Edge heterogeneous internet of things operation and maintenance case data processing method and system

Abstract

The application provides an edge heterogeneous Internet of Things operation and maintenance case data processing method and system. The method comprises the following steps: first, according to the types of multiple source heterogeneous terminals, a unified data collection scheme is formulated, a unified collection data format is defined, an asynchronous multi-thread scheduling mechanism is adopted, and the data collection tasks of each multiple source heterogeneous terminal are managed and executed respectively to collect and generate corresponding collection raw data. The collection raw data in a normal sampling state is standardized to generate a multiple source collection data set. Then, according to the type of each multiple source heterogeneous terminal, the standardized abnormal score of the corresponding multiple source collection data set is calculated, and weighted comprehensive processing is performed to generate a comprehensive abnormal score, and the abnormal grade is determined according to the comprehensive abnormal score. Finally, the abnormal grade of the multiple source heterogeneous terminal is matched with the preset local protection strategy, the local abnormal protection action is executed based on the matching result, the abnormal information is recorded, and the abnormal alarm information is reported to the remote center platform.

Description

TECHNICAL FIELD

[0001] The present application relates to data processing technology, and in particular to an edge heterogeneous Internet of Things operation and maintenance machine case data processing method and system. BACKGROUND

[0002] As an important carrier of edge computing and front-end sensing, the Internet of Things operation and maintenance machine case is widely used in the oil and petrochemical industry, power industry, mining industry, smart city, water industry, highway industry, and rail transit industry, and mainly undertakes tasks such as front-end device power supply and distribution, sensor data acquisition, and edge communication management. At present, various types of Internet of Things terminals deployed in the field have the characteristics of multi-source heterogeneity, diverse communication protocols, and non-uniform data structures. Existing solutions cannot efficiently access and accurately identify operational abnormalities under a unified architecture. In particular, when terminal devices experience communication interruption, state fluctuation, or sensor abnormalities, there is a lack of fine-grained abnormality identification mechanisms and differentiated local response strategies, which can easily lead to hidden trouble accumulation or misoperation, affecting the stable operation of the system and the efficiency of remote operation and maintenance. SUMMARY

[0003] The present application provides an edge heterogeneous Internet of Things operation and maintenance machine case data processing method and system, which solves the problem of non-uniform state monitoring of multi-source heterogeneous terminals, low abnormality identification accuracy, and lack of local rapid response capability.

[0004] In a first aspect, the present application provides an edge heterogeneous Internet of Things operation and maintenance machine case data processing method, comprising:

[0005] Multi-source heterogeneous terminal data acquisition and preprocessing, according to the type of multi-source heterogeneous terminal, a unified data acquisition scheme is formulated, a unified data acquisition format is defined, an asynchronous multi-thread scheduling mechanism is adopted, and the data acquisition tasks of each multi-source heterogeneous terminal are managed and executed respectively, corresponding raw data is collected and generated, and standardized processing is performed on the raw data with normal sampling state to generate a multi-source data set with unified structure;

[0006] Comprehensive abnormality evaluation, according to the type of each multi-source heterogeneous terminal, the standardized abnormality score is calculated for the corresponding multi-source data set, and weighted comprehensive processing is performed to generate a comprehensive abnormality score, and the abnormality level is determined according to the comprehensive abnormality score;

[0007] Abnormal response and local protection decision, the abnormality level of the multi-source heterogeneous terminal is matched with the preset local protection strategy, based on the matching result, the local abnormality protection actions including the early warning mechanism, the local protection measures and the forced local protection measures are executed, the abnormal information is recorded and the abnormal alarm information is reported to the remote center platform.

[0008] In a possible design, the unified data acquisition scheme comprises:

[0009] setting a sampling period, a communication protocol parameter and a communication timeout threshold for each of the plurality of source heterogeneous terminal types;

[0010] configuring a unified collection task list based on the sampling period, the communication protocol parameter and the communication timeout threshold.

[0011] In a possible design, the unified collection data format comprises a terminal unique identifier, a plurality of source heterogeneous terminal types, a set of sampling values, a timestamp and a collection state.

[0012] The set of sampling values comprises a sampling value and a corresponding sampling value unique identifier.

[0013] In a possible design, the comprehensive anomaly evaluation comprises:

[0014] determining a static normalized anomaly score for each of the sampling values in the multi-source collection data set;

[0015] determining a dynamic fluctuation normalized anomaly score for continuous sampling values and enumeration sampling values respectively based on a dynamic fluctuation sliding window for each of the sampling values in the multi-source collection data set;

[0016] selecting a larger value of the static normalized anomaly score and the dynamic fluctuation normalized anomaly score as a single sampling item anomaly score of the sampling value for each sampling value;

[0017] generating a weighted comprehensive anomaly score of each of the plurality of source heterogeneous terminals by performing weighted comprehensive processing based on each of the single sampling item anomaly scores and an importance weight of each of the sampling values;

[0018] determining an anomaly level of the plurality of source heterogeneous terminals based on an anomaly level classification rule according to the weighted comprehensive anomaly score.

[0019] In a possible design, a rule for determining the static normalized anomaly score is as follows:

[0020] when the sampling value is within a corresponding static normal threshold interval, the static normalized anomaly score of the sampling value is 0;

[0021] when the sampling value is outside the corresponding static normal threshold interval, a static normalized anomaly score of the sampling value is calculated;

[0022] a calculation formula of the static normalized anomaly score is as follows:

[0023]

[0024] wherein, i is a sampling value unique identifier, is a static normalized anomaly score of a sampling value unique identifier i, is a sampling value uniquely identified as i, is a sampling value uniquely identified as i, is a latest static normal threshold.

[0025] In a possible design, a rule for determining a dynamic fluctuation normalized anomaly score of the continuous sampling value is:

[0026] When the sliding standard deviation is less than or equal to the sliding standard deviation upper limit value, the dynamic fluctuation normalized anomaly score of the continuous sampling value is 0.

[0027] When the sliding standard deviation is greater than the sliding standard deviation upper limit value, the dynamic fluctuation normalized anomaly score of the continuous sampling value is calculated.

[0028] The calculation formula of the dynamic fluctuation normalized anomaly score of the continuous sampling value is:

[0029]

[0030] wherein i is a sampling value unique identifier, is a dynamic fluctuation normalized anomaly score of a sampling value uniquely identified as i, is a sliding standard deviation, is a sliding standard deviation upper limit value.

[0031] In a possible design, a calculation formula of a dynamic fluctuation anomaly score of the enumeration sampling value is:

[0032]

[0033] wherein i is a sampling value unique identifier, is a dynamic fluctuation normalized anomaly score of a sampling value uniquely identified as i, indicates 1 if the adjacent two sampling values are different, or 0 otherwise.

[0034] In a possible design, the anomaly level includes normal, mild anomaly, moderate anomaly, and severe anomaly.

[0035] The local protection policy is to trigger a local anomaly protection measure and an anomaly reporting mechanism when the anomaly level is mild anomaly, moderate anomaly, and severe anomaly.

