A highway traffic information publishing equipment state monitoring method, equipment and medium

Abstract

The embodiment of the present invention discloses a method for monitoring the state of a highway traffic information publishing device, comprising: obtaining power quality monitoring data during the operation of the highway traffic information publishing device; sampling a voltage time domain waveform and a current time domain waveform from the power quality monitoring data; transforming the voltage time domain waveform and the current time domain waveform into frequency domain features respectively; and determining the operating state of the device according to each time domain waveform and frequency domain feature, as well as historical data. This embodiment can realize real-time monitoring of the operating state of the highway traffic information publishing device.

Description

Technical Field

[0001] The embodiments of the present invention relate to the technical field of electromechanical equipment monitoring, and in particular to a method, equipment and medium for monitoring the status of highway traffic information publishing equipment. Background Art

[0002] Highway traffic information release equipment is a key highway electromechanical equipment. Its failure may have a significant impact on highway traffic operations, which may not only cause huge economic losses but also threaten the public's travel safety.

[0003] Currently, highway traffic information release equipment is still lacking in supervision, and maintenance mainly relies on manual discovery after damage, which is inefficient and cannot achieve real-time monitoring. Summary of the invention

[0004] The embodiments of the present invention provide a method, device and medium for monitoring the status of a highway traffic information publishing device to solve the above technical problems.

[0005] In a first aspect, an embodiment of the present invention provides a method for monitoring the state of a highway traffic information publishing device, comprising:

[0006] Obtain power quality monitoring data during the operation of highway traffic information release equipment;

[0007] Sampling a voltage time domain waveform and a current time domain waveform from the power quality monitoring data;

[0008] The voltage time domain waveform and the current time domain waveform are respectively transformed into frequency domain features;

[0009] The operating status of the device is determined based on the time domain waveforms and frequency domain characteristics, as well as historical data.

[0010] In a second aspect, an embodiment of the present invention provides an electronic device, the electronic device comprising:

[0011] one or more processors;

[0012] a memory for storing one or more programs,

[0013] When the one or more programs are executed by the one or more processors, the one or more processors implement the highway traffic information release equipment status monitoring method described in any embodiment.

[0014] In a third aspect, an embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the method for monitoring the status of highway traffic information release equipment described in any embodiment is implemented.

[0015] The embodiment of the present invention provides a method for monitoring the status of highway traffic information publishing equipment, which automatically monitors the operating status of the equipment based on the power quality monitoring data during the operation of the traffic information publishing equipment. Specifically, the power quality monitoring technology is first used to collect the operating status parameters of the power system in real time, and then the characteristic parameters in the real-time power monitoring information are extracted through time domain and frequency domain analysis, and the relevant power status rules in the database are compared. According to the power quality monitoring historical data of the traffic information publishing equipment, the operating status of the equipment is determined. This method can realize real-time monitoring of the operating status of highway traffic information publishing equipment and improve the good operating rate of highway traffic information publishing equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 It is a main structural diagram of a variable information board provided by an embodiment of the present invention;

[0018] Figure 2 It is a diagram of a light-emitting component of a variable information sign provided by an embodiment of the present invention;

[0019] Figure 3 It is a flow chart of a method for monitoring the state of a highway traffic information publishing device provided by an embodiment of the present invention;

[0020] Figure 4 It is a hardware design diagram of a traffic information release equipment status monitoring device provided by an embodiment of the present invention;

[0021] Figure 5 It is a schematic diagram of a deployment environment of a traffic information release equipment status monitoring device provided by an embodiment of the present invention;

[0022] Figure 6 This is a schematic diagram of normal working measurement values ​​of an F information board display screen provided by an embodiment of the present invention;

[0023] Figure 7 It is a schematic diagram of working measurement values ​​of a small number of lamp group failures of an F information board display screen provided by an embodiment of the present invention;

[0024] Figure 8 It is a schematic diagram of working measurement values ​​of a large number of lamp group failures of an F information board display screen provided by an embodiment of the present invention;

[0025] Fig. 9It is a structural schematic diagram of a fault detection model corresponding to a fault state provided by an embodiment of the present invention;

[0026] Fig.10 is a structural schematic diagram of a fault detection model corresponding to another fault state provided by an embodiment of the present invention;

[0027] Fig.11 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be described clearly and completely below. 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 belong to the scope of protection of the present invention.

