Abnormal signal processing method and device, storage medium and electronic equipment

Abstract

The invention discloses an abnormal signal processing method and device, a storage medium and electronic equipment. The method comprises the steps that an original signal is used for representing the operation state of a target data source; de-noising processing is carried out on the original signals, and noise signals in the de-noised original signals are less than noise signals in the original signals before de-noising; the denoised original signals are input into an abnormal signal recognition model of the target data source for analysis, abnormal signals of the target data source are obtained, and the abnormal signal recognition model is established through denoised historical original signals of the target data source; and executing an exception handling strategy of the target data source on the exception signal, the exception handling strategy being used for representing a rule for performing exception handling on the exception signal. The technical problem that the abnormal signal processing efficiency is low due to the fact that the abnormal signal cannot be accurately recognized is solved.

Description

Technical Field

[0001] The present invention relates to the field of signal processing technology, and in particular to a method, device, storage medium and electronic device for processing abnormal signals. Background Art

[0002] During industrial production and equipment operation, various abnormal signals often appear. Timely identification of these abnormal signals and rapid processing of them are crucial to ensuring the stability and reliability of the production system.

[0003] At present, manually set rules are often used to identify abnormal signals generated by production systems, and the abnormal signals obtained above are further processed. However, manually set rules are prone to omissions or delays, which leads to technical problems such as low efficiency in processing abnormal signals due to the inability to accurately identify abnormal signals.

[0004] With respect to the above-mentioned technical problem of low efficiency in processing abnormal signals due to the inability to accurately identify abnormal signals, no effective solution has been proposed so far. Summary of the invention

[0005] The embodiments of the present invention provide a method, device, storage medium and electronic device for processing abnormal signals, so as to at least solve the technical problem of low efficiency in processing abnormal signals due to inability to accurately identify abnormal signals.

[0006] According to one aspect of the embodiments of the invention, a method for processing abnormal signals is provided. The method includes: the original signal is used to characterize the operating state of the target data source; the original signal is denoised, wherein the noise signal in the original signal after denoising is less than the noise signal in the original signal before denoising; the original signal after denoising is input into the abnormal signal recognition model of the target data source for analysis to obtain the abnormal signal of the target data source, wherein the abnormal signal recognition model is established by the historical original signal after denoising of the target data source; the abnormal signal processing strategy of the target data source is executed on the abnormal signal, wherein the abnormal processing strategy is used to characterize the rules for abnormal processing of the abnormal signal.

[0007] Optionally, the denoised original signal is input into an abnormal signal recognition model of a target data source for analysis to obtain an abnormal signal of the target data source, including: obtaining an abnormal pattern signal corresponding to the denoised original signal; using the abnormal signal recognition model to perform feature extraction on the abnormal pattern signal and the denoised original signal, respectively, to obtain a target abnormal pattern signal corresponding to the abnormal pattern signal, and a target signal corresponding to the denoised original signal; and determining the abnormal signal based on the target abnormal pattern signal and the target signal.

[0008] Optionally, determining the abnormal signal based on the target abnormal pattern signal and the target signal includes: using an abnormal signal recognition model to obtain the similarity between the target abnormal pattern signal and the target signal; and in response to the similarity being the same as a similarity threshold, determining the target signal as an abnormal signal.

[0009] Optionally, executing an exception handling strategy of a target data source on an abnormal signal includes: classifying the abnormal signal according to a target abnormal classification rule to obtain an abnormality level of the abnormal signal, wherein the abnormality level is used to characterize the severity of the abnormality generated by the data source corresponding to the abnormal signal; and executing an exception handling strategy on the abnormal signal based on the abnormality level.

[0010] Optionally, based on the exception level, an exception handling strategy is executed on the exception signal, including: in response to the exception level being a first exception level, a first exception handling strategy of the exception handling strategy is executed on the exception signal, wherein the first exception handling strategy is used to start a backup device to put the target production system corresponding to the target data source into a normal operating state; in response to the exception level being a second exception level, a second exception handling strategy of the exception handling strategy is executed on the exception signal, wherein the severity of the second exception level is higher than the severity of the first exception level, and the second exception handling strategy is used to put the target production system into a normal operating state by switching the first production line corresponding to the generated exception signal to the second production line; in response to the exception level being a third exception level, a third exception handling strategy of the exception handling strategy is executed on the exception signal, wherein the severity of the third exception level is higher than the severity of the second exception level, and the third exception handling strategy is used to stop the control operation of the target production system.

[0011] Optionally, after executing the exception handling strategy of the target data source on the abnormal signal, the method further includes: obtaining an alarm instruction of the target data source; generating alarm notification information in response to the alarm instruction, wherein the alarm notification information is used to notify the target object that an abnormality has occurred in the target data source.

[0012] Optionally, the target interface includes a first target interface and a second target interface, wherein the first target interface is used to represent a hardware interface of the target data source, and the second target interface is used to represent a software protocol interface corresponding to the target data source.

[0013] According to one aspect of an embodiment of the present invention, a device for processing abnormal signals is provided, and the device also includes: an acquisition unit, which is used to acquire original signals of a target data source through a target interface of at least one target data source, wherein the original signals are used to characterize the operating state of the target data source; a denoising unit, which is used to perform denoising on the original signals, wherein the noise signal in the original signals after denoising is less than the noise signal in the original signals before denoising; an analysis unit, which is used to input the denoised original signals into an abnormal signal recognition model of the target data source for analysis to obtain abnormal signals of the target data source, wherein the abnormal signal recognition model is established by using the historical original signals after denoising of the target data source; and an execution unit, which is used to execute an abnormal processing strategy of the target data source on the abnormal signals, wherein the abnormal processing strategy is used to characterize the rules for performing abnormal processing on the abnormal signals.

[0014] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium includes a stored program, wherein when the program is executed by a processor, the device where the storage medium is located is controlled to execute the method in the embodiment of the present invention.

[0015] According to another aspect of an embodiment of the present invention, a processor is provided. The processor is used to run a program, wherein the method in the embodiment of the present invention is executed when the program is run.

