Method and system for reducing invalid alarms of rail transit signal monitoring system

Abstract

The invention provides a method and system for reducing invalid alarms of a rail transit signal monitoring system, and the method comprises the steps: obtaining alarm data which comprises the alarm content, the alarm time, the alarm level and the alarm validity; constructing an alarm prediction model based on a long short-term memory network; pre-processing the alarm data, training the alarm prediction model by taking the pre-processed alarm data as a training set, and monitoring the performance of the alarm prediction model by using a verification set; an alarm module of the rail transit signal monitoring system calls the alarm prediction model, when an alarm is about to be generated, the content of the alarm about to be generated is input into the alarm prediction model, and the alarm prediction model predicts whether the alarm about to be generated is effective or not. And a mode of predicting subsequent alarms by using the LSTM model is used, so that repeated and meaningless alarms are effectively reduced, and the repeated workload of electric service personnel is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of rail transit signal monitoring, and in particular to a method and system for reducing invalid alarms in a rail transit signal monitoring system. Background Art

[0002] In the field of railway signaling, monitoring systems play an important role in the daily maintenance of equipment by electricians, especially with the continuous updating and upgrading of modern railways, the intelligence of various equipment has been improved, and the complexity of equipment has been rising. When equipment failure occurs, the monitoring system is required to issue an alarm in a timely and accurate manner, which requires the signal system to issue an alarm efficiently and accurately. However, due to the complexity of the signal system and the intricate relationships between various alarms, the occurrence of an alarm event may cause multiple alarms to appear at the same time. Therefore, how to know which alarms are valid and which are unimportant or repeated alarms has become a problem.

[0003] Therefore, it is necessary to propose a method and system for reducing invalid alarms in a rail transit signal monitoring system. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for reducing invalid alarms in a rail transit signal monitoring system, so as to reduce the generation of repeated and meaningless alarms, facilitate users to quickly find the real cause of the alarm, promptly and quickly handle the alarm, and reduce the repetitive workload of electrical personnel.

[0005] In order to achieve the above-mentioned object of the invention, in a first aspect, the present invention provides a method for reducing invalid alarms in a rail transit signal monitoring system, characterized in that it includes:

[0006] Acquiring alarm data, wherein the alarm data includes alarm content, alarm time, alarm level and alarm validity;

[0007] Construct an alarm prediction model based on long short-term memory network;

[0008] Preprocessing the alarm data, using the preprocessed alarm data as a training set to train the alarm prediction model, and using a validation set to monitor the performance of the alarm prediction model;

[0009] The alarm module of the rail transit signal monitoring system calls the alarm prediction model. When an alarm is about to be generated, the content of the alarm to be generated is input into the alarm prediction model, and the alarm prediction model predicts whether the alarm to be generated is valid.

[0010] Optionally, obtaining alarm data includes:

[0011] Retrieve all alarm data from the alarm database, generate alarm content data files with device name and alarm content, generate alarm level data files with alarm level, and generate alarm time data files with alarm occurrence time;

[0012] Create a new alarm validity file, analyze whether all alarm data is valid, and mark the validity. Valid alarms are marked as 1 and invalid alarms are marked as 0.

[0013] Optionally, the constructing of an alarm prediction model based on a long short-term memory network includes:

[0014] A long short-term memory network model including time dimension, alarm level dimension and effectiveness dimension is constructed. The time dimension is input as a feature to the fully connected layer, and two output layers are constructed, which are used for the regression of alarm level and the binary classification of whether the alarm is effective.

[0015] Optionally, the preprocessing of the alarm data includes:

[0016] Convert the alarm time into a timestamp format;

[0017] Convert the alarm level into a number, the smaller the number, the higher the level, and the highest level is 1;

[0018] The alarm content is segmented, stop words are removed, stems are extracted, and then converted into word vectors.

[0019] Optionally, the alarm prediction model predicts whether the alarm to be generated is effective, including:

[0020] If the alarm that is about to be generated is predicted to be a valid alarm, and no other alarm is predicted to occur within a certain number of seconds, the alarm that is about to be generated is valid and reported.