[0036] The local anomaly protection measure includes:

[0037] When the anomaly level is mild anomaly, the frequency of the sampling period is adjusted, and a warning mechanism is started.

[0038] When the anomaly level is moderate anomaly, the local protection measure is executed.

[0039] When the abnormality level is a severe abnormality, a mandatory local protection measure is performed;

[0040] The abnormality reporting mechanism comprises:

[0041] When the abnormality level is a mild abnormality, abnormality information is recorded in a local abnormality log, the abnormality information is only synchronized to a local edge server or a regional management node, and is not immediately reported to a remote center platform;

[0042] When the abnormality level is a moderate abnormality, abnormality information is recorded in a local abnormality log, abnormality alarm information is reported to a remote center platform within a set short-time abnormality reporting window, and is identified as a middle-level early warning level;

[0043] When the abnormality level is a severe abnormality, abnormality information is recorded in a local abnormality log, and abnormality alarm information is immediately reported to a remote center platform, and is identified as a high-level alarm level.

[0044] In a second aspect, the application provides an edge heterogeneous Internet of Things operation and maintenance case data processing system, comprising:

[0045] A multi-source heterogeneous terminal data acquisition and preprocessing module formulates a unified data acquisition scheme, defines a unified acquisition data format, adopts an asynchronous multi-thread scheduling mechanism, manages and executes data acquisition tasks of each multi-source heterogeneous terminal respectively, acquires corresponding acquisition raw data, performs standardization processing on the acquisition raw data in a normal sampling state, and generates a multi-source acquisition data set with unified structure;

[0046] A comprehensive abnormality evaluation module calculates a standardized abnormality score for the corresponding multi-source acquisition data set according to each multi-source heterogeneous terminal type, performs weighted comprehensive processing, generates a comprehensive abnormality score, and determines an abnormality level according to the comprehensive abnormality score;

[0047] An abnormality response and local protection decision module performs corresponding matching of the abnormality level of the multi-source heterogeneous terminal and a preset local protection strategy, executes local abnormality protection actions including an early warning mechanism, a local protection measure and a mandatory local protection measure based on the matching result, records abnormality information and reports abnormality alarm information to a remote center platform.

[0048] The edge heterogeneous Internet of Things operation and maintenance case data processing method and system provided by the application realize accurate evaluation of the running state of each multi-source heterogeneous terminal by introducing a dual normalization score mechanism of sampling value static normalization abnormality score and dynamic fluctuation normalization abnormality score, and combining sampling value importance weight to calculate the comprehensive abnormality score of each multi-source heterogeneous terminal, which significantly improves the accuracy and robustness of abnormality identification.

[0049] Meanwhile, according to the local protection strategy matched with the type of multi-source heterogeneous terminals and the difference of abnormal levels, the application has the abilities of hierarchical response, dynamic adjustment and traceable reporting, and can automatically complete the lightweight operation and maintenance response and fault isolation on the edge side without the intervention of the remote center platform, thereby enhancing the rapid response ability to the abnormality and improving the stability of system operation. BRIEF DESCRIPTION OF DRAWINGS

[0050] The accompanying drawings, which are incorporated herein and form a part of the specification, illustrate embodiments consistent with the present application and, together with the description, further serve to explain the principles of the application.

[0051] Figure 1 is a flowchart of a data processing method of an edge heterogeneous Internet of Things operation and maintenance cabinet according to an example embodiment of the application.

[0052] Figure 2 is a flowchart of multi-source heterogeneous terminal data acquisition and preprocessing according to an example embodiment of the application.

[0053] Figure 3 is a flowchart of comprehensive abnormality evaluation according to an example embodiment of the application.

[0054] Figure 4 is a flowchart of abnormality response and local protection decision according to an example embodiment of the application.

[0055] Figure 5 is a structural diagram of a data processing system of an edge heterogeneous Internet of Things operation and maintenance cabinet according to an example embodiment of the application.

[0056] The above-described drawings have shown the specific embodiments of the application, and the following will have a more detailed description. These drawings and the written description are not intended to limit the scope of the concept of the application by any means, but to illustrate the concept of the application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0057] The example embodiments will be described in detail hereinafter with reference to the attached drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following example embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of apparatuses and methods consistent with some aspects of the present application as detailed in the appended claims.

[0058] Figure 1 is a flowchart of a data processing method of an edge heterogeneous Internet of Things operation and maintenance cabinet according to an example embodiment of the application. As shown in Figure 1As shown, the method provided by the embodiment includes:

[0059] Step S101: Multi-source heterogeneous terminal data acquisition and preprocessing, according to the type of multi-source heterogeneous terminal, a unified data acquisition scheme is formulated, a unified data acquisition format is defined, an asynchronous multi-thread scheduling mechanism is adopted, and the data acquisition task of each multi-source heterogeneous terminal is managed and executed respectively, the corresponding acquisition raw data is collected and generated, the acquisition raw data with normal sampling state is standardized, and the multi-source acquisition data set with unified structure is generated.

[0060] In this step, the type of multi-source heterogeneous terminal includes a wireless communication link terminal, a wired serial port terminal, a digital quantity input and output terminal, an analog quantity input and output terminal, and a power distribution terminal.

[0061] The wireless communication link terminal includes a long-distance microwave terminal, a 4G wireless communication terminal, a CDMA wireless communication terminal, and a GPRS wireless communication terminal.

[0062] The wired serial port terminal includes a 4-port RS485 terminal, an 8-port RS484 terminal, and a 4-port RS232.

[0063] The digital quantity input and output terminal includes an 8-channel digital input terminal (8DI terminal), an 8-channel digital output terminal (8DO terminal), and a 4-channel digital input + 4-channel digital output terminal (4DI4DO terminal).

[0064] The analog quantity input and output terminal includes an 8-channel analog input terminal (8AI terminal), an 8-channel analog output terminal (8AO terminal), a 4-channel analog input + 4-channel analog output terminal (4AI4AO terminal).

[0065] The power distribution terminal includes an AC220V power distribution terminal, an AC24V power distribution terminal, a DC48V power distribution terminal, a DC36V power distribution terminal, a DC24V power distribution terminal, and a DC12V power distribution terminal.

[0066] Step S102: Comprehensive anomaly evaluation, according to each type of multi-source heterogeneous terminal, the standardized anomaly score is calculated for the corresponding multi-source acquisition data set, and the weighted comprehensive processing is performed to generate the comprehensive anomaly score, and the anomaly level is determined according to the comprehensive anomaly score.

[0067] Step S103: Abnormal response and local protection decision, the anomaly level of the multi-source heterogeneous terminal is matched with the preset local protection strategy, the local anomaly protection action including the early warning mechanism, the local protection measure and the forced local protection measure is executed based on the matching result, the abnormal information is recorded and the abnormal alarm information is reported to the remote center platform.

[0068] Figure 2is a flowchart of a multi-source heterogeneous terminal data collection and preprocessing process according to an example embodiment of the present application. As shown in Figure 2 The method provided by the embodiment includes the following steps.