[0016] According to another aspect of the embodiments of the present invention, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the method in the embodiments of the present invention is implemented.

[0017] According to another aspect of an embodiment of the present invention, an electronic device is provided, the electronic device comprising a processor and a memory for storing instructions executable by the processor, wherein the processor is configured to execute the instructions to implement the method in the embodiment of the present invention.

[0018] In an embodiment of the present invention, the original signal of the target data source is collected through the target interface of at least one target data source, wherein the original signal is used to characterize the operating state of the target data source; the original signal is denoised, wherein the noise signal in the denoised original signal is less than the noise signal in the original signal before denoising; the denoised original signal is input into the abnormal signal recognition model of the target data source for analysis to obtain the abnormal signal of the target data source, wherein the abnormal signal recognition model is established by the denoised historical original signal of the target data source; the abnormal signal of the target data source is executed on the abnormal signal, wherein the abnormal signal handling strategy is used to characterize the rule for abnormal signal handling. That is to say, in an embodiment of the present invention, the original signal of the target data source can be collected through the target interface of at least one target data source, and then the original signal obtained above is denoised to obtain the denoised original signal, and then the denoised original signal is input into the abnormal signal recognition model of the target data source for analysis to obtain the abnormal signal of the target data source, and finally the abnormal signal can be executed on the abnormal signal. The abnormal signal handling strategy of the target data source is used to achieve the purpose of effectively processing the abnormal signal. Considering that after obtaining the denoised original signal, the denoised original signal can be input into the abnormal signal recognition model to obtain the abnormal signal, and then the abnormal signal is processed by the abnormal processing strategy to ensure the security of the target data source, so as to solve the technical problem of low efficiency in processing abnormal signals due to the inability to accurately identify abnormal signals, and achieve the technical effect of accurately identifying abnormal signals and improving the processing efficiency of abnormal signals. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0020] Figure 1 is a flow chart of a method for processing an abnormal signal according to an embodiment of the present invention;

[0021] Figure 2 is a flow chart of a method for automatically identifying and processing abnormal signals according to an embodiment of the present invention;

[0022] Figure 3 is a schematic diagram of a device for processing abnormal signals according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0025] According to an embodiment of the present invention, a method for processing abnormal signals is provided. It should be noted that in the flowchart of the accompanying drawings, the steps shown therein can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0026] The following is an introduction to a method for processing abnormal signals according to an embodiment of the present invention.

[0027] Figure 1 is a flow chart of a method for processing an abnormal signal according to an embodiment of the present invention. Figure 1 As shown, the method for processing the abnormal signal may include the following steps:

[0028] Step S101: collecting an original signal of a target data source through a target interface of at least one target data source.

[0029] In the technical solution provided in the above step S101 of the present invention, the original signal of the target data source can be collected through at least one target interface of the target data source, wherein the original signal is used to characterize the operating state of the target data source.

[0030] Optionally, the target data source is used to represent a variety of hardware devices that can collect original signals, such as various sensors, various monitoring devices, etc. The target interface may be a preset interface, which may be referred to as a preset data interface.

[0031] For example, signal data is collected from various data sources in real time or periodically through a preset data interface, where the data sources include sensors and monitoring equipment.

[0032] It should be noted that this is only a preferred implementation method for collecting the original signal of the target data source, and does not specifically limit the process and method of collecting the original signal of the target data source. As long as it is through the target interface of at least one target data source, the process and method of collecting the original signal of the target data source are within the protection scope of the present invention and are not listed here.

[0033] Step S102, performing denoising processing on the original signal.

[0034] In the technical solution provided in the above step S102 of the present invention, after the original signal is obtained, the original signal can be subjected to a gold-top denoising process, wherein the noise signal in the original signal after denoising is less than the noise signal in the original signal before denoising.

[0035] Optionally, after obtaining the original signal, it is necessary to clean the original signal to remove invalid or redundant information to obtain a cleaned original signal, and then perform denoising on the cleaned original signal to obtain a denoised original signal.

[0036] Optionally, the denoising process for the original signal may include: using filters, Fourier transform, and wavelet transform, etc. Among them, using filters is used to represent the removal of noise in the signal by using low-pass filters, medium-pass filters or high-pass filters. Fourier transform is used to represent that the signal can be converted into the frequency domain for processing to remove noise with a frequency higher than the signal frequency. Wavelet transform is used to represent the conversion of the signal into the wavelet domain for processing.

[0037] It should be noted that the denoising method of the original signal is only illustrated here, and the denoising method of the original signal is not specifically limited.

[0038] Step S103: input the denoised original signal into the abnormal signal recognition model of the target data source for analysis to obtain the abnormal signal of the target data source.

[0039] In the technical solution provided in the above step S103 of the present invention, the denoised original signal obtained through the above step is input into the abnormal signal recognition model for analysis, so as to achieve the purpose of obtaining the abnormal signal of the target data source, wherein the abnormal signal recognition model is established by the denoised historical original signal of the target data source, and the model is a trained model.

[0040] Optionally, an abnormal signal recognition model is used to perform feature extraction on the denoised original signal to obtain a feature vector of the denoised original signal, and based on the feature vector, the abnormal signal of the target data source is determined, wherein the feature extraction includes a first feature extraction, a second feature extraction and a third feature extraction, the first feature extraction is used to perform feature extraction on numerical data in the denoised original signal, the second feature extraction is used to perform feature extraction on categorical data in the denoised original signal, and the third feature extraction is used to perform feature extraction on text data in the denoised original signal.

[0041] Optionally, the first feature extraction may include the following processing methods: missing value processing, which is used to process missing values ​​by filling (such as mean filling, median filling, mode filling, etc.), deleting or interpolating, so as to ensure the integrity of the data; standardization and normalization processing, which is used to convert numerical features into a form with a mean of 0 and a variance of 1, or a value range of [0,1], so that the model can better understand and process it; discretization processing, which is used to discretize continuous features into multiple intervals or bins for better representing the distribution and trend of the data; statistical feature extraction processing: extracting statistical information of numerical features through statistical methods, such as mean, median, maximum, minimum, variance, and standard deviation.

