Management machine intelligent control system and method based on Internet of Things

Abstract

The invention discloses a management machine intelligent control system and method based on the Internet of Things, and relates to the electrical field. Complete working condition operation data in various environments are collected, key feature points of three-phase voltage are extracted, multi-dimensional analysis of abnormal features of the three-phase voltage is achieved, and an abnormal sample library is constructed; accurate synchronization of time is realized by using the Beidou clock server, timestamp marking is performed on data, a basis is provided for fault traceability when a fault occurs subsequently, an expert sets key feature point weights of three-phase voltage according to an actual scene to generate a comprehensive index, the reliability of a threshold value is ensured, a classification threshold value is set according to the comprehensive index, and the classification accuracy is improved. According to the method, the threshold boundary is dynamically adjusted by introducing the equipment operation duration parameter, when the comprehensive index is in different threshold intervals, different emergency plans are triggered, and the data with the timestamp is uploaded, so that the safety of a power system and the reliability of equipment operation are improved.

Description

Technical Field

[0001] This invention relates to the electrical field, specifically to an intelligent control system and method for management machines based on the Internet of Things (IoT). Background Technology

[0002] In power systems, abnormal three-phase voltage can damage power equipment, leading to power outages or even safety accidents. Therefore, real-time monitoring and accurate early warning of voltage status are crucial. Traditional monitoring methods use fixed thresholds, which cannot adapt to the insufficient tolerance of power equipment due to aging, easily causing misjudgments. Traditional monitoring methods also have limited dimensional extraction and insufficient time synchronization accuracy, resulting in disordered time sequence of multi-source data and a lack of fault tracing capabilities. With the intelligent development of power grids, there is an urgent need for an intelligent control method that can analyze the key characteristics of multiple voltages and dynamically adjust thresholds to address the insufficient tolerance of equipment due to aging over time, ultimately leading to safety accidents, thereby improving the safety of power system operation. Summary of the Invention

[0003] The purpose of this invention is to provide an intelligent control system and method for management machines based on the Internet of Things (IoT) to solve the problems raised in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an intelligent control method for a management machine based on the Internet of Things, the intelligent management method comprising the following steps: Step S1: Use a real-time power system monitoring device to sample the three-phase voltage of the device. The sampling process covers normal and abnormal operating conditions. Based on the sampled data, use time series analysis to plot the dynamic curve of the device's voltage change over time. Step S1-1: Use power monitoring equipment to monitor the voltage data of the equipment under rated load, light load and overload conditions. Use a sampling rate greater than 1000 points / second to synchronously sample the voltage data of the equipment, and the collection time covers a 24-hour cycle. Step S1-2: Under test conditions, simulate abnormal operating conditions and use monitoring equipment to record the transient voltage data of the equipment under abnormal operating conditions; Steps S1-3: Use the moving average filtering method to remove noise from the original sampled data. Use segmented truncation to extract the voltage sequence of the three stages before, during and after the fault, forming independent time series samples. Plot the voltage change curve over time based on the sample data to provide a basis for subsequent data analysis. The system ensured the completeness, accuracy, and analyzability of the collected voltage data, providing reliable data support for subsequent steps.

[0005] Step S2: Use the Beidou clock server to unify the time reference for all voltage data monitored by the monitoring equipment, and timestamp each data point in the dynamic curve using the international standard time format. Step S2-1: Connect the Beidou clock server to the monitoring equipment. The Beidou clock server receives the orbital parameters, atomic clock timestamps and time synchronization information transmitted by the satellite. It extracts the UTC time reference carried by the satellite based on its own coordinates, calculates the satellite time and compares it with the built-in clock, and calibrates the local time through digital phase-locked loop technology. Step S2-2: The monitoring device periodically sends a request message containing the local timestamp to the Beidou clock. After receiving the request, the Beidou clock server responds immediately and sends a message containing the UTC timestamp and its own received timestamp. Step S2-3: When the monitoring equipment performs sampling operations on transformers and transmission lines, it reads the synchronized local clock and generates a response timestamp. The timestamp format is the international standard time format, which provides a basis for tracing and locating the source of subsequent faults. By using timestamps, the consistency of the time base in the sampling of monitoring equipment is ensured, and the problem of misalignment of equipment data time sequence is prevented. By marking timestamps, the faults can be accurately compared when power equipment fails in the future, and the evolution process of equipment abnormality can be restored.