[0029] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0030] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0031] As described in the background technology, highway traffic information release equipment (including information boards, variable speed limit signs, traffic lights, etc.) is an important equipment in the highway traffic system. Once the equipment fails, it will have a great impact on the operation of highway traffic and endanger driving safety. The embodiment of the present invention starts with the structural principle of the highway traffic information release equipment, analyzes the fault information relationship between the equipment "signal-data-phenomenon-equipment status", analyzes the fault type of the highway traffic equipment, and matches different status monitoring methods for different types of faults, completes the status monitoring and fault diagnosis of the traffic information release equipment, and improves the good operation rate of the highway traffic information release equipment.

[0032] Take variable information boards as an example. Figure 1 and Figure 2 They are respectively a main structure diagram and a light-emitting component diagram of a variable information board provided by an embodiment of the present invention. Figure 1 , Figure 2 The largest part of the variable information board is a large display screen with light-emitting devices as the basic display unit. Since the screen itself is generally of great mass and is installed on or beside the road surface of highways, bridges and other road traffic facilities, it is mainly divided into gantry-type variable information boards and F-type variable information boards according to the appearance structure. Other components include display control cards and their built-in software, power supply facilities, and lightning protection equipment. Variable information boards generally use optical fiber backbone networks to achieve network communication with the control center.

[0033] Variable information boards are important information release platforms on highways. Compared with other intelligent field monitoring equipment, variable information boards consume more power and are "big electricity consumers" among field monitoring equipment. The main faults of variable information boards are communication failures, power supply failures, and display hardware failures. Communication failures and power supply failures can be automatically monitored by monitoring the online status of the variable information board. Signs of display hardware failures include: the display does not display, the display displays partial information, the display displays garbled characters, etc. These signs can be automatically monitored by monitoring the power consumption history data and power quality history data of the traffic information release equipment during operation, and by comparing big data to automatically monitor the display hardware failures.

[0034] Figure 3 This is a flow chart of a method for monitoring the state of a highway traffic information publishing device provided by an embodiment of the present invention, which is applicable to the above-mentioned situation of automatically monitoring the state and faults of the device through the historical data of power quality, and is executed by an electronic device. Figure 3 As shown, the method specifically includes:

[0035] S110. Acquire power quality monitoring data during the operation of the highway traffic information release equipment.

[0036] Optionally, the power quality monitoring data of the equipment during operation can be collected from the branch of the power system that supplies power to the highway traffic information release equipment. Figure 4 The highway traffic information release equipment status monitoring device shown in the figure uses the ADE7880 chip as the main monitoring device. The device is connected to the branch of the power system that supplies power to the highway traffic information release equipment, and the two ends of the branch are respectively connected to the 7880+transformer circuit board, so that the real-time voltage signal and real-time current signal of the equipment can be collected.

[0037] S120: sampling a voltage time domain waveform and a current time domain waveform from the power quality monitoring data.

[0038] Optionally, multiple frames of voltage and current signals are extracted from the power quality monitoring data; and the extracted voltage and current signals are normalized respectively to obtain voltage time domain waveforms and current time domain waveforms with higher visibility.

[0039] S130, transforming the voltage time-domain waveform and the current time-domain waveform into frequency-domain features respectively.

[0040] Optionally, the voltage time domain waveform is subjected to Fourier transformation to obtain the frequency domain characteristics of the voltage, i.e., the spectrum sequence; the current time domain waveform is subjected to Fourier transformation to obtain the frequency domain characteristics of the current; and used for time domain analysis and frequency domain analysis of power quality data.

[0041] S140: Determine the operating status of the device according to the time domain waveforms and frequency domain characteristics, as well as historical data.