[0042] Furthermore, the second feature extraction may include the following processing methods: one-hot encoding, which is used to convert categorical features into binary form, with each category corresponding to one binary bit, for model processing; category count encoding, which is used to convert categorical features into codes corresponding to the number of times the category occurs, so as to capture the frequency information of the category; category frequency encoding, which is used to convert categorical features into codes corresponding to the frequency of occurrence of the category, so as to capture the weight information of the category.

[0043] Furthermore, the third feature extraction may include the following processing methods: a bag-of-words model, which is used to convert text into a collection of words or phrases, construct a bag-of-words model, and count the number of occurrences or weights of each word or phrase; a term frequency-inverse document frequency (TF-IDF), which is used to assign a weight to each word by calculating the frequency of occurrence and inverse document frequency of the word to capture the importance of the word in the text; a word vector model (Word2Vec), which is used to convert text into a vector representation, and map each word to a continuous real number vector space by training the word vector model.

[0044] Optionally, an initial abnormal signal recognition model is established according to the denoised historical original signal, and the initial abnormal signal recognition model is trained to obtain the abnormal signal recognition model, wherein a suitable deep learning model can be selected according to the type and characteristics of the denoised historical original signal to construct the initial abnormal signal recognition model. For example, commonly used models include: Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory Network (LSTM), and Autoencoder, etc.

[0045] Alternatively, for time series data or sequence data, models such as RNN and LSTM may be more suitable; for image data, CNN is more commonly used; and autoencoders are frequently used because they can learn compressed representations of data and are used for anomaly detection.

[0046] It should be noted that this is only a preferred implementation method for obtaining the abnormal signal of the target data source, and the process and method for obtaining the abnormal signal of the target data source are not specifically limited. As long as the denoised original signal is input into the abnormal signal recognition model of the target data source for analysis, the process and method for obtaining the abnormal signal of the target data source are within the protection scope of the present invention and are not listed here.

[0047] Step S104: executing the exception handling strategy of the target data source on the exception signal.

[0048] In the technical solution provided in the above step S104 of the present invention, the exception handling strategy of the target data source is executed on the obtained exception signal, wherein the exception handling strategy is used to characterize the rules for performing exception handling on the exception signal.

[0049] Optionally, an abnormal signal recognition model is used to obtain an abnormality type corresponding to the abnormal signal, and a target abnormality processing strategy is executed on the abnormal signal according to the abnormality type, so as to achieve the purpose of effectively processing the abnormal signal.

[0050] It can be understood that this is only a preferred implementation of executing the exception handling strategy of the target data source on the exception signal, and does not specifically limit the process and method of executing the exception handling strategy of the target data source on the exception signal.

[0051] In the above steps S101 to S104 of the present invention, the original signal of the target data source can be first collected through the target interface of at least one target data source, and then the original signal obtained above can be denoised to obtain the denoised original signal, and then the denoised original signal can be input into the abnormal signal recognition model of the target data source for analysis, so as to obtain the abnormal signal of the target data source, and finally the abnormal signal can be processed by the abnormal signal processing strategy of the target data source to achieve the purpose of effectively processing the abnormal signal. Considering that after obtaining the denoised original signal, the denoised original signal can be input into the abnormal signal recognition model to obtain the abnormal signal, and then the abnormal signal is processed by the abnormal signal processing strategy, the security of using the target data source is guaranteed, so as to solve the technical problem of low processing efficiency of abnormal signals due to the inability to accurately identify abnormal signals, and achieve the technical effect of accurately identifying abnormal signals and improving the processing efficiency of abnormal signals.

[0052] The above method of this embodiment is further introduced below.

[0053] As an optional implementation method, the denoised original signal is input into an abnormal signal recognition model of a target data source for analysis to obtain an abnormal signal of the target data source, including: obtaining an abnormal pattern signal corresponding to the denoised original signal; using the abnormal signal recognition model to perform feature extraction on the abnormal pattern signal and the denoised original signal, respectively, to obtain a target abnormal pattern signal corresponding to the abnormal pattern signal, and a target signal corresponding to the denoised original signal; and determining the abnormal signal based on the target abnormal pattern signal and the target signal.

[0054] In this embodiment, an abnormal mode signal corresponding to the original signal after denoising can be obtained, and then the abnormal mode signal and the original signal after denoising are respectively feature extracted using the abnormal signal recognition model to obtain a target abnormal mode signal corresponding to the abnormal mode signal and a target signal corresponding to the original signal after denoising, and the purpose of determining the abnormal signal is achieved according to the target abnormal mode signal and the target signal obtained above. The abnormal mode signal can be a preset abnormal signal.

[0055] For example, an abnormal signal recognition model is used to analyze a preset abnormal signal and the original signal after denoising to determine the target signal. The abnormal signal recognition model can automatically analyze the signal characteristics and compare the original signal after denoising with the preset abnormal signal to determine whether the original signal after denoising is abnormal at this time, thereby achieving the purpose of determining the abnormal signal.

[0056] As an optional implementation method, based on the target abnormal pattern signal and the target signal, determining the abnormal signal includes: using an abnormal signal recognition model to obtain the similarity between the target abnormal pattern signal and the target signal; in response to the similarity and the similarity threshold being the same, determining the target signal as an abnormal signal.

[0057] In this embodiment, after obtaining the target abnormal pattern signal and the target signal, the abnormal signal recognition model can be used to obtain the similarity between the target abnormal pattern signal and the target signal, and when the similarity and the similarity threshold are the same, the target signal at this time is determined to be an abnormal signal; if the similarity and the similarity threshold are not the same, it is considered that the target signal at this time is not an abnormal signal. Wherein, the similarity threshold can be a preset value.