[0069] Step S1011: Formulate a unified data collection scheme according to the types of multi-source heterogeneous terminals and collection requirements.

[0070] In this step, the unified data collection scheme includes setting a sampling period, communication protocol parameters and a communication timeout threshold for each type of multi-source heterogeneous terminal, and configuring a unified collection task list based on the sampling period, communication protocol parameters and communication timeout threshold.

[0071] The configuration of the unified collection scheme ensures the scheduling standardization of multi-source heterogeneous terminal collection tasks and provides a unified basis for subsequent concurrent collection and standardized processing.

[0072] The setting of the sampling period and the communication timeout threshold is as follows:

[0073] For a wireless communication link terminal, the sampling period ranges from 10 seconds to 60 seconds, and the communication timeout threshold ranges from 1 second to 3 seconds.

[0074] For a wired serial port terminal, the sampling period ranges from 30 seconds to 120 seconds, and the communication timeout threshold ranges from 1 second to 5 seconds.

[0075] For a digital input / output terminal, the sampling period ranges from 1 second to 10 seconds, and the communication timeout threshold ranges from 0.5 second to 1 second.

[0076] For an analog input / output terminal, the sampling period ranges from 5 seconds to 30 seconds, and the communication timeout threshold ranges from 0.5 second to 2 seconds.

[0077] For a power distribution terminal, the sampling period ranges from 10 seconds to 60 seconds, and the communication timeout threshold ranges from 1 second to 5 seconds.

[0078] The setting of the communication protocol parameters is used to ensure that the Internet of Things operation and maintenance machine case can correctly establish a communication connection with each type of multi-source heterogeneous terminal and smoothly collect data. Therefore, according to the different types of multi-source heterogeneous terminals, the contents of the communication protocol parameters are also different, including:

[0079] For a wireless communication link terminal, the communication protocol parameters include:

[0080] The wireless communication protocol type includes, but is not limited to, LTE, CDMA or GPRS wireless communication protocol.

[0081] Authentication information, including but not limited to Access Point Name (APN), International Mobile Subscriber Identity (IMSI).

[0082] Wired serial port terminal, the communication protocol parameters include:

[0083] Serial port protocol type, including but not limited to Modbus RTU.

[0084] Communication rate (Baud Rate), unit: bps (bits per second).

[0085] Frame structure parameters, including but not limited to data bits, check bits, stop bits.

[0086] Register address mapping table, used to define the reading or writing position relationship of the device register.

[0087] Digital input and output terminal, the communication protocol parameters include:

[0088] General input and output channel number mapping table (GPIO Channel Mapping Table), used to calibrate the function and address of each input and output channel.

[0089] Digital communication mode, including but not limited to GPIO (General Purpose Input Output) coding, or fixed-length data frame structure protocol based on RS485 interface.

[0090] Analog input and output terminal, the communication protocol parameters include:

[0091] Analog-digital conversion interface protocol, including but not limited to ADC (Analog-to-Digital Converter) or DAC (Digital-to-Analog Converter).

[0092] Sampling resolution (Resolution), unit: bit.

[0093] Reference voltage configuration (Reference Voltage), used to determine the sampling voltage range.

[0094] Power distribution terminal, the communication protocol parameters include:

[0095] Communication protocol type, including but not limited to Modbus TCP or SNMP (Simple Network Management Protocol).

[0096] IP address and communication port number of power distribution terminal, used for defining IP address and communication port number of power distribution terminal in local area network or wide area network.

[0097] Object Identifier (OID), used for identifying access path and data reading item of each functional module of power distribution terminal.

[0098] Step S1012: defining unified collection data format.

[0099] In this step, according to terminal type and collection demand, unified collection data format is defined, so that data of different types of multi-source heterogeneous terminals have unified field composition and structure order.

[0100] The unified collection data format includes terminal unique identifier, multi-source heterogeneous terminal type, sample value set, time stamp and collection state.

[0101] The sample value set includes sample value and corresponding sample value unique identifier.

[0102] The interpretation of each field in the unified collection data format is shown as follows:

[0103] Terminal unique identifier: an identifier for uniquely identifying each multi-source heterogeneous terminal, used for ensuring accurate positioning to specific multi-source heterogeneous terminal in data collection processing and abnormal information reporting process.

[0104] Multi-source heterogeneous terminal type: including wireless communication link terminal, wired serial port terminal, digital input / output terminal, analog input / output terminal and power distribution terminal.

[0105] Sample value set: a two-dimensional array set, including sample value and corresponding sample value unique identifier, and sample value is different according to different multi-source heterogeneous terminal type. The sample value of each multi-source heterogeneous terminal type is shown as follows:

[0106] Wireless communication link terminal: including RTT, packet loss rate, bandwidth utilization rate, link state and signal quality.

[0107] Wired serial port terminal: including working voltage, running power, running state and communication state.

[0108] Digital input / output terminal: including switch state and output state.

[0109] Analog input / output terminal: including continuous physical quantity data. In this embodiment, including temperature, water level and gas concentration.

[0110] The power distribution terminal includes output voltage, output current and out-of-limit abnormality alarm.

[0111] In the embodiment, according to the functional characteristics and operation state monitoring requirements of each type of multi-source heterogeneous terminal, the core sampling value directly reflecting the working state of each multi-source heterogeneous terminal type is selected. The selected sampling value set is an important indicator for determining whether each multi-source heterogeneous terminal is in a normal working state, ensuring the accuracy and efficiency of abnormality detection.

[0112] The expression of the sampling value set is:

[0113]

[0114] Among them, is the unique identification of the sampling value; is the unique identification of the sampling value is the sampling value of

[0115] RTT, packet loss rate, bandwidth utilization rate, link state, signal quality, working voltage, running power, running state, communication state, switch state, output state, continuous physical quantity data, voltage, output current and out-of-limit abnormality alarm. RTT, packet loss rate, bandwidth utilization rate, link state, signal quality, working voltage, running power, running state, communication state, switch state, output state, continuous physical quantity data, voltage, output current and out-of-limit abnormality alarm.

[0116] RTT, packet loss rate, bandwidth utilization rate, link state, signal quality, working voltage, running power, running state, communication state, switch state, output state, continuous physical quantity data, voltage, output current and out-of-limit abnormality alarm. RTT, packet loss rate, bandwidth utilization rate, link state, signal quality, working voltage, running power, running state, communication state, switch state, output state, continuous physical quantity data, voltage, output current and out-of-limit abnormality alarm.

[0117] Timestamp: records the standard timestamp corresponding to the multi-source heterogeneous terminal data collection.

[0118] Collection state: indicates the state result of each multi-source heterogeneous terminal data collection, including normal sampling and abnormal sampling.

[0119] Step S1013: based on the asynchronous multi-thread scheduling mechanism, periodically collect multi-source heterogeneous terminal data to generate collected raw data.