[0042] The power quality data of highway traffic information release equipment in different states are quite different. For example, Figure 5 As shown, the highway traffic information release equipment status monitoring device is connected to the branch of the power system that supplies power to the F information board. The fault of some light groups on the display screen can be simulated by disconnecting the power supply interface of the light group fault on the display screen, and the real machine fault simulation can be carried out. The highway traffic information release equipment and status monitoring device also include a display panel. After the monitoring device collects the power parameters of the F information board under various operating conditions, the key parameters can be displayed on the display panel. Figure 6 Displays the normal operating measurement values ​​of the F information board display. Figure 7 The F information panel display shows a small number of lamp group failure working measurement values, Figure 8A large number of light group fault working measurement values ​​of the F information board display screen are displayed. It can be seen that there are differences between the power parameters when a light group fault occurs on the F information board display screen and the normal working state. This step will perform time domain and frequency domain analysis on the power quality data collected in real time, compare the relevant power status rules in the database, and determine the unstable working state of the equipment (such as harmonics, voltage fluctuation flicker, pulses, oscillations, etc.) based on the historical data of power quality monitoring of the traffic information release equipment, and predict the operating state of the equipment. Optionally, depending on the historical data and analysis methods, this embodiment provides two optional implementation methods:

[0043] In the first optional implementation, the historical data includes representative voltage time domain waveforms, current time domain waveforms, voltage frequency domain features, and current frequency domain features of the device under various fault conditions. At this time, each time domain waveform and each frequency domain feature obtained in S130 can be compared with each representative time domain waveform and frequency domain feature under each fault condition. Specifically, the voltage amplitude in the current voltage time domain waveform is compared with the voltage amplitude in the representative voltage time domain waveform under each fault condition; the change frequency in the current voltage frequency domain feature is compared with the change frequency in the representative voltage frequency domain feature under each fault condition; the current amplitude in the current current time domain waveform is compared with the current amplitude in the representative current time domain waveform under each fault condition; the change frequency in the current current frequency domain feature is compared with the change frequency in the representative current frequency domain feature under each fault condition.

[0044] If all comparison differences corresponding to a fault state are less than the set threshold, it is determined that the device is in the fault state. That is, in the above four comparisons, if the difference between the two data in each comparison is less than the set threshold, it means that the current power quality data is similar enough to the power quality data in the fault state, and it can be determined that the device is in the fault state.

[0045] In a second optional implementation, the historical data includes the power quality data of the device in multiple time periods under various fault conditions. In this case, the voltage time domain waveform, current time domain waveform, voltage frequency domain characteristics and current frequency domain characteristics under the same fault state and from the same time period can together constitute a consistency sample, thereby obtaining a consistency sample library under various fault conditions; then, the voltage time domain waveform, current time domain waveform, voltage frequency domain characteristics and current frequency domain characteristics with inconsistent time periods or inconsistent fault conditions are respectively extracted from the multiple consistency samples to together constitute an inconsistent sample, thereby obtaining an inconsistent sample sample. Then, based on the above samples, the neural network-based state monitoring model is trained, and the trained model takes the voltage time domain waveform, current time domain waveform, voltage frequency domain characteristics and current frequency domain characteristics of the device in a certain time period as input, and takes the operating state or fault state of the device in the time period as output.

[0046] Optionally, the state monitoring model includes multiple neural network-based fault detection models, each fault detection model corresponds to a fault state, and each fault detection model is used to detect whether the device is in a corresponding fault state in a certain period of time. Fig. 9 is a structural diagram of a fault detection model corresponding to a fault state provided by an embodiment of the present invention, such as Fig. 9 As shown, the fault detection model includes four branches, wherein branch 1 composed of encoder 1 and decoder 1 is used to process the voltage time domain waveform, branch 2 composed of encoder 2 and decoder 2 is used to process the current time domain waveform, branch 3 composed of encoder 3 and decoder 3 is used to process the voltage frequency domain features, and branch 4 composed of encoder 4 and decoder 4 is used to process the current frequency domain features. The encoder in each branch is used to extract the deep features of the input signal, and the decoder is used to restore the input signal according to the deep features. Taking branch 1 as an example, encoder 1 is used to extract the deep features of the current time domain waveform, and decoder 1 is used to restore the current time domain waveform according to the deep features, and the remaining branches are similar. Furthermore, the encoder and decoder are inverse operations of each other, and a convolutional neural network structure or a multi-head self-attention structure can be used, which is not specifically limited in this embodiment.