[0058] Optionally, using an abnormal signal recognition model, a first feature map corresponding to the target abnormal pattern signal and a second feature map corresponding to the target signal are obtained; the distance between the first feature map and the second feature map is calculated; in response to the distance being the same as the distance threshold, the target signal is determined to be an abnormal signal. It should be noted that this is only a preferred implementation method for determining an abnormal signal, and the process and method for determining an abnormal signal are not specifically limited.

[0059] For example, when the denoised real-time signal is input into the abnormal signal recognition model, the model will automatically analyze the characteristics of the real-time signal and compare it with the abnormal pattern signal learned during the training process, where the comparison is usually achieved by calculating the similarity or distance between feature vectors or feature graphs; if the characteristics of the real-time signal are highly similar to the characteristics of an abnormal pattern signal learned by the model (or the distance is close), the model will identify the signal as an abnormal signal, and the recognition process also involves threshold setting, that is, only when the similarity or distance exceeds a preset threshold, the real-time signal is considered to be abnormal.

[0060] As an optional implementation method, executing the exception handling strategy of the target data source on the exception signal includes: classifying the exception signal according to the target exception classification rule to obtain the exception level of the exception signal, wherein the exception level is used to characterize the severity of the exception generated by the data source corresponding to the exception signal; based on the exception level, executing the exception handling strategy on the exception signal.

[0061] In this embodiment, the abnormal signal can be classified according to the target abnormal classification rule to obtain the abnormal level of the abnormal signal, and then the abnormal processing strategy is implemented for the abnormal signal according to the abnormal level obtained above. The target abnormal classification rule can be a preset rule, such as a preset classification standard.

[0062] For example, an intelligent control system is used to classify the collected abnormal signals through preset classification standards, identify the abnormal level of the abnormal signal, and then execute different abnormality handling strategies for different abnormal levels.

[0063] As an optional implementation method, based on the exception level, an exception handling strategy is executed on the exception signal, including: in response to the exception level being a first exception level, a first exception handling strategy of executing the exception handling strategy on the exception signal, wherein the first exception handling strategy is used to start a backup device to put the target production system corresponding to the target data source into a normal operating state; in response to the exception level being a second exception level, a second exception handling strategy of executing the exception handling strategy on the exception signal, wherein the severity of the second exception level is higher than the severity of the first exception level, and the second exception handling strategy is used to put the target production system into a normal operating state by switching the first production line corresponding to the generated exception signal to the second production line; in response to the exception level being a third exception level, a third exception handling strategy of executing the exception handling strategy on the exception signal, wherein the severity of the third exception level is higher than the severity of the second exception level, and the third exception handling strategy is used to stop the control operation of the target production system.

[0064] In this embodiment, if the abnormality level is the first abnormality level, the first abnormality handling strategy is executed on the abnormal signal; if the abnormality level at this time is the second abnormality level, the second abnormality handling strategy needs to be executed on the abnormal signal; if the abnormality level at this time is the third abnormality level, the third abnormality handling strategy is executed on the abnormal signal. The first abnormality level can be called a slight abnormality, the second abnormality level can be called a medium abnormality, and the third abnormality level can be called a severe abnormality.

[0065] For example, when the abnormality level of the abnormal signal is a slight abnormality, it means that the failure of the data source that sends the abnormal signal is not very serious. Then the backup device can be directly started to replace the data source at this time to continue working, so as to ensure that the production system is in normal operation.

[0066] Furthermore, when the abnormality level of the abnormal signal is medium abnormal, it means that the failure of the data source that sends the abnormal signal has affected the entire production line. In this case, the production line needs to be switched to a normal production line so that the production system can continue to operate normally.

[0067] For another example, when the abnormality level of the abnormal signal is a serious abnormality, the control operation of the production system must be stopped at this time, and relevant personnel within the fault range must be evacuated.

[0068] As an optional implementation mode, after executing the exception handling strategy of the target data source on the abnormal signal, the method also includes: obtaining an alarm instruction of the target data source; generating alarm notification information in response to the alarm instruction, wherein the alarm notification information is used to notify the target object that an abnormality has occurred in the target data source.

[0069] In this embodiment, the alarm instruction of the target data source can be obtained, and the alarm notification information can be generated according to the alarm instruction, wherein the alarm notification information can be implemented in a variety of ways, such as text messages, emails, and sound and light alarms. It should be noted that only examples are given here to illustrate the implementation method of the alarm notification information, and the implementation method of the alarm notification information is not specifically limited.

[0070] For example, after executing the exception handling strategy, an alarm notification is triggered to notify relevant personnel to pay attention to the abnormal situation, where the alarm notification includes SMS, email, and sound and light alarm; at the same time, according to the abnormal level of the abnormal signal, the production parameters can be automatically adjusted, the backup equipment can be started, or other emergency measures can be implemented to quickly restore the system so that the system can operate normally; the production system records the results of each abnormal identification and the subsequent processing effects, which are used as feedback data for model retraining or parameter adjustment. Through continuous iterative optimization, the model can gradually adapt to new abnormal patterns, improve recognition accuracy and processing efficiency, and form a virtuous circle. Among them, the production system can be referred to as the system.

[0071] It should be noted that this is only a preferred implementation method for generating alarm notification information, and does not specifically limit the process and method for generating alarm notification information.

[0072] As an optional implementation manner, the target interface includes a first target interface and a second target interface, wherein the first target interface is used to represent a hardware interface of the target data source, and the second target interface is used to represent a software protocol interface corresponding to the target data source.

[0073] In this embodiment, the target interface may include a first target interface and a second target interface, wherein the first target interface may include a serial communication interface and a parallel bus interface. For example, the serial communication interface may be a Universal Asynchronous Receiver / Transmitter (UART), a Serial Peripheral Interface (SPI), etc.

[0074] Optionally, the second target interface is used to characterize the protocol specifications used when various data sources communicate with each other, such as Modular Digital Communication Protocol (MODBUS), Transmission Control Protocol / Internet Protocol (TCP / IP), etc. The software protocol specifies the data format, communication process, error handling and other specifications of the communicating parties to ensure that data exchange and communication can be carried out correctly between various data sources. Selecting a suitable software protocol can improve the compatibility and communication efficiency between various data sources.