[0006] Step S3: Collect the three-phase voltage data of the equipment under abnormal scenarios, extract key features, and construct abnormal three-phase voltage feature samples; Step S3-1: Classify abnormal scenarios into three-phase voltage imbalance, single-phase grounding fault, two-phase short circuit and three-phase short circuit; Step S3-2: Mark the fault type and timestamp on the abnormal operating condition information collected in step S1; Step S3-3: Extract the time-domain and frequency-domain features of the abnormal state voltage. The time-domain features are divided into the voltage amplitude deviation rate of each phase and the three-phase voltage imbalance. The formula for calculating the voltage amplitude deviation rate of each phase is as follows: ; In the formula, U is the voltage amplitude deviation rate of the k-th phase. k U is the actual measured value of the voltage of phase k. nom This is the rated value of the voltage; The formula for calculating three-phase voltage unbalance is: ; ; In the formula, U represents the three-phase voltage imbalance. max U is the maximum value among the three-phase voltages.min U is the minimum value among the three-phase voltages. avg U is the arithmetic mean of the three-phase voltages. A U B U C , which are the individual single-phase voltages in the three-phase voltage; Frequency domain analysis features are divided into transient and harmonic features. Transient features are specifically manifested as voltage fluctuations, and the formula is: ; In the formula, ∆Uk(t) is the sudden change in the voltage of the k-th phase at time t, Uk(t) represents the instantaneous value of the voltage of the k-th phase at time t, and Uk(t-1) represents the instantaneous value of the voltage of the k-th phase at time t-1. Harmonics specifically refer to harmonic distortion rate, which is calculated using the following formula: ; In the formula, THD k Let be the total harmonic distortion rate of the k-th phase voltage. U is the amplitude of the h-th harmonic voltage of the k-th phase voltage. k,1 Let be the harmonic value of the k-th phase voltage; Step S3-4: Store the extracted features in a structured manner according to the fault type to build a sample library; By analyzing the key characteristics of voltage from multiple dimensions, we can achieve precise voltage analysis. Through the resume sample library, we can provide a basis for subsequent threshold setting.

[0007] Step S4: Analyze the normal operating voltage range of the equipment under normal operating conditions, identify and set graded early warning tripping thresholds by combining deep learning algorithms, and introduce the equipment age parameter to correct the threshold boundaries in real time through an adaptive filtering algorithm. Step S4-1: Based on historical normal operating condition information, set the three-phase voltage data with no fault alarm records within the load rate range of 20%-80% as the three-phase voltage data under normal operating conditions. Step S4-2: Calculate the mean, standard deviation, and range of each voltage; Step S4-3: Normalize the voltage amplitude deviation rate of each phase, the three-phase voltage imbalance, the voltage mutation, the harmonic distortion rate, the voltage mean, standard deviation, and range. The calculation formula is as follows: ; In the formula, F is the original eigenvalue, F norm The value is the normalized value; Step S4-4: The weight ratio of each feature is manually set by professionals according to the actual scenario. Weights are assigned to the seven features, namely, the phase voltage amplitude deviation rate, three-phase voltage imbalance, voltage surge, harmonic distortion rate, voltage mean, standard deviation, and range. 1, 2, 3, ..., 7. The sum of the weights of the seven features under the constraint condition equals 1; Step S4-5: Calculate the comprehensive index based on the weights and normalized values, specifically the sum of the products of each normalized value and the assigned weight; Steps S4-6: Based on historical work scenarios and expert evaluation comprehensive index, set three threshold levels S1, S2, and S3. [S1, S2] is the first-level warning, [S2, S3] is the second-level warning, and exceeding S3 is a trip warning. Step S4-7: Establish a time-threshold decay function to describe the decreasing trend of equipment tolerance over time. The calculation formula is as follows: ; In the formula, y represents the number of years the equipment has been in use, Y max The design life is given by α, which is the aging factor. Step S4-8, the threshold correction factor is: ; Step S4-9: Use recursive least squares filtering to fuse the equipment age parameter with real-time monitoring data, dynamically adjust the threshold boundary, and calculate the formula as follows: ; In the formula, The threshold value is the threshold value of the previous time step, z is the real-time feature value, specifically the normalized value in step S4-3, and K is the filter gain. This formula is used to update the three threshold values ​​in step S4-6. By introducing the equipment's operating life parameter and dynamically setting the threshold, the threshold is adapted to the equipment's operating time, thus improving the sensitivity and reliability of the early warning system.

