors, comprises:
arranging the operation parameters of the single preset time point into an input vector;
respectively converting the input vector into a query vector and a key vector through a learning embedding matrix;
calculating the product between the query vector and the transpose vector of the key vector to obtain a self-attention correlation matrix;
carrying out standardization processing on the self-attention association matrix to obtain a standardized self-attention association matrix;
inputting the standardized self-attention association matrix into a Softmax activation function to activate so as to obtain a self-attention feature matrix; and
and multiplying the self-attention characteristic matrix with each operation parameter in the operation parameters at the single preset time point as a value vector to obtain a plurality of single-point operation parameter characteristic vectors.
10. The method of diagnosing a state of a hydrogen fuel cell according to claim 9, wherein passing the full-time-series context operation parameter semantic understanding matrix through a convolutional neural network model using a spatial attention mechanism to obtain a full-time-series context operation parameter semantic understanding feature matrix, comprises:
performing depth convolution coding on the full-time sequence context operation parameter semantic understanding matrix by using a convolution coding part of the convolution neural network model to obtain an initial convolution feature map;
inputting the initial convolution feature map into a spatial attention portion of the convolution neural network model to obtain a spatial attention map;
-passing said spatial attention map through a Softmax activation function to obtain a spatial attention profile;
calculating the position-based points of the space attention feature map and the initial convolution feature map to obtain a full-time sequence context operation parameter semantic understanding feature map; and
and carrying out global averaging treatment along the channel dimension on the full-time sequence context operation parameter semantic understanding feature map to obtain the full-time sequence context operation parameter semantic understanding feature matrix.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310376056.6A CN116130721B (en) | 2023-04-11 | 2023-04-11 | Status diagnostic system and method for hydrogen fuel cell |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310376056.6A CN116130721B (en) | 2023-04-11 | 2023-04-11 | Status diagnostic system and method for hydrogen fuel cell |
Publications (2)
Publication Number | Publication Date |
---|---|
CN116130721A true CN116130721A (en) | 2023-05-16 |
CN116130721B CN116130721B (en) | 2023-06-16 |
Family
ID=86299391
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310376056.6A Active CN116130721B (en) | 2023-04-11 | 2023-04-11 | Status diagnostic system and method for hydrogen fuel cell |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116130721B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116309446A (en) * | 2023-03-14 | 2023-06-23 | 浙江固驰电子有限公司 | Method and system for manufacturing power module for industrial control field |
CN116404212A (en) * | 2023-05-22 | 2023-07-07 | 中国电建集团江西省电力建设有限公司 | Capacity equalization control method and system for zinc-iron flow battery system |
CN117148165A (en) * | 2023-09-15 | 2023-12-01 | 东莞市言科新能源有限公司 | Testing and analyzing method and system for polymer lithium ion battery |
CN118098482A (en) * | 2024-04-22 | 2024-05-28 | 吉林大学 | Intelligent medical management system and method based on 5G technology |
CN118098482B (en) * | 2024-04-22 | 2024-06-28 | 吉林大学 | Intelligent medical management system and method based on 5G technology |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20200142421A1 (en) * | 2018-11-05 | 2020-05-07 | GM Global Technology Operations LLC | Method and system for end-to-end learning of control commands for autonomous vehicle |
WO2021216313A1 (en) * | 2020-04-23 | 2021-10-28 | Zoox, Inc. | Vehicle health monitor |
CN114566190A (en) * | 2022-02-28 | 2022-05-31 | 上海亭章通信技术有限公司 | Early warning method and system based on voice characteristics and electronic equipment thereof |
CN114647198A (en) * | 2022-03-09 | 2022-06-21 | 深圳市经纬纵横科技有限公司 | Intelligent home control method and system based on Internet of things and electronic equipment |
CN115098999A (en) * | 2022-05-25 | 2022-09-23 | 同济大学 | Multi-mode fusion fuel cell system performance attenuation prediction method |
CN115147655A (en) * | 2022-07-12 | 2022-10-04 | 温州宁酷科技有限公司 | Oil gas gathering and transportation monitoring system and method thereof |
CN115291646A (en) * | 2022-07-08 | 2022-11-04 | 福建龙氟化工有限公司 | Energy management control system for lithium fluoride preparation and control method thereof |
CN115392370A (en) * | 2022-08-24 | 2022-11-25 | 诸暨市萤朵贸易有限公司 | Fault diagnosis method and system for thermal power generation equipment |
CN115693918A (en) * | 2022-09-07 | 2023-02-03 | 浙江心友机电设备安装有限公司 | Comprehensive intelligent power utilization system and method for building |
CN115797708A (en) * | 2023-02-06 | 2023-03-14 | 南京博纳威电子科技有限公司 | Power transmission and distribution synchronous data acquisition method |
-
2023
- 2023-04-11 CN CN202310376056.6A patent/CN116130721B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20200142421A1 (en) * | 2018-11-05 | 2020-05-07 | GM Global Technology Operations LLC | Method and system for end-to-end learning of control commands for autonomous vehicle |