[0075] In this embodiment, the original signal of the target data source can be first collected through the target interface of at least one target data source, and then the original signal obtained above can be denoised to obtain the denoised original signal, and then the denoised original signal can be input into the abnormal signal recognition model of the target data source for analysis, so as to obtain the abnormal signal of the target data source, and finally the abnormal signal can be processed by the abnormal signal processing strategy of the target data source to achieve the purpose of effectively processing the abnormal signal. Considering that after obtaining the denoised original signal, the denoised original signal can be input into the abnormal signal recognition model to obtain the abnormal signal, and then the abnormal signal is processed by the abnormal signal processing strategy, the security of using the target data source is guaranteed, so as to solve the technical problem of low processing efficiency of abnormal signals due to the inability to accurately identify abnormal signals, and achieve the technical effect of accurately identifying abnormal signals and improving the processing efficiency of abnormal signals.

[0076] The technical solution of the embodiment of the present invention is illustrated below in conjunction with preferred implementation modes.

[0077] In the process of industrial production and equipment operation, various abnormal signals often appear, such as fault alarm signals, fault display signals, etc. Timely identification of these abnormal signals and rapid processing of the abnormal signals are crucial to ensure the stability and reliability of the production system. At present, artificially set rules are often used to identify the abnormal signals generated by the production system, and the obtained abnormal signals are further processed. However, when processing abnormal signals involves manual intervention, omissions or delays are prone to occur, which leads to technical problems such as low efficiency in processing abnormal signals due to the inability to accurately identify abnormal signals.

[0078] Therefore, in order to solve the above problems, the present invention proposes a method for processing abnormal signals, which can first collect the original signal of the target data source through the target interface of at least one target data source, and then perform denoising on the original signal obtained above to obtain the denoised original signal, and then input the denoised original signal into the abnormal signal recognition model of the target data source for analysis, so as to obtain the abnormal signal of the target data source, and finally execute the abnormal processing strategy of the target data source on the abnormal signal, so as to achieve the purpose of effectively processing the abnormal signal. Considering that after obtaining the denoised original signal, the denoised original signal can be input into the abnormal signal recognition model to obtain the abnormal signal, and then execute the abnormal processing strategy on the abnormal signal, the security of using the target data source is guaranteed, so as to solve the technical problem of low processing efficiency of abnormal signals due to the inability to accurately identify abnormal signals, and achieve the technical effect of accurately identifying abnormal signals and then improving the processing efficiency of abnormal signals.

[0079] Figure 2 is a flow chart of a method for automatically identifying and processing abnormal signals according to an embodiment of the present invention. Figure 2 As shown, the method mainly includes the following steps:

[0080] Step S201, through a preset data interface or protocol, collect original signals from various data sources in real time or periodically, clean the original signals, remove invalid or redundant information, obtain the cleaned original signals, and then denoise the cleaned original signals to obtain the denoised original signals.

[0081] In this embodiment, signal data is collected from various data sources in real time or periodically through a preset data interface, and then the collected signal data is cleaned to remove invalid or redundant information to obtain cleaned signal data; the cleaned signal data is denoised to reduce the interference of noise on signal analysis; feature information useful for abnormality identification is extracted from the original data through feature extraction technology, laying the foundation for subsequent analysis. Among them, data sources include sensors and monitoring equipment, etc.

[0082] Optionally, feature information useful for abnormal signal identification is extracted from the original data through feature extraction technology, and the above implementation methods include: feature extraction of numerical data, feature extraction of categorical data and feature extraction of text data.

[0083] Optionally, the implementation methods of feature extraction of numerical data mainly include: missing value processing, which is used to process missing values ​​by filling, deleting or interpolating to ensure the integrity of the data; standardization and normalization processing, which is because numerical features often need to be standardized or normalized, and is converted into a form with a mean of 0 and a variance of 1, or a value range between [0,1], so that the model can better understand and process it; discretization processing, for continuous features, this method can discretize them into multiple intervals or bins to better represent the distribution and trend of the data; statistical feature extraction processing: extracting statistical information of numerical features through statistical methods, such as mean, median, maximum, minimum, variance, and standard deviation.

[0084] Optionally, the implementation methods of feature extraction of categorical data mainly include: one-hot encoding, which is used to convert categorical features into binary form, with each category corresponding to one binary bit, for model processing; category count encoding, which is used to convert categorical features into codes corresponding to the number of times the category occurs, so as to capture the frequency information of the category; category frequency encoding, which is used to convert categorical features into codes corresponding to the frequency of occurrence of the category, so as to capture the weight information of the category.

[0085] Optionally, the implementation methods of feature extraction of text data mainly include: bag-of-words model, which is used to convert text into a collection of words or phrases, build a bag-of-words model, and count the number of occurrences or weights of each word or phrase; word frequency-inverse document frequency, which is used to assign a weight to each word by calculating the frequency of occurrence and inverse document frequency of the word to capture the importance of the word in the text; word vector model, which is used to convert text into vector representation, and map each word to a continuous real number vector space by training the word vector model.

[0086] Step S202, using the denoised historical original signal to train a deep learning model so that the model can learn and identify abnormal signals, thereby obtaining a trained deep learning model.

[0087] In this embodiment, the denoised historical original signal is used to train a deep learning model so that it can learn and identify abnormal pattern signals in the data.

[0088] Optionally, the main process of training a deep learning model is: dividing the preprocessed data into a training set, a validation set, and a test set, wherein the training set is used for model training, the validation set is used to adjust model parameters and prevent overfitting, and the test set is used to evaluate the final performance of the model.

[0089] Furthermore, for certain types of data (such as images or text), data augmentation can be used to increase data diversity and improve the generalization ability of the model. For example, image data can be enhanced by rotation, scaling, cropping, etc.