[0120] In this step, the thread pool is initialized based on the generated unified collection task list, and the data collection task of each multi-source heterogeneous terminal is allocated to an independent thread according to the type of multi-source heterogeneous terminal. Each independent thread internally starts the data collection task asynchronously in a timer manner according to the set sampling period, completes the data interaction according to the corresponding communication protocol parameters, and waits for the response within the set communication timeout threshold, completes the periodic data collection task, and generates the collected raw data.

[0121] Step S1014: Validity verification is performed on each of the collected raw data, and the collection state is marked.

[0122] In this step, validity detection is performed on each of the collected raw data, and the validity detection includes:

[0123] Communication detection: whether the communication between the Internet of Things operation machine case and the multi-source heterogeneous terminal is successful. If it is successful, the communication detection is successful, otherwise it is failed.

[0124] Effective value detection: whether each of the collected values in the sample value set is an effective value. If all the sample values are effective values, the effective value detection is successful, otherwise it is failed.

[0125] Timestamp detection: comparing the timestamp of the collected value with the timestamp of the previous collected value, if the interval meets the set sampling period and the timestamp does not appear to be reversed, it is determined that the timestamp is normal, otherwise it is failed.

[0126] In the detection process, if any one of the communication detection, the effective value detection, and the timestamp detection is failed, the collection state of the collected raw data is marked as sampling abnormal. Otherwise, the collection state of the collected raw data is marked as sampling normal.

[0127] Step S1015: For the collected raw data marked as sampling normal, standardization processing is performed to generate a multi-source collected data set.

[0128] In this step, for each of the collected raw data marked as sampling normal, field encapsulation is performed according to a uniformly defined encapsulation template to generate standardized collected data.

[0129] The field encapsulation includes: encapsulating the terminal unique identifier, the multi-source heterogeneous terminal type, the sample value set, the timestamp, and the collection state information in sequence according to a uniform field order.

[0130] After the field encapsulation is completed, the standardized collected data is uniformly organized according to a preset data output format to form a multi-source collected data set with consistent structure. The multi-source collected data set is output in a standardized JSON format to ensure the normativity and parsing consistency of the data structure.

[0131] Figure 3 is a flowchart of comprehensive abnormality evaluation according to an example embodiment of the present application. As shown in Figure 3 The method provided by the embodiment includes:

[0132] Step S1021: For each of the sample values in the multi-source collected data set, a static normalized abnormality score is determined.

[0133] In this step, the rule for determining the static normalized anomaly score is:

[0134] When the sampling value is within the corresponding static normal threshold interval, the static normalized anomaly score of the sampling value is 0.

[0135] When the sampling value is outside the corresponding static normal threshold interval, the static normalized anomaly score of the sampling value is calculated.

[0136] The formula for calculating the static normalized anomaly score is:

[0137]

[0138] wherein the static normal threshold interval is , is the static normal lower limit value of the sampling value uniquely identified as i, is the static normal upper limit value of the sampling value uniquely identified as i, and is a preset value, which is different according to different types of multi-source heterogeneous terminals; and are also different; i is the unique identification of the sampling value; is the static normalized anomaly score of the sampling value uniquely identified as i; is the sampling value uniquely identified as i; is the static normal threshold closest to the sampling value , i.e. the value of or the value of

[0139] The setting of the static normal lower limit value and the static normal upper limit value is as follows:

[0140] 1. Wireless communication link terminal

[0141] RTT: The static normal lower limit value is 0 ms, and the static normal upper limit value is 500 ms. The link delay of a normally operating wireless communication link terminal is usually less than 500 ms.

[0142] Packet loss rate: The static normal lower limit value is 0%, and the static normal upper limit value is 5%. The packet loss rate of a normally operating wireless communication link terminal should not exceed 5%.

[0143] Bandwidth utilization rate: The static normal lower limit value is 10%, and the static normal upper limit value is 90%. Too low or too high bandwidth utilization rate is determined as abnormal.

[0144] Link state: Its value is an enumeration type, and the legal value is 0 or 1, 1 for normal and 0 for abnormal. The static normal lower limit value and the static normal upper limit value are both 1, and only 1 is allowed.

[0145] Signal quality: the static normal lower limit value is -80 dBm, and the static normal upper limit value is 0 dBm. In the range of -80 dBm to 0 dBm, the signal strength is in the stable working interval.

[0146] 2. Wired serial port terminal

[0147] Operating voltage: the static normal lower limit value is 198V, and the static normal upper limit value is 242V. The normal range is centered on 220V rated voltage ±10%.

[0148] Operating power: the static normal lower limit value is 5% rated power, and the static normal upper limit value is 100% rated power. The operating power in the range of no load to full load is considered normal, and the actual power of the specific device can be set.

[0149] Operating state: its value is enumerated type, and the legal value is 0 or 1, 1 is normal and 0 is abnormal. The static normal lower limit value and the static normal upper limit value are both 1, and only 1 is allowed.

[0150] Communication state: its value is enumerated type, and the legal value is 0 or 1, 1 is normal and 0 is abnormal. The static normal lower limit value and the static normal upper limit value are both 1, and only 1 is allowed.

[0151] 3. Digital input and output terminal

[0152] Switch state: its value is enumerated type, and the legal value is 0 or 1, 1 is on and 0 is off. The static normal lower limit value is 0, and the static normal upper limit value is 1.

[0153] Output state: its value is enumerated type, and the legal value is 0 or 1, 1 is on and 0 is off. The static normal lower limit value is 0, and the static normal upper limit value is 1.

[0154] 4. Analog input and output terminal

[0155] Continuous physical quantity data: preset according to actual business scenarios.

[0156] In this embodiment, it includes:

[0157] Temperature: the static normal lower limit value is -20℃, and the static normal upper limit value is 80℃.

[0158] Water level: the static normal lower limit value is 0m, and the static normal upper limit value is 5m.

[0159] Gas concentration: the static normal lower limit value is 0ppm, and the static normal upper limit value is the preset upper limit of safe concentration ppm.

[0160] 5. Power distribution terminal

[0161] Output voltage: the lower limit of static normal value is 90% rated voltage, and the upper limit of static normal value is 110% rated voltage.

[0162] Output current: the lower limit of static normal value is 0 A, and the upper limit of static normal value is the rated current value of the power distribution terminal.

[0163] Out-of-limit abnormality alarm: the value is an enumeration type, and the legal values are 0 or 1, 1 is normal, and 0 is abnormal. The lower limit of static normal value and the upper limit of static normal value are both 1, and only 1 is allowed.

[0164] Step S1022: For each sampling value in the multi-source collected data set, the dynamic fluctuation normalized abnormality scores of continuous sampling values and enumeration sampling values are respectively determined based on a dynamic fluctuation sliding window.

[0165] The continuous sampling values include RTT, packet loss rate, bandwidth utilization, signal quality, operating voltage, operating power, continuous physical quantity data, output voltage, and output current.

[0166] The enumeration sampling values include link state, operating state, communication state, switch state, output state, and out-of-limit abnormality alarm.

[0167] In this step, it specifically includes:

[0168] 1. Set the dynamic fluctuation sliding window length k, where k is a positive integer.