[0008] Step S5: Activate early warning mechanisms at all levels for abnormal states near each threshold boundary. Once the tripping threshold is exceeded, trigger the relay protection device to directly issue a tripping command and simultaneously upload the fault timestamp and waveform data to the main station system.

[0009] Step S5-1: Set different levels of early warning plans according to the grading thresholds set in step S4. Step S5-2: When the comprehensive characteristic value index is in [S1, S2], the first-level emergency plan is activated; when it is in [S2, S3], the second-level emergency plan is activated. When the comprehensive characteristic value index exceeds the threshold S3, a trip warning is issued, triggering the relay protection device to directly issue a trip command. Step S5-3: When the equipment trips, the monitoring equipment extracts three-phase voltage data 500ms before and after the fault based on the UTC timestamp, and encapsulates the timestamp and model data into a structured data packet. Step S5-4: The monitoring device uploads fault data packets to the main station system through an encrypted VPN channel using the TCP protocol. If the network is interrupted, the local cache is used, and the data is uploaded after the network is restored. After receiving the data, the main station automatically parses and stores it.

[0010] The system includes a data acquisition and time synchronization module, a data processing module, an early warning module, and a data upload and storage module; The data acquisition and time synchronization module is responsible for acquiring, synchronizing, and timestampting the three-phase voltage data of the power equipment. The data processing module is responsible for voltage feature point extraction and sample construction; The early warning module is responsible for setting thresholds, activating early warning plans, and uploading fault data to the main station.

[0011] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention provides richer and more accurate data support for subsequent processes by collecting data from various scenarios and extracting multi-dimensional voltage features.

[0012] 2. This invention sets feature weight ratios, generates comprehensive indicators based on features, sets graded thresholds according to the indicators, and introduces a lifespan parameter to achieve dynamic adjustment of the thresholds, thereby accurately matching the tolerance of the equipment's life cycle.

[0013] 3. This invention provides data support for post-fault analysis by intercepting, storing, and uploading data before and after the trip. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating an intelligent control method for a management machine based on the Internet of Things according to the present invention. Figure 2 This is a schematic diagram of the structure of an intelligent control system for management machines based on the Internet of Things according to the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Step S1: Use a real-time power system monitoring device to sample the three-phase voltage of the device. The sampling process covers normal and abnormal operating conditions. Based on the sampled data, use time series analysis to plot the dynamic curve of the device's voltage change over time. Step S1-1: Use power monitoring equipment to monitor the voltage data of the equipment under rated load, light load and overload conditions. Use a sampling rate greater than 1000 points / second to synchronously sample the voltage data of the equipment, and the collection time covers a 24-hour cycle. Step S1-2: Under test conditions, simulate abnormal operating conditions and use monitoring equipment to record the transient voltage data of the equipment under abnormal operating conditions; Steps S1-3: Use the moving average filtering method to remove noise from the original sampled data. Use segmented truncation to extract the voltage sequence of the three stages before, during and after the fault, forming independent time series samples. Plot the voltage change curve over time based on the sample data to provide a basis for subsequent data analysis. The system ensured the completeness, accuracy, and analyzability of the collected voltage data, providing reliable data support for subsequent steps.

[0017] Step S2: Use the Beidou clock server to unify the time reference for all voltage data monitored by the monitoring equipment, and timestamp each data point in the dynamic curve using the international standard time format. Step S2-1: Connect the Beidou clock server to the monitoring equipment. The Beidou clock server receives the orbital parameters, atomic clock timestamps and time synchronization information transmitted by the satellite. It extracts the UTC time reference carried by the satellite based on its own coordinates, calculates the satellite time and compares it with the built-in clock, and calibrates the local time through digital phase-locked loop technology. Step S2-2: The monitoring device periodically sends a request message containing the local timestamp to the Beidou clock. After receiving the request, the Beidou clock server responds immediately and sends a message containing the UTC timestamp and its own received timestamp. Step S2-3: When the monitoring equipment performs sampling operations on transformers and transmission lines, it reads the synchronized local clock and generates a response timestamp. The timestamp format is the international standard time format, which provides a basis for tracing and locating the source of subsequent faults. By using timestamps, the consistency of the time base in the sampling of monitoring equipment is ensured, and the problem of misalignment of equipment data time sequence is prevented. By marking timestamps, the faults can be accurately compared when power equipment fails in the future, and the evolution process of equipment abnormality can be restored.