[0090] In addition, it is also necessary to determine the model's learning rate, batch size, number of iterations and other hyperparameters. The selection of these parameters has an important impact on the model's training effect and performance. Use the training set data to train the deep learning model. During the training process, it is necessary to monitor the model's loss function and performance indicators (such as accuracy, recall rate, etc.) to evaluate the model's training effect. Adjust the model parameters according to the performance indicators of the validation set to prevent overfitting, and use regularization techniques, optimizers and other strategies to optimize model performance. Save the model's weights and parameters during the training process for subsequent loading and continued training.

[0091] Optionally, according to the type and characteristics of the abnormal signal data, a suitable deep learning model can be selected. Commonly used models include convolutional neural networks, recurrent neural networks, long short-term memory networks, and autoencoders, and for time series data or sequence data, models such as RNN and LSTM may be more suitable; for image data, CNN is more commonly used; and autoencoders are frequently used because they can learn compressed representations of data and are used in anomaly detection.

[0092] Step S203: input the denoised original signal into the trained deep learning model, analyze whether the denoised original signal is an abnormal signal, and if the denoised original signal is an abnormal signal, determine the abnormal level corresponding to the abnormal signal.

[0093] In this embodiment, the real-time signal is input into the trained model, and the model automatically analyzes the signal features and compares them with the learned abnormal pattern signals to determine whether the input real-time signal is abnormal and give an abnormality level.

[0094] Optionally, the model automatically analyzes signal features and compares them with learned abnormal pattern signals in the following process: real-time signals are acquired through sensors or other data acquisition devices and input into a trained deep learning model; necessary preprocessing is performed on the input signal, such as denoising, filtering, normalization, etc., to improve the signal quality and the recognition efficiency of the model, wherein the deep learning model has the ability to automatically learn features from data; during the training process, the model learns how to extract key features from the signal and represent these features as high-level abstract features; the extracted features are then nonlinearly transformed and combined through the hidden layer of the model to form a feature vector or feature map that can reflect the signal characteristics.

[0095] Furthermore, the comparison is usually achieved by calculating the similarity or distance between feature vectors or feature graphs; if the feature of the real-time signal is highly similar to the feature of a certain abnormal pattern signal learned by the model (or the distance is close), the model will identify the signal as an abnormal signal, and the identification process also involves threshold setting, that is, only when the similarity or distance exceeds a certain preset threshold, the real-time signal at this time is considered abnormal. In addition, the model performs the final abnormality identification of the real-time signal based on the above comparison results and preset decision rules, and outputs abnormality identification information, such as whether the real-time signal at this time is abnormal. In the case of an abnormality in the real-time signal, the abnormality level of the real-time signal is obtained. The above information can be used for subsequent alarms, troubleshooting, system optimization and other applications.

[0096] Step S204, triggering an alarm notification to notify relevant personnel to pay attention to the abnormality. At the same time, according to the abnormality level, relevant emergency measures are executed to quickly restore the system.

[0097] In this embodiment, an alarm notification can be triggered according to preset processing rules and processes to notify relevant personnel to pay attention to abnormal situations, wherein the alarm notification includes text messages, emails, and sound and light alarms; at the same time, according to the abnormality level of the abnormal signal, production parameters can be automatically adjusted, backup equipment can be started, or other emergency measures can be implemented to quickly restore the system so that the system can operate normally.

[0098] Optionally, an intelligent control system is used to classify the collected abnormal signals according to preset classification standards, identify the abnormal levels of the abnormal signals, and then execute different abnormality handling strategies for different abnormal levels.

[0099] For example, abnormal signals are classified according to preset classification standards to obtain the abnormal level of the abnormal signal. According to the abnormal level, the intelligent control system automatically adjusts the production parameters to correct the abnormal signal and maintain the stability of the production process. The abnormal levels include minor abnormalities, moderate abnormalities and severe abnormalities. Detailed emergency plans are formulated for different types of abnormalities. The emergency plans include starting backup equipment, switching production lines and evacuating personnel, which can clarify the emergency response process, division of responsibilities and implementation of emergency measures.

[0100] In step S205, the production system records the results of each abnormality identification and the subsequent processing effects as feedback data for model retraining or parameter adjustment. Through continuous iterative optimization, the model can gradually adapt to new abnormal patterns, improve recognition accuracy and processing efficiency, and form a virtuous circle.

[0101] In this embodiment, the original signal of the target data source can be first collected through the target interface of at least one target data source, and then the original signal obtained above can be denoised to obtain the denoised original signal, and then the denoised original signal can be input into the abnormal signal recognition model of the target data source for analysis, so as to obtain the abnormal signal of the target data source, and finally the abnormal signal can be processed by the abnormal signal processing strategy of the target data source to achieve the purpose of effectively processing the abnormal signal. Considering that after obtaining the denoised original signal, the denoised original signal can be input into the abnormal signal recognition model to obtain the abnormal signal, and then the abnormal signal is processed by the abnormal signal processing strategy, the security of using the target data source is guaranteed, so as to solve the technical problem of low processing efficiency of abnormal signals due to the inability to accurately identify abnormal signals, and achieve the technical effect of accurately identifying abnormal signals and improving the processing efficiency of abnormal signals.

[0102] According to an embodiment of the present invention, a device for processing an abnormal signal is provided. It should be noted that the device for processing an abnormal signal can be used to execute a method for processing an abnormal signal in an embodiment.

[0103] Figure 3 is a schematic diagram of a device for processing abnormal signals according to an embodiment of the present invention, such as Figure 3 As shown, the abnormal signal processing device 300 may include: a collection unit 301 , a denoising unit 302 , an analysis unit 303 and an execution unit 304 .

[0104] The collection unit 301 is used to collect original signals of a target data source through a target interface of at least one target data source, wherein the original signals are used to represent the operating status of the target data source.

[0105] The denoising unit 302 is used to perform denoising on the original signal, wherein the noise signal in the original signal after denoising is less than the noise signal in the original signal before denoising.