[0169] Based on the historical sampling values, the dynamic fluctuation sliding window is constructed in combination with the time stamp and the dynamic fluctuation sliding window length k. That is, the historical sampling values with the sampling value identifier i are sorted in ascending order of time stamp, the t-th sampling value in the sequence is selected as the current sample, and the last k sampling values are extracted forward The dynamic fluctuation sliding window is constructed. In this embodiment, the value of k is set to 5.

[0170] 2. Determine the dynamic fluctuation normalized abnormality score of the continuous sampling value, including:

[0171] For each continuous sampling value, set the corresponding sliding standard deviation upper limit value to limit the short-term fluctuation amplitude and prevent occasional disturbance from being misjudged as abnormal.

[0172] The sliding standard deviation upper limit value of the continuous sampling value is set as follows:

[0173] Wireless communication link terminal

[0174] RTT: the sliding standard deviation upper limit value is 50 seconds.

[0175] Packet loss rate: the sliding standard deviation upper limit value is 2%.

[0176] Bandwidth utilization: sliding standard deviation upper limit value is 20%.

[0177] Signal quality: sliding standard deviation upper limit value is 5dB.

[0178] Serial terminal

[0179] Operating voltage: sliding standard deviation upper limit value is 2V.

[0180] Operating power: sliding standard deviation upper limit value is 10% of rated power.

[0181] Analog input / output terminal

[0182] Continuous physical quantity data: preset according to actual business scenarios.

[0183] In this embodiment, it includes:

[0184] Temperature: sliding standard deviation upper limit value is 3℃.

[0185] Water level: sliding standard deviation upper limit value is 0.2m.

[0186] Gas concentration: sliding standard deviation upper limit value is 5% of the upper limit of safe concentration.

[0187] Power distribution terminal

[0188] Output voltage: sliding standard deviation upper limit value is 2% of rated voltage.

[0189] Output current: sliding standard deviation upper limit value is 10% of rated current.

[0190] Determine the dynamic fluctuation normalized anomaly score of the continuous sampling value. The rule for determining the dynamic fluctuation normalized anomaly score of the continuous sampling value is:

[0191] When the sliding standard deviation is less than or equal to the sliding standard deviation upper limit value, the dynamic fluctuation normalized anomaly score of the continuous sampling value is 0;

[0192] When the sliding standard deviation is greater than the sliding standard deviation upper limit value, the dynamic fluctuation normalized anomaly score of the continuous sampling value is calculated.

[0193] The calculation formula of the dynamic fluctuation normalized anomaly score of the continuous sampling value is:

[0194]

[0195] Where i is the unique identifier of the sampling value; is the dynamic fluctuation normalized anomaly score of the sampling value with unique identifier i; is the sliding standard deviation; is the sliding standard deviation upper limit value.

[0196] The calculation formula of the sliding standard deviation is:

[0197]

[0198] Wherein, is the sliding standard deviation of the sampling value uniquely identified as i; k is the length of the dynamic fluctuation sliding window; is the t-jth sampling value uniquely identified as i; is the average value of the sampling value uniquely identified as i in the dynamic fluctuation sliding window.

[0199] 3. Determine the dynamic fluctuation anomaly score of the enumerated sampling value.

[0200] The dynamic fluctuation anomaly score of the enumerated sampling value is calculated according to the state jump frequency, which is used to reflect its short-term fluctuation. The calculation formula of the dynamic fluctuation anomaly score of the enumerated sampling value is:

[0201]

[0202] Wherein, i is the unique identification of the sampling value; is the dynamic fluctuation normalized anomaly score of the sampling value uniquely identified as i; If the adjacent two sampling values are different, it is 1, otherwise it is 0. It is used to count the number of jump events, which can reflect the state instability.

[0203] Step S1023: For each sampling value, select the larger value between its static normalized anomaly score and dynamic fluctuation normalized anomaly score as the single sampling item anomaly score of the sampling value.

[0204] The expression for determining the single sampling item anomaly score is:

[0205]

[0206] Wherein, i is the unique identification of the sampling value; is the single sampling item anomaly score of the sampling value uniquely identified as i, is the static normalized anomaly score of the sampling value uniquely identified as i, is the dynamic fluctuation normalized anomaly score of the sampling value uniquely identified as i.

[0207] It should be noted that the static normalized abnormal score is used to identify whether the sampling value exceeds the preset normal range, reflecting the absolute abnormality of the sampling value itself. The dynamic fluctuation normalized abnormal score is used to identify the sharp fluctuation of the sampling value in a short period of time, reflecting the stability problem of the sampling value. Through the double mechanism of the static normalized abnormal score and the dynamic fluctuation normalized abnormal score, the comprehensive identification of the abnormal state of the multi-source heterogeneous terminal sampling is realized, the accuracy and sensitivity are taken into account, the missed detection and false alarm are avoided, and the robustness and reliability of the multi-source heterogeneous terminal abnormal detection are improved.

[0208] Step S1024: Based on each single sampling item abnormal score, the importance weight of each sampling value is combined for weighted comprehensive processing to generate a weighted comprehensive abnormal score of each multi-source heterogeneous terminal.

[0209] In this step, when the multi-source heterogeneous terminal type is a wireless communication link terminal, the calculation formula of the weighted comprehensive abnormal score is:

[0210]

[0211] wherein, is the weighted comprehensive abnormal score of the wireless communication link terminal; , , , and are importance weights; is the single sampling item abnormal score of the RTT; is the single sampling item abnormal score of the packet loss rate; is the single sampling item abnormal score of the bandwidth utilization rate; is the single sampling item abnormal score of the link state; is the single sampling item abnormal score of the signal quality.

[0212] The value of is 0.25; The value of is 0.25; The value of is 0.2; The value of is 0.15; The value of is 0.15.

[0213] When the multi-source heterogeneous terminal type is a wired serial port terminal, the weighted comprehensive abnormal score is:

[0214]

[0215] wherein, is the weighted comprehensive abnormal score of the wired serial port terminal; , , and are importance weights; is the single-sample-item abnormal score of the working voltage; is the single-sample-item abnormal score of the running power; is the single-sample-item abnormal score of the running state; is the single-sample-item abnormal score of the communication state.

[0216] is 0.3; is 0.25; is 0.25; is 0.2.

[0217] When the multi-source heterogeneous terminal type is a digital input and output terminal, the weighted comprehensive abnormal score is:

[0218]

[0219] wherein, is the weighted comprehensive abnormal score of the digital input and output terminal; , are importance weights; is the single-sample-item abnormal score of the switch state; is the single-sample-item abnormal score of the output state.

[0220] is 0.5; is 0.5.

[0221] When the multi-source heterogeneous terminal type is an analog input and output terminal, the weighted comprehensive abnormal score is:

[0222]

[0223] wherein, is the weighted comprehensive abnormal score of the analog input and output terminal; is the single-sample-item abnormal score of the continuous physical quantity data.

[0224] When the multi-source heterogeneous terminal type is a power distribution terminal, the weighted comprehensive abnormal score is:

[0225]

[0226] wherein, is the weighted comprehensive abnormal score of the power distribution terminal; , and are importance weights; is the single-sample-item abnormal score of the output voltage; is the single-sample-item abnormal score of the output current; is the single sampling item abnormal score of the out-of-limit anomaly alarm.