[0018] Step S3: Collect the three-phase voltage data of the equipment under abnormal scenarios, extract key features, and construct abnormal three-phase voltage feature samples; Step S3-1: Classify abnormal scenarios into three-phase voltage imbalance, single-phase grounding fault, two-phase short circuit and three-phase short circuit; Step S3-2: Mark the fault type and timestamp on the abnormal operating condition information collected in step S1; Step S3-3: Extract the time-domain and frequency-domain features of the abnormal state voltage. The time-domain features are divided into the voltage amplitude deviation rate of each phase and the three-phase voltage imbalance. The formula for calculating the voltage amplitude deviation rate of each phase is as follows: ; In the formula, U is the voltage amplitude deviation rate of the k-th phase. k U is the actual measured value of the voltage of phase k. nom This is the rated value of the voltage; The formula for calculating three-phase voltage unbalance is: ; ; In the formula, U represents the three-phase voltage imbalance. max U is the maximum value among the three-phase voltages. min U is the minimum value among the three-phase voltages. avg U is the arithmetic mean of the three-phase voltages. A U B U C , which are the individual single-phase voltages in the three-phase voltage; Frequency domain analysis features are divided into transient and harmonic features. Transient features are specifically manifested as voltage fluctuations, and the formula is: ; In the formula, ∆Uk(t) is the sudden change in the voltage of the k-th phase at time t, Uk(t) represents the instantaneous value of the voltage of the k-th phase at time t, and Uk(t-1) represents the instantaneous value of the voltage of the k-th phase at time t-1. Harmonics specifically refer to harmonic distortion rate, which is calculated using the following formula: ; In the formula, THD k Let be the total harmonic distortion rate of the k-th phase voltage. U is the amplitude of the h-th harmonic voltage of the k-th phase voltage. k,1 Let be the harmonic value of the k-th phase voltage; Step S3-4: Store the extracted features in a structured manner according to the fault type to build a sample library; By analyzing the key characteristics of voltage from multiple dimensions, we can achieve precise voltage analysis. Through the resume sample library, we can provide a basis for subsequent threshold setting.

[0019] Step S4: Analyze the normal operating voltage range of the equipment under normal operating conditions, identify and set graded early warning tripping thresholds by combining deep learning algorithms, and introduce the equipment age parameter to correct the threshold boundaries in real time through an adaptive filtering algorithm. Step S4-1: Based on historical normal operating condition information, set the three-phase voltage data with no fault alarm records within the load rate range of 20%-80% as the three-phase voltage data under normal operating conditions. Step S4-2: Calculate the mean, standard deviation, and range of each voltage; Step S4-3: Normalize the voltage amplitude deviation rate of each phase, the three-phase voltage imbalance, the voltage mutation, the harmonic distortion rate, the voltage mean, standard deviation, and range. The calculation formula is as follows: ; In the formula, F is the original eigenvalue, F norm The value is the normalized value; Step S4-4: The weight ratio of each feature is manually set by professionals according to the actual scenario. Weights are assigned to the seven features, namely, the phase voltage amplitude deviation rate, three-phase voltage imbalance, voltage surge, harmonic distortion rate, voltage mean, standard deviation, and range. 1, 2, 3, ..., 7. The sum of the weights of the seven features under the constraint condition equals 1; Step S4-5: Calculate the comprehensive index based on the weights and normalized values, specifically the sum of the products of each normalized value and the assigned weight; Steps S4-6: Based on historical work scenarios and expert evaluation comprehensive index, set three threshold levels S1, S2, and S3. [S1, S2] is the first-level warning, [S2, S3] is the second-level warning, and exceeding S3 is a trip warning. Step S4-7: Establish a time-threshold decay function to describe the decreasing trend of equipment tolerance over time. The calculation formula is as follows: ; In the formula, y represents the number of years the equipment has been in use, Y max The design life is given by α, which is the aging factor. Step S4-8, the threshold correction factor is: ; Step S4-9: Use recursive least squares filtering to fuse the equipment age parameter with real-time monitoring data, dynamically adjust the threshold boundary, and calculate the formula as follows: ; In the formula, The threshold value is the threshold value of the previous time step, z is the real-time feature value, specifically the normalized value in step S4-3, and K is the filter gain. This formula is used to update the three threshold values ​​in step S4-6. By introducing the equipment's operating life parameter and dynamically setting the threshold, the threshold is adapted to the equipment's operating time, thus improving the sensitivity and reliability of the early warning system.