[0106] The analysis unit 303 is used to input the denoised original signal into the abnormal signal recognition model of the target data source for analysis to obtain the abnormal signal of the target data source, wherein the abnormal signal recognition model is established by the denoised historical original signal of the target data source.

[0107] The execution unit 304 is used to execute the exception handling strategy of the target data source on the exception signal, wherein the exception handling strategy is used to represent the rules for performing exception handling on the exception signal.

[0108] Optionally, the analysis unit 303 may include: a first acquisition module, used to acquire an abnormal pattern signal corresponding to the denoised original signal; a second acquisition module, used to use an abnormal signal recognition model to perform feature extraction on the abnormal pattern signal and the denoised original signal, respectively, to obtain a target abnormal pattern signal corresponding to the abnormal pattern signal, and a target signal corresponding to the denoised original signal; a determination module, used to determine the abnormal signal based on the target abnormal pattern signal and the target signal.

[0109] Optionally, the determination module may include: an acquisition submodule, used to use the abnormal signal recognition model to obtain the similarity between the target abnormal pattern signal and the target signal; and a determination submodule, used to determine the target signal as an abnormal signal in response to the similarity and the similarity threshold being the same.

[0110] Optionally, the execution unit 304 may include: a classification module, used to classify the abnormal signal according to the target abnormal classification rule to obtain the abnormal level of the abnormal signal, wherein the abnormal level is used to characterize the severity of the abnormality generated by the data source corresponding to the abnormal signal; and an execution module, used to execute an abnormality handling strategy on the abnormal signal based on the abnormal level.

[0111] Optionally, the execution module may include: a first execution submodule, for executing a first exception handling strategy of the exception handling strategy on the exception signal in response to the exception level being a first exception level, wherein the first exception handling strategy is used to start a backup device to put the target production system corresponding to the target data source into a normal operating state; a second execution submodule, for executing a second exception handling strategy of the exception handling strategy on the exception signal in response to the exception level being a second exception level, wherein the severity of the second exception level is higher than the severity of the first exception level, and the second exception handling strategy is used to put the target production system into a normal operating state by switching the first production line corresponding to the generated exception signal to the second production line; a third execution submodule, for executing a third exception handling strategy of the exception handling strategy on the exception signal in response to the exception level being a third exception level, wherein the severity of the third exception level is higher than the severity of the second exception level, and the third exception handling strategy is used to stop the control operation of the target production system.

[0112] Optionally, after executing the exception handling strategy of the target data source on the abnormal signal, the device also includes: an acquisition unit, used to obtain an alarm instruction of the target data source; a generation unit, used to generate alarm notification information in response to the alarm instruction, wherein the alarm notification information is used to notify the target object that an abnormality has occurred in the target data source.

[0113] Optionally, the target interface includes a first target interface and a second target interface, wherein the first target interface is used to represent a hardware interface of the target data source, and the second target interface is used to represent a software protocol interface corresponding to the target data source.

[0114] In this embodiment, an acquisition unit acquires an original signal of a target data source through a target interface of at least one target data source, wherein the original signal is used to characterize the operating state of the target data source; a denoising unit performs denoising on the original signal, wherein the noise signal in the denoised original signal is less than the noise signal in the original signal before denoising; an analysis unit inputs the denoised original signal into an abnormal signal recognition model of the target data source for analysis to obtain an abnormal signal of the target data source, wherein the abnormal signal recognition model is established by using the denoised historical original signal of the target data source; an execution unit executes an abnormal processing strategy of the target data source on the abnormal signal, wherein the abnormal processing strategy is used to characterize the rules for performing abnormal processing on the abnormal signal, thereby solving the technical problem of low processing efficiency of abnormal signals due to failure to accurately identify abnormal signals, and achieving the technical effect of accurately identifying abnormal signals and thus improving processing efficiency of abnormal signals.

[0115] According to an embodiment of the present invention, a computer-readable storage medium is also provided, which includes a stored program, wherein when the program is executed by a processor, the device where the computer-readable storage medium is located is controlled to execute the abnormal signal processing method in the embodiment.

[0116] Optionally, the storage medium is used for the original signal to characterize the operating status of the target data source; for denoising the original signal, wherein the noise signal in the original signal after denoising is less than the noise signal in the original signal before denoising; for inputting the denoised original signal into an abnormal signal recognition model of the target data source for analysis to obtain the abnormal signal of the target data source, wherein the abnormal signal recognition model is established through the denoised historical original signal of the target data source; and for executing the abnormal handling strategy of the target data source on the abnormal signal, wherein the abnormal handling strategy is used to characterize the rules for performing abnormal handling on the abnormal signal.

[0117] Optionally, the storage medium is used to input the denoised original signal into an abnormal signal recognition model of a target data source for analysis to obtain an abnormal signal of the target data source, including: obtaining an abnormal pattern signal corresponding to the denoised original signal; using the abnormal signal recognition model to perform feature extraction on the abnormal pattern signal and the denoised original signal, respectively, to obtain a target abnormal pattern signal corresponding to the abnormal pattern signal, and a target signal corresponding to the denoised original signal; and determining the abnormal signal based on the target abnormal pattern signal and the target signal.

[0118] Optionally, the storage medium is used to determine the abnormal signal based on the target abnormal pattern signal and the target signal, including: using an abnormal signal recognition model to obtain the similarity between the target abnormal pattern signal and the target signal; in response to the similarity and the similarity threshold being the same, determining the target signal as an abnormal signal.

[0119] Optionally, the storage medium is used to execute an exception handling strategy of a target data source on an abnormal signal, including: classifying the abnormal signal according to a target abnormal classification rule to obtain an abnormality level of the abnormal signal, wherein the abnormality level is used to characterize the severity of the abnormality generated by the data source corresponding to the abnormal signal; and executing an exception handling strategy on the abnormal signal based on the abnormality level.