[0227] the value of is 0.4; the value of is 0.4; the value of is 0.2.

[0228] Step S1025: determining the abnormal level of the multi-source heterogeneous terminal according to the weighted comprehensive abnormal score based on an abnormal level classification rule.

[0229] In this step, the abnormal level classification rule comprises:

[0230] when the weighted comprehensive abnormal score is less than 0.05, the abnormal level is normal;

[0231] when the weighted comprehensive abnormal score is greater than or equal to 0.05 and less than 0.15, the abnormal level is mild abnormality;

[0232] when the weighted comprehensive abnormal score is greater than or equal to 0.15 and less than 0.35, the abnormal level is moderate abnormality;

[0233] when the weighted comprehensive abnormal score is greater than or equal to 0.35, the abnormal level is severe abnormality.

[0234] Figure 4 is a flowchart of an abnormal response and a local protection decision according to an example embodiment of the present application. As shown in Figure 4 the method provided by the present embodiment comprises:

[0235] Step S1031: corresponding matching the abnormal level of the multi-source heterogeneous terminal with a preset local protection strategy.

[0236] In this step, the local protection strategy is to trigger a local abnormal protection measure and an abnormal reporting mechanism when the abnormal level is mild abnormality, moderate abnormality and severe abnormality.

[0237] the local abnormal protection measure comprises:

[0238] when the abnormal level is mild abnormality, adjusting the frequency of the sampling period and starting a warning mechanism.

[0239] when the abnormal level is moderate abnormality, executing a local protection measure.

[0240] when the abnormal level is severe abnormality, executing a forced local protection measure.

[0241] In the present embodiment, according to different types of multi-source heterogeneous terminals, the frequency of the sampling period, the early warning mechanism, the local protection measure and the mandatory local protection measure are all different, and the details are shown as follows:

[0242] 1. Wireless communication link terminal:

[0243] When the abnormality level is mild abnormality, the frequency of the sampling period is adjusted to 6-40 seconds. The early warning mechanism is to start the local link quality detection, and the Internet of Things operation and maintenance box performs Ping test periodically and counts the packet loss rate to evaluate the current link quality state.

[0244] When the abnormality level is moderate abnormality, the local protection measure is to perform wireless module soft restart, and the Internet of Things operation and maintenance box issues a control instruction to drive the wireless communication link terminal to perform module restart operation, which includes re-registering the base station and rebuilding the link connection.

[0245] When the abnormality level is severe abnormality, the mandatory local protection measure is to perform primary and backup link switching, and the Internet of Things operation and maintenance box issues a control instruction to trigger the wireless communication link terminal to switch to the backup communication module or configure the backup access point name (APN) to ensure communication recovery.

[0246] 2. Wired serial port terminal:

[0247] When the abnormality level is mild abnormality, the frequency of the sampling period is adjusted to 20-80 seconds. The early warning mechanism is to strengthen the serial port communication state detection, and the Internet of Things operation and maintenance box counts the retransmission times, timeout count and CRC check failure for evaluation.

[0248] When the abnormality level is moderate abnormality, the local protection measure is serial port module soft restart, and the Internet of Things operation and maintenance box issues a control instruction to trigger the wired serial port terminal to reinitialize the serial port communication stack.

[0249] When the abnormality level is severe abnormality, the mandatory local protection measure is to disconnect the serial port communication connection, and the Internet of Things operation and maintenance box issues a control instruction to control the wired serial port terminal to realize communication disconnection by pulling down the hardware interface level or triggering the interrupt pin.

[0250] 3. Digital input and output terminal:

[0251] When the abnormality level is mild abnormality, the frequency of the sampling period is adjusted to 0.6-6 seconds. The early warning mechanism is to poll the state change quickly , and the Internet of Things operation and maintenance box realizes it by shortening the sampling period to read the port state periodically and recording the number of jumps.

[0252] When the abnormality level is moderate abnormality, the local protection measure is to restart the digital input and output terminal The acquisition module is triggered by a control instruction issued by the Internet of Things operation and maintenance case to execute the module restart operation of the digital input and output terminal.

[0253] When the abnormality level is severe abnormality, the mandatory local protection measure is to close the digital output driving circuit of the digital input and output terminal to prevent error action propagation. A control instruction is issued by the Internet of Things operation and maintenance case to trigger the digital input and output terminal to close its internal driving to block the false control signal output.

[0254] 4, analog input and output terminal:

[0255] When the abnormality level is mild abnormality, the frequency of the sampling period is adjusted to 6-40 seconds. The early warning mechanism is to filter the sampling value, and the Internet of Things operation and maintenance case uses the moving average method to smooth the sampling value, reduces the influence of mutation or jitter, and more accurately reflects the terminal running state. Specifically, based on the filter processing sliding window length h, the average value of the current sampling value and the previous h-1 sampling values is calculated to replace the current value for subsequent analysis. Wherein, h is a positive integer.

[0256] When the abnormality level is moderate abnormality, the local protection measure is to reset the analog interface module, and a control instruction is issued by the Internet of Things operation and maintenance case to trigger the analog input and output terminal to initialize its internal ADC module to restore the sampling function.

[0257] When the abnormality level is severe abnormality, the mandatory local protection measure is to disconnect the analog signal path to prevent abnormal output from causing failure. An instruction is issued by the Internet of Things operation and maintenance case to trigger the analog input and output terminal to execute the disconnection operation to prevent the abnormal signal from continuing to output.

[0258] 5, power distribution terminal:

[0259] When the abnormality level is mild abnormality, the frequency of the sampling period is adjusted to 20-80 seconds. The early warning mechanism is to monitor the output voltage and output current changes in real time, and the Internet of Things operation and maintenance case continuously collects the corresponding data based on the updated sampling period and performs fluctuation analysis.

[0260] When the abnormality level is moderate abnormality, the local protection measure is to run locally limited current and automatically reduce the output power. A control instruction is issued by the Internet of Things operation and maintenance case to trigger the power distribution terminal to reduce the output power.

[0261] When the abnormality level is severe abnormality, the mandatory local protection measure is to quickly disconnect the power protection, and a control instruction is issued by the Internet of Things operation and maintenance case to trigger the power distribution terminal to automatically reduce the output power to the set safety threshold.

[0262] The abnormality reporting mechanism comprises:

[0263] When the abnormality level is mild abnormality, the abnormality information is recorded in the local abnormality log, the abnormality information is only synchronized to the local edge server or the regional management node, and is not immediately reported to the remote center platform. The abnormality information is reported after being aggregated in batches or in a timing manner, thereby reducing the occupation of communication resources.

[0264] When the abnormality level is moderate abnormality, the abnormality information is recorded in the local abnormality log, and the abnormality alarm information is reported to the remote center platform within a set short-time abnormality reporting window, and is identified as a middle-level early warning level, so that the remote center platform can timely perceive potential risks. In the present example, the short-time abnormality reporting window is 5 minutes.