[0021] Optionally, the alarm prediction model predicts whether the alarm to be generated is effective, and further includes:

[0022] If the alarm that is about to be generated is predicted to be invalid, and a new alarm is predicted to occur after a certain number of seconds, the current alarm will be temporarily withheld and not reported, and a certain number of seconds will be waited for the alarm to time out.

[0023] Optionally, the alarm prediction model predicts whether the alarm to be generated is effective, and further includes:

[0024] If no new alarm is generated after waiting for several seconds, the current alarm is judged to be valid and reported.

[0025] Optionally, the alarm prediction model predicts whether the alarm to be generated is effective, and further includes:

[0026] If a new alarm is generated during the waiting process of several seconds, and it is determined that the new alarm level is lower, the new alarm is discarded and not reported, and the waiting process continues for several seconds to expire.

[0027] Optionally, the alarm prediction model predicts whether the alarm to be generated is effective, and further includes:

[0028] If a new alarm is generated during the waiting process of several seconds, and it is determined that the new alarm level is higher, the higher level alarm is retained and the waiting process continues for several seconds to time out.

[0029] In a second aspect, the present invention provides a readable storage medium, wherein the readable storage medium stores a computer program, and when the computer program is executed by a processor, any one of the methods for reducing invalid alarms in a rail transit signal monitoring system is implemented.

[0030] In a third aspect, the present invention provides a rail transit signal monitoring system, comprising:

[0031] Alarm module;

[0032] Alarm prediction model;

[0033] The alarm module is used to input the content of the alarm to be generated into the alarm prediction model when an alarm is about to be generated, and the alarm prediction model predicts the effectiveness of the alarm to be generated.

[0034] Compared with the prior art, the technical solution of the present invention has at least the following beneficial effects:

[0035] The present invention uses a long short-term memory network (LSTM) to analyze and model a large amount of alarm content including the time dimension, the alarm level dimension, and the effective alarm dimension, so as to achieve the purpose of predicting alarms and reducing the occurrence of repeated alarms and invalid alarms. The method of using the LSTM model to predict subsequent alarms effectively reduces the generation of repeated and meaningless alarms, facilitates users to quickly find the real cause of the alarm, and can promptly and quickly handle the alarm, reducing the repetitive workload of electricians. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A schematic flow chart of a method for reducing invalid alarms in a rail transit signal monitoring system according to an embodiment of the present invention;

[0037] Figure 2 A schematic diagram of a process flow of alarm data collection and processing according to an embodiment of the present invention;

[0038] Figure 3 A schematic diagram of a construction process of an alarm prediction model according to an embodiment of the present invention;

[0039] Figure 4The figure is a schematic diagram of the prediction and sending process when an alarm occurs according to an embodiment of the present invention. DETAILED DESCRIPTION

[0040] The following is a further detailed description of the method for reducing invalid alarms in the rail transit signal monitoring system of the present invention in combination with the accompanying drawings and specific embodiments. According to the following description, the advantages and features of the present invention will be clearer. It should be noted that the accompanying drawings are in a very simplified form and use non-precise proportions, which are only used to conveniently and clearly assist in explaining the purpose of the embodiments of the present invention. In order to make the purposes, features and advantages of the present invention more obvious and easy to understand, please refer to the accompanying drawings. It should be noted that the structure, proportion, size, etc. illustrated in the accompanying drawings of this specification are only used to match the contents disclosed in the specification for people familiar with this technology to understand and read, and are not used to limit the limiting conditions for the implementation of the present invention, so they have no technical substantive significance. Any structural modification, change in proportional relationship or adjustment of size, without affecting the effects that the present invention can produce and the purposes that can be achieved, should still fall within the scope of the technical content disclosed by the present invention.

[0041] See also Figure 1 , Figure 1 The present invention is a flow chart of a method for reducing invalid alarms in a rail transit signal monitoring system according to an embodiment of the present invention, comprising:

[0042] Step S1: Acquire alarm data, wherein the alarm data includes alarm content, alarm time, alarm level and alarm validity;

[0043] Step S2: constructing an alarm prediction model based on long short-term memory network;

[0044] Step S3: preprocessing the alarm data, and using the preprocessed alarm data as a training set to train the alarm prediction model, and using a validation set to monitor the performance of the alarm prediction model;

[0045] Step S4: The alarm module of the rail transit signal monitoring system calls the alarm prediction model. When an alarm is about to be generated, the content of the alarm to be generated is input into the alarm prediction model, and the alarm prediction model predicts whether the alarm to be generated is valid.

[0046] The present invention uses a long short-term memory network (LSTM) to analyze and model a large amount of alarm content including the time dimension, the alarm level dimension, and the effective alarm dimension, so as to achieve the purpose of predicting alarms and reducing the occurrence of repeated alarms and invalid alarms. The method of using the LSTM model to predict subsequent alarms effectively reduces the generation of repeated and meaningless alarms, facilitates users to quickly find the real cause of the alarm, and can promptly and quickly handle the alarm, reducing the repetitive workload of electricians.

[0047] This embodiment uses a long short-term memory network to model the alarm content, and converts the alarm content into a format that can be processed by the neural network through a series of preprocessing steps such as word segmentation, word vector representation, and serialization.

[0048] In step S1, prepare alarm data, see Figure 2 , retrieve all alarm data from the real database; generate the device name and alarm content into an alarm content data file, generate the alarm level into an alarm level data file, generate the alarm occurrence time into an alarm event time data file, and add a file for whether the alarm is valid. All alarm analysis marks are valid alarms. These files are used as input sources, the alarm level is used as the importance dimension label of the alarm content, the alarm time is used as the time dimension label of the alarm content, and whether the alarm is valid is used as a binary dimension label to indicate whether this alarm is valid.

[0049] In step S2, please refer to Figure 3 , the construction of the alarm prediction model based on the long short-term memory network includes:

[0050] TensorFlow and Keras are two popular open source libraries for building and training deep learning models. TensorFlow is a powerful machine learning framework developed by Google, and Keras is a high-level neural network API (Application Programming Interface) that can run on top of TensorFlow. Use TensorFlow and Keras to build a long short-term memory network model that includes time dimension, alarm level dimension, and effectiveness dimension. The time dimension is used as a feature input to the fully connected layer, and two output layers are built, which are used for the regression of the alarm level and the binary classification of whether the alarm is effective.

[0051] In step S3, the alarm data is preprocessed, see Figure 2 , convert the alarm time into a timestamp format; convert the alarm level into a number, the smaller the number, the higher the level, and the highest level is 1; segment the alarm content, remove stop words, extract stems, etc., and then convert it into a word vector.

[0052] LSTM is used to process text sequences, and the time dimension and alarm level dimension are integrated into the LSTM layer as additional features.

[0053] Input the preprocessed data into the model for training, use the validation set to monitor the model performance, and make appropriate adjustments.

[0054] Integrate the trained alarm prediction model into the alarm module.

[0055] The alarm prediction model in step S4 predicts whether the alarm to be generated is effective. Figure 4 When an alarm is about to be generated, the content of the alarm to be generated is first input into the prediction model, and the model is used to predict whether this alarm is a valid alarm and which alarms may appear in the next few seconds.

[0056] If the model determines that this alarm content is a valid alarm and predicts that there will be no other alarms in a few seconds, this alarm will be issued.

[0057] If the model determines that the alarm content may be an invalid alarm and predicts that other alarms may appear within a few seconds, the alarm will be temporarily withheld and wait for a few seconds to time out.

[0058] If a new alarm is generated at this time, and the alarm content is within the range predicted by the model, if the alarm level to be generated is higher than the previous one, then use this higher-level alarm as the input sequence, and the previous alarm will be terminated without being reported; if the new alarm level is lower than the previous one, then terminate the current new alarm and continue waiting.

[0059] If no new alarm is issued within the effective time range predicted by the model, this withheld alarm will be issued.

[0060] Specifically, the method for reducing invalid alarms in the rail transit signal monitoring system provided in this embodiment can be implemented by the following code:

[0061] Segment and encode the alarm content:

[0062] Tokenizer(num_words=10000),tokenizer.fit_on_texts()

[0063] Convert the time, alarm level, and whether the alarm is valid into a numpy array:

[0064] alarmTimeArr=np.array(alarmTimes), alarmLevelArr=np.array(alarmLevels), alarmValidArr=np.array(alarmValids)

[0065] Construct the model input layer:

[0066] alarmContentInput=Input(shape=(200,),name='alarmContentInput');

[0067] embeddingLayer=Embedding(input_dim=10000, output_dim=64, input_length=200)(alarmContentInput); LSTMLayer=LSTM(64, return_sequences=False)(embeddingLayer)

[0068] Establish the alarm time dimension input layer:

[0069] alarmTimeInput=Input(shape=(1,),name='alarmTimeInput')

[0070] Create the alarm level dimension input layer:

[0071] alarmLevelInput=Input(shape=(1,),name='alarmLevelInput')

[0072] Establish whether the alarm is valid input layer:

[0073] alarmValidInput=Input(shape=(1,),name='alarmValidInput')

[0074] Merge the input layer and use the sigmoid activation function:

[0075] concatenated=Concatenate()([lstm_layer, alarmTimeInput, alarmLevelInput, alarmValidInput]); dropout=Dropout(0.5)(concatenated), dense_layer=Dense(64, activation='relu')(dropout); output_layer=Dense(1, activation='sigmoid')(dense_layer)

[0076] Build the complete model:

[0077] model=Model(inputs=[alarmContentInput, alarmTimeInput, alarmLevelInput, alarmValidInput], outputs=output_layer); model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])

[0078] Train the model:

[0079] model.fit([alarmContentData,alarmTimeData,alarmLevelData,alarmValidData],epochs=10,batch_size=32)

[0080] This embodiment further provides a readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, any one of the methods for reducing invalid alarms in a rail transit signal monitoring system is implemented.

[0081] On the other hand, this embodiment also provides a rail transit signal monitoring system, including:

[0082] Alarm module;

[0083] Alarm prediction model;

[0084] The alarm module is used to input the content of the alarm to be generated into the alarm prediction model when an alarm is about to be generated, and the alarm prediction model predicts the effectiveness of the alarm to be generated.

[0085] The rail transit signal monitoring system provided in this embodiment, the alarm module can call the alarm prediction model to implement the method of reducing invalid alarms in the rail transit signal monitoring system, and use the LSTM model to predict subsequent alarms, thereby effectively reducing the generation of repeated and meaningless alarms, making it convenient for users to quickly find the real cause of the alarm, and being able to handle the alarm in a timely and rapid manner, reducing the repetitive workload of electricians.

[0086] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0087] It should be noted that the devices and methods disclosed in the embodiments of this article can also be implemented in other ways. The device implementation described above is only schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to multiple embodiments of this article. In this regard, each box in the flowchart or block diagram can represent a part of a module, program or code, and the module, program segment or part of the code contains one or more executable instructions for implementing the specified logical function, and the module, program segment or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented by a dedicated hardware-based system for performing a specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions.

[0088] In addition, the functional modules in the various embodiments of this document may be integrated together to form an independent part, or each module may exist independently, or two or more modules may be integrated to form an independent part.

[0089] Although the content of the present invention has been described in detail through the above preferred embodiments, it should be appreciated that the above description should not be considered as a limitation of the present invention. After reading the above content, it will be apparent to those skilled in the art that various modifications and substitutions of the present invention will occur. Therefore, the protection scope of the present invention should be limited by the appended claims.

Claims

1. A method for reducing invalid alarms in a rail transit signal monitoring system, characterized in that: include: Acquiring alarm data, wherein the alarm data includes alarm content, alarm time, alarm level and alarm validity; Construct an alarm prediction model based on long short-term memory network; Preprocessing the alarm data, using the preprocessed alarm data as a training set to train the alarm prediction model, and using a validation set to monitor the performance of the alarm prediction model; The alarm module of the rail transit signal monitoring system calls the alarm prediction model. When an alarm is about to be generated, the content of the alarm to be generated is input into the alarm prediction model, and the alarm prediction model predicts the effectiveness of the alarm to be generated.

2. The method for reducing invalid alarms in a rail transit signal monitoring system according to claim 1, characterized in that: The obtaining of alarm data comprises: Retrieve all alarm data from the alarm database, generate alarm content data files with device name and alarm content, generate alarm level data files with alarm level, and generate alarm time data files with alarm occurrence time; Create a new alarm validity file, analyze whether all alarm data is valid, and mark the validity. Valid alarms are marked as 1 and invalid alarms are marked as 0.

3. The method for reducing invalid alarms in a rail transit signal monitoring system according to claim 1, characterized in that: The construction of the alarm prediction model based on the long short-term memory network includes: A long short-term memory network model including time dimension, alarm level dimension and effectiveness dimension is constructed. The time dimension is input as a feature to the fully connected layer, and two output layers are constructed, which are used for the regression of alarm level and the binary classification of whether the alarm is effective.

4. The method for reducing invalid alarms in a rail transit signal monitoring system according to claim 1, characterized in that: The preprocessing of the alarm data comprises: Convert the alarm time into a timestamp format; Convert the alarm level into a number, the smaller the number, the higher the level, and the highest level is 1; The alarm content is segmented, stop words are removed, stems are extracted, and then converted into word vectors.

5. The method for reducing invalid alarms in a rail transit signal monitoring system according to claim 1, characterized in that: The alarm prediction model predicts whether the alarm to be generated is effective, including: If the alarm that is about to be generated is predicted to be a valid alarm, and no other alarm is predicted to occur within a certain number of seconds, the alarm that is about to be generated is valid and reported.

6. The method for reducing invalid alarms in a rail transit signal monitoring system according to claim 5, characterized in that: The alarm prediction model predicts whether the alarm to be generated is effective, and also includes: If the alarm that is about to be generated is predicted to be invalid, and a new alarm is predicted to occur after a certain number of seconds, the current alarm will be temporarily withheld and not reported, and a certain number of seconds will be waited for the alarm to time out.

7. The method for reducing invalid alarms in a rail transit signal monitoring system according to claim 6, characterized in that: The alarm prediction model predicts whether the alarm to be generated is effective, and also includes: If no new alarm is generated after waiting for several seconds, the current alarm is judged to be valid and reported.

8. The method for reducing invalid alarms in a rail transit signal monitoring system according to claim 5, characterized in that: The alarm prediction model predicts whether the alarm to be generated is effective, and also includes: If a new alarm is generated during the waiting process of several seconds, and it is determined that the new alarm level is lower, the new alarm is discarded and not reported, and the waiting process continues for several seconds to expire.

9. The method for reducing invalid alarms in a rail transit signal monitoring system according to claim 5, characterized in that: The alarm prediction model predicts whether the alarm to be generated is effective, and also includes: If a new alarm is generated during the waiting process of several seconds, and it is determined that the new alarm level is higher, the higher level alarm is retained and the waiting process continues for several seconds to time out.

10. A readable storage medium, characterized in that: The readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for reducing invalid alarms in a rail transit signal monitoring system according to any one of claims 1 to 9 is implemented.

11. A rail transit signal monitoring system, characterized in that: include: Alarm module; Alarm prediction model; The alarm module is used to input the content of the alarm to be generated into the alarm prediction model when an alarm is about to be generated, and the alarm prediction model predicts the effectiveness of the alarm to be generated.

Images