[0020] Step S5: Activate early warning mechanisms at all levels for abnormal states near each threshold boundary. Once the tripping threshold is exceeded, trigger the relay protection device to directly issue a tripping command and simultaneously upload the fault timestamp and waveform data to the main station system.

[0021] Step S5-1: Set different levels of early warning plans according to the grading thresholds set in step S4. Step S5-2: When the comprehensive characteristic value index is in [S1, S2], the first-level emergency plan is activated; when it is in [S2, S3], the second-level emergency plan is activated. When the comprehensive characteristic value index exceeds the threshold S3, a trip warning is issued, triggering the relay protection device to directly issue a trip command. Step S5-3: When the equipment trips, the monitoring equipment extracts three-phase voltage data 500ms before and after the fault based on the UTC timestamp, and encapsulates the timestamp and model data into a structured data packet. Step S5-4: The monitoring device uploads fault data packets to the main station system through an encrypted VPN channel using the TCP protocol. If the network is interrupted, the local cache is used, and the data is uploaded after the network is restored. After receiving the data, the main station automatically parses and stores it.

[0022] The system includes a data acquisition and time synchronization module, a data processing module, an early warning module, and a data upload and storage module; The data acquisition and time synchronization module is responsible for acquiring, synchronizing, and timestampting the three-phase voltage data of the power equipment. The data processing module is responsible for voltage feature point extraction and sample construction; The early warning module is responsible for setting thresholds, activating early warning plans, and uploading fault data to the main station.

[0023] Example 1: The sampling rate was set to 1024 points / second, and continuous sampling was conducted for 24 hours, resulting in a total of 88,473,600 sampling points. The measured value of the three-phase voltage U... A 220.5V, U B It is 219.8V, U C The voltage is 221.2V, simulating a three-phase short-circuit fault. The voltage during the fault is U. A It is 105.3V, U B It is 102.7V, U CThe voltage is 104.5V. It receives satellite UTC time from a BeiDou clock server. The local clock is calibrated via a digital phase-locked loop with an error of less than or equal to 100ns. The monitoring equipment sends a time request every 100ms, and the response message includes a UTC timestamp in an international standard format. Analysis of the three-phase voltage short-circuit fault voltage deviation rate yields... =-52.14%, =-52.14%, =-52.14%, three-phase voltage imbalance ℇ≈2.49%, voltage surge is 105.2V, and the harmonic distortion rate assumes a fundamental amplitude U k,1 =100V, U k,5 =8V, THD k =8%, which is structured and stored. The average three-phase voltage under normal operating conditions is 220.3V, the standard deviation is 0.8V, and the range is 2.5V. Its key characteristics are normalized to obtain =0.113, ∆U norm =0.8, THD norm =0.8, norm =0, norm =0.15, R norm =0.1, ℇ norm =0.249 was assigned a weight, respectively 1=0.15 , 2=0.2 , 3=0.15 , 4=0.2 , 5=0.1 , 6=0.1 , 7=0.1 The comprehensive index equals 0.403. The threshold values ​​are set sequentially as S1=0.3, S2=0.6, and S3=0.8. The equipment has been used for five years, with an aging coefficient of 0.03. The result is obtained from the function. =0.977. According to the formula, the new threshold after correction is 0.782. Then, the judgment and warning are made according to the range of the warning value, and the data is uploaded after the fault occurs.