[0265] When the abnormality level is severe abnormality, the abnormality information is recorded in the local abnormality log, and the abnormality alarm information is immediately reported to the remote center platform, and is identified as a high-level alarm level.

[0266] It should be noted that the abnormality information comprises a terminal unique identifier, a multi-source heterogeneous terminal type and an abnormality level.

[0267] The abnormality alarm information further comprises a local protection measure or a mandatory local protection measure that has been executed on the basis of the abnormality information.

[0268] When the abnormality level is normal, the local abnormality protection measure and the abnormality reporting mechanism are not triggered. Only the multi-source heterogeneous terminal periodic data acquisition task is executed, and the collected raw data is standardized to generate a multi-source acquisition data set, which is output and saved to the local sampling log in a standardized JSON format.

[0269] Step S1032: performing execution result monitoring based on the execution results of the early warning mechanism, the local protection measure and the mandatory local protection measure.

[0270] Since the corresponding early warning mechanism, local protection measure and mandatory local protection measure are different according to different multi-source heterogeneous terminal categories, the method of performing execution result monitoring based on the execution results of the early warning mechanism, the local protection measure and the mandatory local protection measure is also different.

[0271] 1. Wireless communication link terminal:

[0272] After the execution of the early warning mechanism, it is monitored whether the Ping delay is reduced and whether the packet loss rate is restored to the normal range.

[0273] After the execution of the local protection measure, it is checked whether the base station registration state is successful and whether the link is re-established.

[0274] When the mandatory local protection measures are executed, check whether the standby link is switched successfully and whether the communication is resumed.

[0275] 2. Wired serial port terminal:

[0276] When the early warning mechanism is executed, monitor whether the retransmission frequency is reduced, whether the communication timeout rate is decreased, and whether the CRC check pass rate is restored.

[0277] When the local protection measures are executed, check whether the serial port communication is successfully reconnected and whether the communication data is normally transmitted.

[0278] When the mandatory local protection measures are executed, check whether the communication is re-established after the physical interface is reset and there is no abnormal interruption.

[0279] 3. Digital input and output terminal:

[0280] When the early warning mechanism is executed, check whether the IO state changes tend to be stable, without frequent toggling or abnormal jitter.

[0281] When the local protection measures are executed, check whether the digital quantity cache is successfully flushed and whether the digital input and output is restored to normal communication.

[0282] When the mandatory local protection measures are executed, check whether the terminal is physically isolated after the control line is disconnected to avoid false output.

[0283] 4. Analog input and output terminal:

[0284] When the early warning mechanism is executed, check whether the sampling value fluctuation is reduced and whether the data tends to be stable.

[0285] When the local protection measures are executed, check whether the sampling value of the analog input and output terminal is within the normal range.

[0286] When the mandatory local protection measures are executed, check whether the analog output is set to zero or the path is disconnected to prevent abnormal control signal output.

[0287] 5. Power distribution terminal:

[0288] When the early warning mechanism is executed, check whether the output voltage and current are within the normal range and whether the fluctuation is smooth.

[0289] When the local protection measures are executed, check whether the current limiting measures are effective and whether the output power is reduced to a safe range.

[0290] When the mandatory local protection measures are executed, check whether the output end is successfully powered off to prevent further electrical failure.

[0291] Figure 5is a structural schematic diagram of an edge heterogeneous Internet of Things operation and maintenance case data processing system according to an example embodiment of the present application. As shown in Figure 5 The edge heterogeneous Internet of Things operation and maintenance case data processing system 500 provided by the present embodiment includes:

[0292] The multi-source heterogeneous terminal data acquisition and preprocessing module 510 formulates a unified data acquisition scheme, defines a unified acquisition data format, adopts an asynchronous multi-thread scheduling mechanism, respectively manages and executes the data acquisition task of each multi-source heterogeneous terminal, acquires and generates corresponding acquisition raw data, and performs standardization processing on the acquisition raw data in a normal sampling state to generate a multi-source acquisition data set with unified structure.

[0293] The comprehensive anomaly evaluation module 520 calculates a standardized anomaly score for the corresponding multi-source acquisition data set according to each multi-source heterogeneous terminal type, performs weighted comprehensive processing, generates a comprehensive anomaly score, and determines an anomaly level according to the comprehensive anomaly score.

[0294] The anomaly response and local protection decision module 530 matches the anomaly level of the multi-source heterogeneous terminal with a preset local protection strategy, executes local anomaly protection actions including a warning mechanism, a local protection measure, and a mandatory local protection measure based on the matching result, records anomaly information, and reports anomaly alarm information to a remote center platform.

[0295] Other embodiments of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. The application is intended to cover any variations, uses or adaptations of the application following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice in the art to which the application pertains. The specification and examples are to be regarded as illustrative only, and the true scope and spirit of the application are indicated by the following claims.

[0296] It should be understood that the present application is not limited to the precise construction that has been described above and illustrated in the accompanying drawings, and that various modifications and changes can be made without departing from the scope thereof. The scope of the application is limited only by the claims that follow.

Claims

1. A method for processing edge heterogeneous IoT operation and maintenance chassis data, characterized in that: include: Data collection and preprocessing for multi-source heterogeneous terminals: Develop a unified data collection plan based on the type of multi-source heterogeneous terminals, define a unified data collection format, and use an asynchronous multi-threaded scheduling mechanism to manage and execute data collection tasks for each multi-source heterogeneous terminal. Collect and generate corresponding raw data. Standardize the raw data with normal sampling status to generate a multi-source collection data set with a unified structure. Comprehensive anomaly assessment: based on each of the multi-source heterogeneous terminal types, for the corresponding multi-source collected data set, calculate a standardized anomaly score, perform weighted comprehensive processing, generate a comprehensive anomaly score, and determine the anomaly level based on the comprehensive anomaly score; Abnormal response and local protection decision-making, matching the abnormal level of the multi-source heterogeneous terminals with the preset local protection strategy, executing local abnormal protection actions including early warning mechanism, local protection measures and mandatory local protection measures based on the matching results, recording abnormal information and reporting abnormal alarm information to the remote central platform; The unified data collection format includes a unique terminal identifier, a multi-source heterogeneous terminal type, a sampling value set, a timestamp, and a collection status; The sampling value set includes sampling values ​​and corresponding sampling value unique identifiers; The comprehensive abnormality assessment includes: Determining a static normalized anomaly score for each of the sample values ​​in the multi-source acquisition data set; For each of the sample values ​​in the multi-source acquisition data set, determining a dynamic fluctuation normalized anomaly score of a continuous sample value and an enumerated sample value based on a dynamic fluctuation sliding window; For each sampling value, according to its static normalized anomaly score and dynamic fluctuation normalized anomaly score, the larger value of the two is selected as the single sampling item anomaly score of the sampling value; Based on the anomaly score of each single sampling item and in combination with the importance weight of each sampling value, a weighted comprehensive processing is performed to generate a weighted comprehensive anomaly score of each multi-source heterogeneous terminal; Determining the abnormality level of the multi-source heterogeneous terminal based on the weighted comprehensive abnormality score and an abnormality level classification rule; The rule for determining the static normalized anomaly score is: When the sample value is within the corresponding static normal threshold range, the static normalized anomaly score of the sample value is 0; When the sample value exceeds the corresponding static normal threshold interval, the static normalized anomaly score of the sample value is calculated; The calculation formula of the static normalized anomaly score is: Among them, i is the unique identifier of the sampling value, is the static normalized anomaly score of the sample value uniquely identified as i, is the sampling value uniquely identified as i, The sampling value the nearest static normal threshold; The rule for determining the dynamic fluctuation normalized anomaly score of the continuous sampling value is: When the sliding standard deviation is less than or equal to the upper limit of the sliding standard deviation, the dynamic fluctuation normalized anomaly score of the continuous sampling value is 0; When the sliding standard deviation is greater than the upper limit of the sliding standard deviation, the dynamic fluctuation normalized anomaly score of the continuous sampling value is calculated; The calculation formula of the dynamic fluctuation normalized anomaly score of the continuous sampling value is: Among them, i is the unique identifier of the sampling value, is the dynamic fluctuation normalized anomaly score of the sample value uniquely identified as i, is the sliding standard deviation of the sample value uniquely identified as i, is the upper limit of the sliding standard deviation; The calculation formula for determining the dynamic fluctuation anomaly score of the enumerated sampling value is: Among them, i is the unique identifier of the sampling value, is the dynamic fluctuation normalized anomaly score of the sample value uniquely identified as i, k is the length of the dynamic fluctuation sliding window, If the two adjacent sampling values ​​are different, the value is 1, otherwise the value is 0. The sampling value is uniquely identified as i A sampling value.