[0047] Based on the above model structure, before state monitoring, the fault detection models corresponding to each fault state are first trained separately. Taking the fault detection model corresponding to one of the fault states as an example, this embodiment takes the fault state as the target fault state. During training, multiple consistent samples can be input into the fault detection model, and the four branches in the model are used to encode and decode each input signal respectively, and the model parameters are updated according to the following loss function:

[0048]

[0049] Where M represents the number of consistent samples of the same batch input fault detection model, i≠j, F i，k and F j，k denote the deep features extracted from branch i and branch j in the kth consistent sample, sim(F i，k , F j，k ) indicates F i，k and F j，k The similarity, sim(F i，k , F j，k ) has a value in the interval [0, 1]. represents the average similarity between the deep features of each two different branches (there are 12 combinations of two branches); P i，k and They represent the time domain waveform or frequency domain features of the input branch i in the kth consistent sample, and the restored signal output by branch i; α1 and β1 are the weight ratios of the loss terms. When updating the model parameters by minimizing L1, the loss term The average similarity in It can achieve maximization and ensure that the deep features extracted from the same consistent sample in each branch are similar enough. Since the sample data comes from the same period under the same fault state, whether it is voltage data, current data, time domain data, or frequency domain data, they all reflect the same device state. This embodiment maximizes the average similarity to ensure the consistency of state recognition. At the same time, when updating the model parameters by minimizing L1, the loss term It also achieves minimization, ensuring the restored signal With the original signal P i，k Similar enough to avoid losing important information of the original signal during the encoding process. α1 and β1 can be set as needed to balance the importance of the two loss terms.

[0050] At the same time, multiple inconsistent samples can also be input into the fault detection model, and the four branches in the model are used to encode and decode each input signal respectively, and the model parameters are updated according to the following loss function:

[0051]

[0052] Where M represents the number of inconsistent samples in the same batch input into the fault detection model, i≠j, F i，k and F j，k denote the deep features extracted from branch i and branch j in the kth inconsistent sample, sim(F i，k , F j，k ) indicates F i，k and F j，k The similarity, sim(F i，k , F j，k ) has a value in the interval [0, 1]; represents the average similarity between the deep features of each two different branches (there are 12 combinations of two branches); P i，k and They represent the time domain waveform or frequency domain features of the input branch i in the kth inconsistent sample, and the restored signal output by branch i; α2 and β2 are the weight ratios of the loss terms. When updating the model parameters by minimizing L2, the loss term Minimization can ensure that the deep features of each branch are quite different and match the inconsistency of the sample data; loss term It also achieves minimization, ensuring the restored signal With the original signal P i，k Similar enough to avoid losing important information of the original signal during the encoding process. α2 and β2 can be set as needed to balance the importance of the two loss terms.

[0053] At the same time, multiple consistent samples under the target fault state and multiple consistent samples under other states can also be input into the fault detection model, and the four branches in the model are used to encode and decode each input signal respectively, and the model parameters are updated according to the following loss function:

[0054]

[0055] Where M1 and M2 represent the number of consistent samples in the target fault state and the number of consistent samples in other states of the same batch input fault detection model respectively; m≠l, F m and F l They represent the concatenated features of the deep features extracted from each branch for the mth and lth consistent samples under the target fault state, respectively. (F m -F l ) 2 Represents the difference in features after concatenating the deep features of two different consistent samples in each branch under the target fault state; It represents the concatenation of the deep features extracted from each branch of the nth consistent sample in other states. represents the difference between the consistent samples under the target fault state and the consistent samples under other states after the deep features extracted from each branch are concatenated; P i，k and They represent the time domain waveform or frequency domain features of the input branch i in the kth consistent sample under any state, and the restored signal output by branch i; α3, β3 and c3 are the weight proportions of the loss terms. When updating the model parameters by minimizing L3, Minimization can ensure that the deep features corresponding to the consistent samples under the target fault state are similar enough; Maximization can ensure that there is a large difference between the deep features of the target fault state and other states, thereby playing a role in identifying the fault state; Also used to ensure the restored signal With the original signal P i，k Similar enough to avoid losing important information of the original signal during the encoding process. α3, β3 and c3 can be set as needed to balance the importance of the loss items.