WO2021216313A1 (en) * | 2020-04-23 | 2021-10-28 | Zoox, Inc. | Vehicle health monitor |
CN114566190A (en) * | 2022-02-28 | 2022-05-31 | 上海亭章通信技术有限公司 | Early warning method and system based on voice characteristics and electronic equipment thereof |
CN114647198A (en) * | 2022-03-09 | 2022-06-21 | 深圳市经纬纵横科技有限公司 | Intelligent home control method and system based on Internet of things and electronic equipment |
CN115098999A (en) * | 2022-05-25 | 2022-09-23 | 同济大学 | Multi-mode fusion fuel cell system performance attenuation prediction method |
CN115291646A (en) * | 2022-07-08 | 2022-11-04 | 福建龙氟化工有限公司 | Energy management control system for lithium fluoride preparation and control method thereof |
CN115147655A (en) * | 2022-07-12 | 2022-10-04 | 温州宁酷科技有限公司 | Oil gas gathering and transportation monitoring system and method thereof |
CN115392370A (en) * | 2022-08-24 | 2022-11-25 | 诸暨市萤朵贸易有限公司 | Fault diagnosis method and system for thermal power generation equipment |
CN115693918A (en) * | 2022-09-07 | 2023-02-03 | 浙江心友机电设备安装有限公司 | Comprehensive intelligent power utilization system and method for building |
CN115797708A (en) * | 2023-02-06 | 2023-03-14 | 南京博纳威电子科技有限公司 | Power transmission and distribution synchronous data acquisition method |
Non-Patent Citations (2)
Title |
---|
BEI SUN等: "Short-term performance degradation prediction of a commercial vehicle fuel cell system based on CNN and LSTM hybrid neural network", SCIENCEDIRECT, vol. 18, no. 23, pages 8613 - 8628 * |
何礼明: "基于机器学习模型融合的SOFC故障诊断", 中国优秀硕士学位论文全文数据库工程科技Ⅰ辑, no. 3, pages 015 - 303 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116309446A (en) * | 2023-03-14 | 2023-06-23 | 浙江固驰电子有限公司 | Method and system for manufacturing power module for industrial control field |
CN116404212A (en) * | 2023-05-22 | 2023-07-07 | 中国电建集团江西省电力建设有限公司 | Capacity equalization control method and system for zinc-iron flow battery system |
CN116404212B (en) * | 2023-05-22 | 2024-02-27 | 中国电建集团江西省电力建设有限公司 | Capacity equalization control method and system for zinc-iron flow battery system |
CN117148165A (en) * | 2023-09-15 | 2023-12-01 | 东莞市言科新能源有限公司 | Testing and analyzing method and system for polymer lithium ion battery |
CN117148165B (en) * | 2023-09-15 | 2024-04-12 | 东莞市言科新能源有限公司 | Testing and analyzing method and system for polymer lithium ion battery |
CN118098482A (en) * | 2024-04-22 | 2024-05-28 | 吉林大学 | Intelligent medical management system and method based on 5G technology |
CN118098482B (en) * | 2024-04-22 | 2024-06-28 | 吉林大学 | Intelligent medical management system and method based on 5G technology |
Also Published As
Publication number | Publication date |
---|---|
CN116130721B (en) | 2023-06-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN116130721B (en) | Status diagnostic system and method for hydrogen fuel cell | |
CN115796173B (en) | Data processing method and system for supervising reporting requirements | |
Wang et al. | Fault identification and diagnosis based on KPCA and similarity clustering for nuclear power plants | |
CN116247824B (en) | Control method and system for power equipment | |
CN115834433B (en) | Data processing method and system based on Internet of things technology | |
CN116257406A (en) | Gateway data management method and system for smart city | |
CN115859437A (en) | Jacket underwater stress detection system based on distributed optical fiber sensing system | |
CN116579618B (en) | Data processing method, device, equipment and storage medium based on risk management | |
CN116055293B (en) | Remote fault monitoring method of router and router | |
CN116307624A (en) | Resource scheduling method and system of ERP system | |
CN117077075A (en) | Water quality monitoring system and method for environmental protection | |
CN117248967A (en) | Mine safety monitoring system and method based on intelligent perception | |
CN116086790A (en) | Performance detection method and system for high-pressure valve of hydrogen fuel cell | |
CN116819379A (en) | Fault monitoring system and method for power adapter | |
CN116759053A (en) | Medical system prevention and control method and system based on Internet of things system | |
CN116309580A (en) | Oil and gas pipeline corrosion detection method based on magnetic stress | |
Telford et al. | Fault classification and diagnostic system for unmanned aerial vehicle electrical networks based on hidden Markov models | |
CN116404212B (en) | Capacity equalization control method and system for zinc-iron flow battery system | |
CN117231590A (en) | Fault prediction system and method for hydraulic system | |
CN117316462A (en) | Medical data management method | |
CN116739219A (en) | Melt blown cloth production management system and method thereof | |
Li et al. | Open set recognition fault diagnosis framework based on convolutional prototype learning network for nuclear power plants | |
CN116714437A (en) | Hydrogen fuel cell automobile safety monitoring system and monitoring method based on big data | |
Tsalapati et al. | Enhancing polymer electrolyte membrane fuel cell system diagnostics through semantic modelling | |
CN115293265A (en) | Non-invasive monitoring method and system for building electricity utilization |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |