Gateway that uses AI models to automatically identify monitoring points

Abstract

The present invention discloses a gateway which adopts an AI model to automatically identify monitoring points, and relates to the technical field of artificial intelligence. The present invention comprises communication address acquisition, simulated PLC access, communication message capture, and point determination. The communication message of an existing data acquisition terminal on site is captured to determine which communication protocol is used, and then the message of the corresponding protocol is parsed to obtain the continuous address of the communication point and its value. Then, by simulating PLC, the parsed addresses are set to different values ​​in sequence or in batches, and the address and name of the specific point are determined by visually comparing the values ​​from the existing point display terminal or configuration on site, thereby improving the accuracy of point identification and the accuracy of the identification result.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a gateway that uses an AI model to automatically identify monitoring points. Background Art

[0002] The rise of the Internet of Things has led to a large number of devices and sensors being connected to the network, generating massive amounts of data. In order to effectively manage and process this data, the gateway, as a key device connecting the perception layer and the network layer, has become increasingly important. It is responsible for aggregating and forwarding sensor data, and realizing conversion between different protocols. However, with the increase in monitoring points and the increase in data complexity, traditional gateways face challenges in processing and identifying data sources. The use of AI models to automatically identify monitoring points is the result of the combined effect of technological development and industry needs, and has important practical significance. Therefore, it is particularly important to invent a gateway that uses AI models to automatically identify monitoring points.

[0003] The prior art, such as the invention application patent with announcement number: CN118801919A, discloses a fast SF identification method for a gateway, which includes the following steps: Step 1: First, the received signal is The method comprises the following steps: performing grouping, dividing the spreading factor from 6 to 12, and then calculating the intermediate variable value during the first search; step two: calculating the intermediate variable value when the spreading factor is equal to 6 and the intermediate variable value when the spreading factor is equal to 12; step three: comparing the intermediate variable value of the first search with the intermediate variable value when the spreading factor is equal to 6 and the intermediate variable value when the spreading factor is equal to 12, and outputting the result; step four: setting the spreading factor equal to the second search, comparing the intermediate variable value of the second search with the intermediate variable value when the spreading factor is equal to 6 and the intermediate variable value when the spreading factor is equal to 12, and outputting the result, and the process ends; the method solves the technical problems in the prior art of slow spreading factor identification and high complexity of threshold adjustment during spreading factor identification.

[0004] The existing technology still has the following defects, which are specifically reflected in: 1. In the existing technology, human experience is relied upon to identify monitoring points. When checking equipment information and judging point addresses, humans are easily affected by subjective factors such as fatigue and negligence, which leads to deviations in identification results. It is difficult to ensure that each point can be accurately identified and located. Abnormal situations that occur during communication cannot be detected in a timely and accurate manner, resulting in a decrease in identification accuracy.

[0005] 2. In the existing technology, there is a lack of clear steps and traceability. When problems such as point identification errors or communication failures occur, it is difficult to quickly and accurately locate the problem link. A comprehensive investigation of the entire system is required, which consumes a lot of time and energy, increases the difficulty and cost of troubleshooting, increases maintenance costs and equipment downtime, and reduces production efficiency. Summary of the invention

[0006] The purpose of the present invention is to provide a gateway that uses an AI model to automatically identify monitoring points, thereby solving the problems existing in the background technology.

[0007] In order to solve the above technical problems, the present invention adopts the following technical solution: The present invention provides a gateway that uses an AI model to automatically identify monitoring points, including: step one, obtaining a communication address.

[0008] Step 2: Simulate PLC access.

[0009] Step 3: Capture communication messages.

[0010] Step 4: Determine the point.

[0011] Preferably, the communication address is obtained by checking the on-site equipment configuration and the default serial port configuration of some equipment to obtain the serial port configuration, and determining the IP address and communication port of the PLC by network packet capture.

[0012] Preferably, the simulated PLC access is implemented by: based on the serial port configuration, using the simulated PLC software to create a simulated PLC with the corresponding configuration; based on the IP address of the PLC, setting the computer IP where the packet capture software is located to the IP of the PLC; opening the PLC simulation software to create the corresponding simulated PLC.

[0013] Preferably, the communication message capture is implemented by checking the communication log of the simulated PLC software to determine whether the communication is normal. If the communication is normal, a point analysis tool is used to capture the communication message for point analysis.

[0014] Preferably, the point determination is implemented as follows: when the field data acquisition terminal and the analog PLC can communicate normally and the packet capture tool is used to parse the point continuous address, the captured continuous address is set to an incomprehensible value, and the field point display interface is visually checked to determine the address and name of the specific point.

[0015] The beneficial effects of the present invention are: 1. In the present invention, through detailed steps, such as first obtaining the communication address, then simulating a series of operations such as PLC access, and finally determining the point, compared with the traditional reliance on manual experience judgment or a rough identification method, the specific address and name of each monitoring point can be determined more accurately, reducing errors caused by human negligence, inaccurate judgment, etc., improving the accuracy of point identification, and improving the accuracy of identification results.

[0016] 2. In the present invention, each step has a corresponding judgment basis and implementation method. When problems such as inaccurate point identification or communication abnormalities occur, it is possible to trace back and troubleshoot along each step to quickly locate whether there is an error in the communication address acquisition link, or the simulated PLC access is abnormal, or there is a problem with the message capture, etc., which reduces the difficulty and cost of troubleshooting, reduces maintenance costs and equipment downtime, and increases production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 The present invention is a schematic flow chart of the steps for implementing the method.

[0019] Figure 2 Arrange the graph for the data.

[0020] Figure 3 A forecast diagram. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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 are within the scope of protection of the present invention.

[0022] Reference Figure 1 As shown, the present invention provides a gateway that uses an AI model to automatically identify monitoring points, including: step one, obtaining a communication address.

[0023] In a specific embodiment, the communication address is obtained by checking the on-site device configuration and the default serial port configuration of some devices, obtaining the serial port configuration, and determining the IP address and communication port of the PLC by network packet capture.

[0024] It should be noted that most devices using the modbus protocol have default serial port settings. When the default serial port settings cannot communicate normally, you can use the serial port oscilloscope method to determine the serial port configuration based on the real-time band. The serial port configuration includes serial port baud rate, data bit, check bit, stop bit, etc. The network packet capture method is specifically implemented as follows: connect the computer where the packet capture tool is located to the PLC LAN, and then disconnect the PLC network. At this time, an ARP message searching for the PLC status will be generated, which contains the IP address of the PLC, and the IP address of the PLC can be determined.

[0025] In the present invention, through detailed steps, such as first obtaining the communication address, then simulating a series of operations such as PLC access, and finally determining the point, compared with the traditional reliance on manual experience judgment or a rough identification method, the specific address and name of each monitoring point can be determined more accurately, reducing errors caused by human negligence, inaccurate judgment, etc., improving the accuracy of point identification, and improving the accuracy of identification results.

[0026] Step 2: Simulate PLC access.

[0027] In a specific embodiment, the simulated PLC access is implemented by: based on the serial port configuration, using the simulated PLC software to create a simulated PLC with the corresponding configuration, based on the IP address of the PLC, setting the IP of the computer where the packet capture software is located to the IP of the PLC, opening the PLC simulation software, and creating the corresponding simulated PLC. Normal communication with the field acquisition terminal can be achieved.

[0028] It should be noted that the simulated PLC, including the baud rate, data bit, check bit, and stop bit, can communicate normally with the field acquisition terminal after being connected to the simulated PLC host using a serial port line.

[0029] Step 3: Capture communication messages.

[0030] In a specific embodiment, the communication message capture is implemented by checking the communication log of the simulated PLC software to determine whether the communication is normal. If the communication is normal, a point analysis tool is used to capture the communication message for point analysis.

[0031] It should be noted that the specific analysis method for judging whether the communication is normal is as follows: check the communication log of the simulated PLC software. If there is continuous communication data interaction, the communication is judged to be normal. Otherwise, the communication is judged to be abnormal. The point analysis tool is used to capture the communication message for point analysis. The implementation method is as follows: a specific wiring method is required for serial communication packet capture. The serial port line of the computer where the packet capture tool is located is connected in parallel with the simulated PLC or the acquisition end, and special equipment such as hubs cannot be used. During network communication, the NPACP tool is used to capture the network communication data packets between the simulated PLC and the acquisition end in real time.

[0032] Step 4: Determine the point.

[0033] In a specific embodiment, the point determination is implemented as follows: when the field data acquisition terminal and the analog PLC can communicate normally and the packet capture tool is used to parse the point continuous address, the captured continuous address is set to an incomprehensible value, and the field point display interface is visually checked to determine the address and name of the specific point.

[0034] In the present invention, each step has a corresponding judgment basis and implementation method. When problems such as inaccurate point identification or communication abnormalities occur, it is possible to trace back and troubleshoot along each step to quickly locate whether there is an error in the communication address acquisition link, or the simulated PLC access is abnormal, or there is a problem with the message capture, etc., which reduces the difficulty and cost of troubleshooting, reduces maintenance costs and equipment downtime, and increases production efficiency.

[0035] The main idea of ​​the algorithm is: first, organize the variable names of each project into the data pair format required by the model; use the open source model to train the data set; and finally predict by comparing the variable names with all labels one by one to get the score. The label with the largest score is the one to which it belongs.

[0036] nli-roberta-base is a natural language inference (NLI) model based on RoBERTa. It is specially trained for natural language inference tasks, aiming to judge the logical relationship between two sentences, and retrains the large model based on the RoBERTa large model to enable it to recognize the meaning of industry professional terms.

[0037] refer to Figure 2 For example, the big model does not know the relationship between "semi-enclosed screw parallel unit", "storage temperature online" and "start-up signal". After training the model, it can be known that "start-up signal" belongs to "semi-enclosed screw parallel unit" and "storage temperature online" does not.

[0038] The data is organized into the data pair format required by the model, 0 indicates matching, 1 indicates neutrality, and 2 indicates mismatching. The model's embedding vector can be trained through different matching relationships so that the two vectors calculated by the label corresponding to name and 0 are as close as possible. For example, "start-up signal_1" and "semi-enclosed screw parallel unit (1-3)" are matching relationship 0, "start-up signal_1" and "semi-enclosed screw parallel unit (4-6)" are neutral relationship 1, and "start-up signal_1" and "frozen food cold storage room 4" are mismatching relationship 2.

[0039] The model training, open source model nli-roberta-base, performs word segmentation, word vectorization, Transformer encoder encoding, aggregate embedding vectors on two Chinese sentences, and calculates the cosine similarity of the two embedding vectors. The larger the value, the more similar they are. This model is based on general knowledge training, and cannot accurately understand the semantics of the cold storage industry. It is necessary to train its word vectors with specific vocabulary to make it understand the semantics of the cold storage industry.

[0040] The prediction, reference Figure 3 , input the variable and the label into the model at the same time, and the relationship between the variable and the label will be obtained, corresponding to 0, 1, and 2. The second value has the largest explanation relationship of 1, which means neutrality, and the first value has the largest explanation relationship of 0, which means matching, and finally determines whether the variable belongs to the label.

[0041] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they shall all fall within the protection scope of the present invention.

Claims

1. The AI ​​model is used to automatically identify the gateway of the monitoring point, which is characterized by: include: Step 1: Obtain the communication address; Step 2: Simulate PLC access; Step 3: Capture communication messages; Step 4: Determine the point.

2. The gateway for automatically identifying monitoring points using an AI model according to claim 1, characterized in that: The communication address is obtained by checking the on-site equipment configuration and the default serial port configuration of some equipment, obtaining the serial port configuration, and determining the IP address and communication port of the PLC by network packet capture.

3. The gateway for automatically identifying monitoring points using an AI model according to claim 1, characterized in that: The simulated PLC access is implemented as follows: based on the serial port configuration, use the simulated PLC software to create a simulated PLC with the corresponding configuration; based on the IP address of the PLC, set the IP of the computer where the packet capture software is located to the IP of the PLC; open the PLC simulation software to create the corresponding simulated PLC.

4. The gateway for automatically identifying monitoring points using an AI model according to claim 1, characterized in that: The communication message capture is implemented by checking the communication log of the simulated PLC software to determine whether the communication is normal. If the communication is normal, the point analysis tool is used to capture the communication message for point analysis.

5. The gateway for automatically identifying monitoring points using an AI model according to claim 1, characterized in that: The point determination is implemented as follows: when the field data acquisition terminal and the simulation PLC can communicate normally and the packet capture tool is used to parse the point continuous address, the captured continuous address is set to an incomprehensible value, and the field point display interface is visually checked to determine the address and name of the specific point.

Images