[0120] Optionally, the storage medium is used to execute an exception handling strategy on the exception signal based on the exception level, including: in response to the exception level being a first exception level, executing a first exception handling strategy on the exception signal, wherein the first exception handling strategy is used to start a backup device to put the target production system corresponding to the target data source into a normal operating state; in response to the exception level being a second exception level, executing a second exception handling strategy on the exception signal, wherein the severity of the second exception level is higher than the severity of the first exception level, and the second exception handling strategy is used to put the target production system into a normal operating state by switching the first production line corresponding to the generated exception signal to the second production line; in response to the exception level being a third exception level, executing a third exception handling strategy on the exception signal, wherein the severity of the third exception level is higher than the severity of the second exception level, and the third exception handling strategy is used to stop the control operation of the target production system.

[0121] Optionally, after executing the exception handling strategy of the target data source on the abnormal signal, the storage medium is also used to obtain an alarm instruction of the target data source; and to generate alarm notification information in response to the alarm instruction, wherein the alarm notification information is used to notify the target object that an abnormality has occurred in the target data source.

[0122] Optionally, the target interface includes a first target interface and a second target interface, wherein the first target interface is used to represent a hardware interface of the target data source, and the second target interface is used to represent a software protocol interface corresponding to the target data source.

[0123] According to an embodiment of the present invention, a processor is further provided. The processor is used to run a program, wherein the method for processing an abnormal signal in the embodiment is executed when the program is run.

[0124] According to an embodiment of the present invention, a computer program product is further provided, including a computer program, and when the computer program is executed by a processor, the method for processing an abnormal signal in the embodiment of the present invention is implemented.

[0125] According to an embodiment of the present invention, an electronic device is further provided, the electronic device comprising a processor and a memory for storing instructions executable by the processor, wherein the processor is configured to execute the instructions to implement the abnormal signal processing method in the embodiment of the present invention.

[0126] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0127] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0128] In the several embodiments provided by the present invention, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic, for example, the division of units can be a logical function division, and there can be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0129] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed over multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0130] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0131] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0132] The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for processing abnormal signals, characterized in that: include: Collecting an original signal of the target data source through a target interface of at least one target data source, wherein the original signal is used to characterize an operating state of the target data source; Performing denoising on the original signal, wherein the noise signal in the original signal after denoising is less than the noise signal in the original signal before denoising; Inputting the denoised original signal into the abnormal signal recognition model of the target data source for analysis to obtain the abnormal signal of the target data source, wherein the abnormal signal recognition model is established by the denoised historical original signal of the target data source; An exception handling strategy of the target data source is executed on the exception signal, wherein the exception handling strategy is used to characterize a rule for performing exception handling on the exception signal.

2. The method according to claim 1, characterized in that: The denoised original signal is input into the abnormal signal recognition model of the target data source for analysis to obtain the abnormal signal of the target data source, including: Acquire an abnormal mode signal corresponding to the original signal after denoising; Using the abnormal signal recognition model, feature extraction is performed on the abnormal pattern signal and the original signal after denoising, respectively, to obtain a target abnormal pattern signal corresponding to the abnormal pattern signal and a target signal corresponding to the original signal after denoising; The abnormal signal is determined based on the target abnormal pattern signal and the target signal.

3. The method according to claim 2, characterized in that Determining the abnormal signal based on the target abnormal mode signal and the target signal includes: Using the abnormal signal recognition model, obtaining the similarity between the target abnormal pattern signal and the target signal; In response to the similarity being the same as the similarity threshold, the target signal is determined to be the abnormal signal.

4. The method according to claim 1, characterized in that Executing the exception handling strategy of the target data source on the exception signal includes: According to the target abnormality classification rule, the abnormal signal is classified to obtain the abnormality level of the abnormal signal, wherein the abnormality level is used to characterize the severity of the abnormality generated by the data source corresponding to the abnormal signal; Based on the exception level, the exception handling strategy is executed on the exception signal.

5. The method according to claim 4, characterized in that Based on the abnormality level, executing the abnormality handling strategy on the abnormal signal includes: In response to the abnormality level being the first abnormality level, executing a first abnormality handling strategy of the abnormality handling strategy on the abnormal signal, wherein the first abnormality handling strategy is used to enable the target production system corresponding to the target data source to be in a normal operating state by starting a backup device; In response to the abnormality level being a second abnormality level, executing a second abnormality handling strategy of the abnormality handling strategy on the abnormal signal, wherein the severity of the second abnormality level is higher than the severity of the first abnormality level, and the second abnormality handling strategy is used to switch the first production line corresponding to the generation of the abnormal signal to the second production line, so that the target production system is in the normal operating state; In response to the exception level being a third exception level, executing a third exception handling strategy of the exception handling strategy on the exception signal, wherein the severity of the third exception level is higher than the severity of the second exception level, and the third exception handling strategy is used to stop the control operation of the target production system.

6. The method according to claim 1, characterized in that After executing the exception handling strategy of the target data source on the exception signal, the method further includes: Obtaining an alarm instruction of the target data source; In response to the alarm instruction, an alarm notification message is generated, wherein the alarm notification message is used to notify the target object that an abnormality has occurred in the target data source.

7. The method according to claim 1, characterized in that The target interface includes a first target interface and a second target interface, wherein the first target interface is used to represent a hardware interface of the target data source, and the second target interface is used to represent a software protocol interface corresponding to the target data source.

8. A device for processing abnormal signals, characterized in that: include: A collection unit, configured to collect an original signal of the target data source through a target interface of at least one target data source, wherein the original signal is used to characterize an operating state of the target data source; A denoising unit, configured to perform denoising processing on the original signal, wherein the noise signal in the original signal after denoising is less than the noise signal in the original signal before denoising; An analysis unit, configured to input the denoised original signal into an abnormal signal recognition model of the target data source for analysis, so as to obtain an abnormal signal of the target data source, wherein the abnormal signal recognition model is established by using the denoised historical original signal of the target data source; An execution unit is used to execute the exception handling strategy of the target data source on the exception signal, wherein the exception handling strategy is used to characterize a rule for performing exception handling on the exception signal.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed by a processor, the device where the storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: The invention comprises one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any one of claims 1 to 7.

Images