[0024] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. An Internet of Things-based management machine intelligent control method, characterized in that: The method comprises the following steps: Step S1, using a power system real-time monitoring device to sample the three-phase voltage of the device, the sampling process covering normal and abnormal operating conditions, based on the sampling data, using time series analysis method to draw dynamic curve diagram of the voltage of the device changing with time; Step S2, using Beidou clock server to unify the time reference for all voltage data monitored by the monitoring device, and using international standard time format to time stamp each data in the dynamic curve diagram; Step S3, collecting the voltage data of the three-phase voltage of the device under abnormal scenarios, extracting key features, and constructing abnormal three-phase voltage feature samples; Step S4, analyzing the normal operating voltage interval of the device under normal conditions, combining deep learning algorithm to identify the set graded early warning trip threshold, and introducing the device age parameter to correct the threshold boundary in real time through adaptive filtering algorithm; Step S5, starting the early warning mechanism of each level for the abnormal state near each threshold boundary, triggering the relay protection device when exceeding the trip threshold, directly issuing the trip command, and synchronously uploading the fault time stamp and waveform data to the master station system. 2.The intelligent control method of a management machine based on an Internet of Things according to claim 1, characterized in that: The specific steps of step S1 are as follows: Step S1-1, using the power monitoring device to monitor the voltage data of the device under rated load, light load and overload state, synchronously sampling the voltage data of the device at a sampling rate greater than 1000 points / second, and collecting the voltage data for 24 hours; Step S1-2, simulating abnormal conditions in the test environment, and recording the transient voltage data of the device under abnormal conditions by using the monitoring device; Step S1-3, using the moving average filtering method to remove noise from the original sampling data, using segmented interception to extract the voltage sequence of the three stages before, during and after the fault under abnormal conditions, forming independent time series samples, and drawing the voltage-time curve according to the sample data. 3.The intelligent control method of a management machine based on an Internet of Things according to claim 2, characterized in that: The specific steps of step S2 are as follows: Step S2-1, connecting the Beidou clock server to the monitoring device, the Beidou clock server receiving the satellite transmitted orbit parameters, atomic clock time stamp and time synchronization information, extracting the UTC time reference carried by the satellite according to its own coordinates, comparing the satellite time with the built-in clock through calculation, and calibrating the local time through digital phase-locked loop technology; Step S2-2, the monitoring device sends a request message containing the local time stamp to the Beidou clock in a periodic manner, and the Beidou clock server responds immediately after receiving the request, and sends a message containing the UTC time stamp and the received time stamp; Step S2-3, when the monitoring device performs sampling operation on the transformer and transmission line, the synchronized local clock is read to generate a response time stamp, and the time stamp format is international standard time format, which provides a basis for subsequent fault tracing and positioning.