[0056] In practical applications, we can first use consistent samples for one training (corresponding to loss function L1), and then use inconsistent samples for a second training (corresponding to loss function L2) based on the first training. Based on the second training, we can use consistent samples in the target state and consistent samples in other target states for a third training (corresponding to loss function L3). The above three trainings together constitute the initial training process of the model. We can also input consistent samples in the target fault state, consistent samples in other states, and inconsistent samples into the model (M, M1 and M2 are adjusted according to the actual number of samples of each type), and complete the initial training of the model by minimizing the loss function L = L1 + L2 + L3. The model after preliminary training can ensure that the deep features of consistent samples in the target fault state tend to be consistent, the deep features of consistent samples in the target fault state and other states are different enough, and the deep features extracted from the consistent samples in the four branches are relatively similar, while the average similarity of the deep features extracted from the inconsistent samples in the four branches is relatively low. This average similarity can characterize the reliability of the input data in reflecting the operating status of the equipment.

[0057] After completing the initial training, Fig. 9 In the model shown in Figure 1, the decoders of each branch are removed, and the feature concatenation layer, the fully connected layer and the activation function, as well as the similarity calculation layer are added to obtain the following: Fig.10 The model shown. Among them, the feature concatenation layer is used to concatenate the deep features of the four branches into one feature. The concatenation method is not limited. The concatenation method of the four branches is consistent and the same as the feature concatenation method in L3. The fully connected layer is used to continue deep operations on the concatenated deep features, and finally outputs the classification result of whether the input sample has the target fault state through the activation function. The similarity calculation layer is used to calculate the similarity of every two deep features in the deep features of the four branches, and outputs the average value of all similarities as the confidence of the above classification results. In order to make Fig.10 The model shown has the above functions. After completing the preliminary training, the encoder parameters of each branch can be fixed and each sample can be input Fig.10 The fully connected layer is trained with the target fault state as the label for each sample, so that the activation function can output the prediction result (consistent with the label) of whether each sample is in the target fault state, and the average similarity output by the similarity calculation layer is used as the confidence of the prediction result.

[0058] After training, Fig.10The model shown is used as the final fault detection model of the target fault state. When actually performing state monitoring, the voltage time domain waveform, current time domain model, voltage frequency domain characteristics and current frequency domain characteristics of the current period can be respectively input into each branch of the target fault state to obtain the prediction result of whether the device is in the target fault state in the current period, and the confidence of the prediction result. The confidence reflects the reliability of using the input data as the basis for judging the operation state of the device from the perspective of data consistency. If the confidence is less than the set threshold, it indicates that the input data of the current period may have misoperation in the data collection, data extraction, data matching and other links, resulting in inconsistent data characteristics, and the possibility of misjudgment is high. According to the same method, the voltage time domain waveform, current time domain model, voltage frequency domain characteristics and current frequency domain characteristics of the current period are respectively input into the final fault detection model of each other fault state, and the prediction result of whether the device is in each other fault state in the current period and the confidence of the prediction result can be obtained, thereby realizing equipment state monitoring.

[0059] In summary, this embodiment provides a method for monitoring the status of highway traffic information publishing equipment, starting from the power quality monitoring data during the operation of the traffic information publishing equipment, and automatically monitoring the operating status of the equipment. Specifically, firstly, the power quality monitoring technology is used to collect the operating status parameters of the power system in real time, and then the characteristic parameters in the real-time power monitoring information are extracted through time domain and frequency domain analysis, and the relevant power status rules in the database are compared. According to the power quality monitoring historical data of the traffic information publishing equipment, the operating status of the equipment is determined. This method can realize real-time monitoring of the operating status of highway traffic information publishing equipment and improve the good operating rate of highway traffic information publishing equipment.

[0060] Fig.11 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention is shown in FIG. Fig.11 As shown, the device includes a processor 60, a memory 61, an input device 62 and an output device 63; the number of processors 60 in the device can be one or more. Fig.11 A processor 60 is taken as an example; the processor 60, the memory 61, the input device 62 and the output device 63 in the device can be connected by a bus or other means. Fig.11 The example of connecting through bus is taken in the following.

[0061] The memory 61 is a computer-readable storage medium that can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the highway traffic information release device state monitoring method in the embodiment of the present invention. The processor 60 executes various functional applications and data processing of the device by running the software programs, instructions and modules stored in the memory 61, that is, realizing the above-mentioned highway traffic information release device state monitoring method.