2. The edge heterogeneous IoT operation and maintenance chassis data processing method according to claim 1 is characterized in that: The unified data collection solution includes: Setting a sampling period, communication protocol parameters, and communication timeout threshold for each of the multi-source heterogeneous terminal types; Based on the sampling period, communication protocol parameters and communication timeout threshold, a unified collection task list is configured.

3. The edge heterogeneous IoT operation and maintenance chassis data processing method according to claim 1 is characterized in that: The abnormality levels include normal, mild abnormality, moderate abnormality and severe abnormality; The local protection strategy is to trigger local abnormality protection measures and abnormality reporting mechanisms when the abnormality level is mild abnormality, moderate abnormality and severe abnormality; The local abnormality protection measures include: When the abnormality level is a mild abnormality, the frequency of the sampling period is adjusted and the early warning mechanism is activated; When the abnormality level is moderate, local protection measures are implemented; When the abnormality level is severe, mandatory local protection measures are implemented; The abnormality reporting mechanism includes: When the abnormality level is a mild abnormality, the abnormality information is recorded in the local abnormality log, and the abnormality information is only synchronized to the local edge server or regional management node, and is not immediately reported to the remote central platform; When the abnormality level is moderate, the abnormality information is recorded in the local abnormality log, and the abnormality alarm information is reported to the remote center platform within the set short-term abnormality reporting window, marked as a medium warning level; When the abnormality level is a severe abnormality, the abnormality information is recorded in the local abnormality log, and the abnormality alarm information is immediately reported to the remote center platform, marked as a high alarm level.

4. An edge heterogeneous IoT operation and maintenance chassis data processing system, characterized in that: include: The multi-source heterogeneous terminal data collection and preprocessing module formulates a unified data collection plan and defines a unified collection data format based on the type of multi-source heterogeneous terminal. It adopts an asynchronous multi-threaded scheduling mechanism to manage and execute the data collection tasks of each multi-source heterogeneous terminal, collect and generate the corresponding collection raw data, and perform standardization processing on the collection raw data with normal sampling status to generate a multi-source collection data set with a unified structure. a comprehensive anomaly assessment module, which calculates a standardized anomaly score for each of the multi-source heterogeneous terminal types and the corresponding multi-source collected data set, performs weighted comprehensive processing to generate a comprehensive anomaly score, and determines an anomaly level based on the comprehensive anomaly score; The abnormal response and local protection decision module matches the abnormal level of the multi-source heterogeneous terminals with the preset local protection strategy, executes local abnormal protection actions including early warning mechanism, local protection measures and mandatory local protection measures based on the matching results, records abnormal information and reports abnormal alarm information to the remote central platform; The unified data collection format includes a unique terminal identifier, a multi-source heterogeneous terminal type, a sampling value set, a timestamp, and a collection status; The sampling value set includes sampling values ​​and corresponding sampling value unique identifiers; The comprehensive abnormality assessment includes: Determining a static normalized anomaly score for each of the sample values ​​in the multi-source acquisition data set; For each of the sample values ​​in the multi-source acquisition data set, determining a dynamic fluctuation normalized anomaly score of a continuous sample value and an enumerated sample value based on a dynamic fluctuation sliding window; For each sampling value, according to its static normalized anomaly score and dynamic fluctuation normalized anomaly score, the larger value of the two is selected as the single sampling item anomaly score of the sampling value; Based on the anomaly score of each single sampling item and in combination with the importance weight of each sampling value, a weighted comprehensive processing is performed to generate a weighted comprehensive anomaly score of each multi-source heterogeneous terminal; Determining the abnormality level of the multi-source heterogeneous terminal based on the weighted comprehensive abnormality score and an abnormality level classification rule; The rule for determining the static normalized anomaly score is: When the sample value is within the corresponding static normal threshold range, the static normalized anomaly score of the sample value is 0; When the sample value exceeds the corresponding static normal threshold interval, the static normalized anomaly score of the sample value is calculated; The calculation formula of the static normalized anomaly score is: Among them, i is the unique identifier of the sampling value, is the static normalized anomaly score of the sample value uniquely identified as i, is the sampling value uniquely identified as i, The sampling value the nearest static normal threshold; The rule for determining the dynamic fluctuation normalized anomaly score of the continuous sampling value is: When the sliding standard deviation is less than or equal to the upper limit of the sliding standard deviation, the dynamic fluctuation normalized anomaly score of the continuous sampling value is 0; When the sliding standard deviation is greater than the upper limit of the sliding standard deviation, the dynamic fluctuation normalized anomaly score of the continuous sampling value is calculated; The calculation formula of the dynamic fluctuation normalized anomaly score of the continuous sampling value is: Among them, i is the unique identifier of the sampling value, is the dynamic fluctuation normalized anomaly score of the sample value uniquely identified as i, is the sliding standard deviation of the sample value uniquely identified as i, is the upper limit of the sliding standard deviation; The calculation formula for determining the dynamic fluctuation anomaly score of the enumerated sampling value is: Among them, i is the unique identifier of the sampling value, is the dynamic fluctuation normalized anomaly score of the sample value uniquely identified as i, k is the length of the dynamic fluctuation sliding window, If the two adjacent sampling values ​​are different, the value is 1, otherwise the value is 0. The sampling value is uniquely identified as i A sampling value.

Images