4. The intelligent control method for management machine based on Internet of Things according to claim 3, characterized in that: The specific steps of collecting the voltage data of the three-phase voltage of the device under abnormal scenarios in step S3 are as follows: Step S3-1, classifying abnormal scenarios into three-phase voltage imbalance, single-phase ground fault, two-phase short circuit and three-phase short circuit; Step S3-2, marking the fault type and time stamp for the abnormal operating condition information collected in step S1. 5.The intelligent control method of a management machine based on an Internet of Things according to claim 4, characterized in that: The specific steps of the step S3 of key feature extraction and sample library establishment are as follows: Step S3-3, time domain and frequency domain features of the abnormal state voltage are extracted, the time domain features include phase voltage amplitude deviation rate and three-phase voltage unbalance degree, the phase voltage amplitude deviation rate calculation formula is: ； wherein Ukis the kth phase voltage amplitude deviation rate, U k Ukis the kth phase voltage amplitude deviation rate, U nom Ukis the kth phase voltage amplitude deviation rate, U The three-phase voltage unbalance degree calculation formula is: ； ； wherein is the three-phase voltage unbalance degree, U max is the maximum value in the three-phase voltage, U min is the minimum value in the three-phase voltage, U avg is the arithmetic mean in the three-phase voltage, U A , U B , U C are the individual single-phase voltages in the three-phase voltage, respectively The frequency domain analysis features include transient and harmonic features, the transient is specifically represented by voltage mutation, the formula is: ； In the formula, ΔUk(t) is the mutation of the k-phase voltage at t, Uk(t) represents the instantaneous value of the k-phase voltage at t, and Uk(t-1) represents the instantaneous value of the k-phase voltage at t-1; The harmonic is specifically represented by harmonic distortion rate, the calculation formula is: ； where THD k is the total harmonic distortion of the kth phase voltage, is the hth harmonic voltage magnitude of the kth phase voltage, U k,1 is the harmonic value of the kth phase voltage; Step S3-4, the extracted features are stored according to the fault types, and a sample library is constructed. 6.The intelligent control method of a management machine based on an Internet of Things according to claim 5, characterized in that: The specific steps of the step S4 of analyzing the normal working voltage interval are as follows: Step S4-1, according to the historical normal working condition information, the three-phase voltage data without fault alarm records in the load rate range of 20%-80% is set as the three-phase voltage data of the normal working condition; Step S4-2, the mean value, standard deviation and range of each voltage are calculated. 7.The intelligent control method of a management machine based on an Internet of Things according to claim 6, characterized in that: The specific steps of the step S4 of setting the hierarchical early warning tripping threshold are as follows: Step S4-3, the phase voltage amplitude deviation rate, three-phase voltage unbalance degree, voltage mutation, harmonic distortion rate, voltage mean value, standard deviation and range are normalized, and the calculation formula is: ； In the formula, F is the original feature value, F norm is the normalized value; Step S4-4, the weight ratio of each feature is manually set by a professional according to the actual scene, and each feature is assigned a weight, respectively, the three-phase voltage amplitude deviation rate, the three-phase voltage unbalance degree, the voltage mutation, the harmonic distortion rate, the voltage average, the standard deviation and the range are respectively 1, 2, 3, 7, the sum of the weights of the seven features is equal to 1; Step S4-5, the comprehensive index is calculated according to the weight and the normalized value, specifically, the sum of each normalized value multiplied by the assigned weight; Step S4-6, according to the historical working scene and the expert evaluation comprehensive index, three threshold values S1, S2 and S3 are set, [S1, S2] is the first level early warning, [S2, S3] is the second level early warning, and when the value exceeds S3, it is the tripping early warning. 8.The intelligent control method of a management machine based on an Internet of Things according to claim 7, characterized in that: The specific steps of the step S4 of self-adaptive correction of the threshold boundary are as follows: Step S4-7, a service life-threshold decay function is established to describe the downward trend of the equipment tolerance capacity with time, and the calculation formula is: ； In the formula, y is the service life of the equipment, Y max is the design life, and a is the aging coefficient. Step S4-8, the threshold correction factor is: ； Step S4-9, the equipment service life parameter and the real-time monitoring data are fused by using the recursive least square filter, the threshold boundary is dynamically adjusted, and the calculation formula is: ； In the formula, is the threshold value of the last time, z is the real-time characteristic value, specifically the normalized value in step S4-3, and K is the filtering gain. The three threshold values in step S4-6 are updated using the formula. 9.The intelligent control method of a management machine based on an Internet of Things according to claim 8, characterized in that: The specific steps of the step S5 are as follows: Step S5-1, different levels of early warning plans are set according to the hierarchical threshold set in the step S4; Step S5-2, when the characteristic value comprehensive index is in [S1, S2], the first level plan is started, when the characteristic value comprehensive index is in [S2, S3], the second level plan is started, when the characteristic value comprehensive index exceeds the threshold S3, it is the tripping early warning, the relay protection device is triggered, and the tripping command is directly performed; Step S5-3, when the equipment trips, the monitoring equipment intercepts the three-phase voltage data 500 ms before and after the fault based on the UTC time stamp, and encapsulates the time stamp and the model data into a structured data packet; Step S5-4, the monitoring equipment uploads the fault data packet to the master station system through an encrypted VPN channel, adopts the TCP protocol, and uses the local cache when the network is interrupted, and uploads after the network is restored, after the master station receives the data, automatically analyzes and stores the data.

10. An Internet of Things based management machine intelligent control system, characterized in that: The system comprises the following modules: The system comprises a data acquisition and time synchronization module, a data processing module, an early warning module and a data uploading and storage module; The data acquisition and time synchronization module is responsible for the acquisition of three-phase voltage data of power equipment, time synchronization and timestamp marking of data; The data processing module is responsible for voltage feature point extraction and sample construction; The early warning module is responsible for threshold setting, starting of early warning plan and uploading of fault data to the master station.

Images