[0062] The memory 61 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 61 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 61 may further include a memory remotely arranged relative to the processor 60, and these remote memories may be connected to the device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0063] The input device 62 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the device. The output device 63 may include a display device such as a display screen.

[0064] An embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method for monitoring the state of highway traffic information publishing equipment of any embodiment is implemented.

[0065] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, a device or a device or used in combination with it.

[0066] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0067] The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0068] Computer program code for performing the operations of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as C or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0069] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring the status of a highway traffic information publishing device, characterized in that: include: Obtain power quality monitoring data during the operation of highway traffic information release equipment; Sampling a voltage time domain waveform and a current time domain waveform from the power quality monitoring data; The voltage time domain waveform and the current time domain waveform are respectively transformed into frequency domain features; Extracting the voltage time domain waveform, current time domain waveform, voltage frequency domain feature and current frequency domain feature of the device in the same period under various fault conditions from the historical data to form a consistency sample; The neural network-based condition monitoring model is trained using each consistent sample, so that the output of the model continuously approaches the fault state of the device in each period; Input each time domain waveform and frequency domain feature of the current period into the trained model to obtain the current operating status of the device; Wherein, the state monitoring model includes a plurality of fault detection models based on neural networks, each fault detection model corresponds to a fault state, and each fault detection model is used to detect whether the device is in a corresponding fault state in a certain period of time; Each fault detection model includes four branches, each branch is used to process the voltage time domain waveform, the current time domain waveform, the voltage frequency domain characteristics and the current frequency domain characteristics; Each branch includes an encoder and a decoder, the encoder and the decoder are inverse operations of each other, the encoder is used to extract the depth features of the input signal, and the decoder is used to restore the input signal according to the depth features; The method of using the consistency samples to train the neural network-based state monitoring model so that the output of the model continuously approaches the fault state of the device in each time period includes: Extracting voltage time domain waveforms, current time domain waveforms, voltage frequency domain features, and current frequency domain features with inconsistent time periods or inconsistent fault states from multiple consistent samples, and jointly forming a non-consistent sample; Using a plurality of first consistent samples in any fault state, a plurality of second consistent samples in other states, and a plurality of inconsistent samples, a preliminary training is performed on a fault detection model corresponding to any fault state; during the preliminary training process, the average similarity of the depth features extracted from each branch of the same consistent sample is maximized, the average similarity of the depth features extracted from each branch of the same inconsistent sample is minimized, the difference between the first depth features of each first consistent sample is minimized, the difference between each first depth feature and the second depth feature of each second consistent sample is maximized, and the difference between the restored signal output by the decoder of each branch and the input signal of each branch is minimized; In the model after preliminary training, the decoders of each branch are removed, and a similarity calculation layer, a feature splicing layer and a fully connected layer are added, wherein the similarity calculation layer is used to calculate the average similarity of the deep features extracted from each branch of the same sample, the feature splicing layer is used to splice the deep features extracted from each branch of the same sample, and the fully connected layer is used to predict whether the current sample is in any of the fault states according to the spliced ​​features; The encoder parameters of each branch are fixed, and the fully connected layer is trained with whether each sample is in any of the fault states as a label, so that the fully connected layer can output a prediction result of whether each sample is in any of the fault states, and the average similarity output by the similarity calculation layer is used as the confidence of the prediction result.

2. The method according to claim 1, characterized in that: The obtaining of the power quality monitoring data during the operation of the highway traffic information release device includes: The power quality monitoring data of the highway traffic information release equipment during operation is collected from the branch of the power system that supplies power to the highway traffic information release equipment.

3. The method according to claim 1, characterized in that The step of sampling the voltage time domain waveform and the current time domain waveform from the power quality monitoring data comprises: Extracting multiple frames of voltage and current signals from the power quality monitoring data; The extracted voltage and current signals are normalized respectively to obtain the voltage time domain waveform and the current time domain waveform.

4. An electronic device, characterized in that: include: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the highway traffic information release equipment status monitoring method described in any one of claims 1-3.

5. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the program is executed by a processor, the method for monitoring the state of highway traffic information publishing equipment as described in any one of claims 1-3 is implemented.

Images