Weighted integral-based DC system insulation state monitoring method

Abstract

The invention relates to the technical field of insulation state monitoring, in particular to a DC system insulation state monitoring method based on weighted integration, which comprises the following steps: S1, respectively collecting positive pole to ground original current and negative pole to ground original current in real time through a positive pole MTR sensor module and a negative pole MTR sensor module; s2, calibrating the original current to obtain calibrated positive electrode to ground current and negative electrode to ground current; s3, sliding window weighted integral processing is carried out on the calibrated current, and positive pole weighted integral and negative pole weighted integral are calculated according to a weighting function which is a function attenuating along with time; s4, judging grounding types based on the weighted integral ratio and the weighted average current, wherein the grounding types comprise positive electrode grounding, negative electrode grounding and bipolar grounding; s5, whether insulation is gradually reduced or not is judged through the weighted integral change rate and linear regression significance test; and S6, triggering graded alarm according to the grounding type and severity. According to the invention, the problems of weak anti-interference capability, lag in trend identification, fuzzy polarity judgment and low calculation precision are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of insulation state monitoring, and particularly relates to a DC system insulation state monitoring method based on weighted integration. BACKGROUND

[0002] The working power supply of control systems, signal systems and automatic control devices of relay protection of power plants, transformer substations and the like is a DC system, which is a very large multi-branch power supply network, and the insulation performance of the power distribution line directly relates to the safe operation of the power plants, transformer substations and the like, and it is of great significance to monitor the insulation on-line and effectively detect single-point grounding faults in time.

[0003] At present, the existing DC system insulation state monitoring has the following problems: 1. Weak anti-interference capability: the monitoring method based on instantaneous current difference is easily affected by electromagnetic interference and instantaneous noise, and has a high false alarm rate (>5%); 2. Trend recognition lag: the sensitivity is low to slowly developing insulation decline (such as the insulation resistance gradually decreases from 500kΩ to 200kΩ), and the alarm is given only when the fault is serious; 3. Ambiguous polarity judgment: the boundary condition definition of single-pole grounding and double-pole grounding is not clear, and the single-pole grounding is easily misjudged when the grounding current is low; 4. Insufficient calculation accuracy: the sensor error and current fluctuation are not considered, and the insulation resistance calculation error is >10%, which affects the fault positioning accuracy. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a DC system insulation state monitoring method based on weighted integration, which solves the problems of weak anti-interference capability, trend recognition lag, ambiguous polarity judgment and low calculation accuracy.

[0005] The present application is realized by the following technical solutions: A DC system insulation state monitoring method based on weighted integration is provided, comprising the following steps: S1, real-time acquisition of the positive pole MTR sensor module and the negative pole MTR sensor module respectively acquires the positive pole to ground original current and the negative pole to ground original current ; S2, calibration of the above original current to eliminate zero drift and gain error, to obtain the calibrated positive pole to ground current and the negative pole to ground current ; S3, sliding window weighted integration processing of the calibrated current, the window length is =30s, and the weight function is a function decaying with time ; wherein: , ,satisfy ; Calculate the positive electrode weighted integral and negative weighted integral ; S4. Determine the grounding type based on the weighted integral ratio and weighted average current, including positive grounding, negative grounding and bipolar grounding; S5. Determine whether the insulation gradually decreases by using the weighted integral rate of change and the significance test of linear regression; S6. Trigger graded alarms based on grounding type and severity.

[0006] Furthermore, in step S2, the calibration step is performed using the following formula: ; (2); in, This represents the gain calibration coefficient for the positive electrode MTR sensor module. This is the gain calibration coefficient for the negative electrode MTR sensor module; This refers to the zero-point calibration coefficient of the positive electrode MTR sensor module. Zero-point calibration coefficient of the negative electrode MTR sensor module.

[0007] Furthermore, in step S3, the weighted integral is calculated using the following formula: ; ; in, for The positive calibration current seconds ago, when When =0, the calibrated positive-to-ground current Highest weight; for The negative calibration current seconds ago, when When =0, the calibrated negative electrode-to-ground current It has the highest weight.

[0008] Furthermore, in step S4, the grounding type determination includes: Positive grounding determination: and ; Negative grounding determination: and ; Bipolar grounding determination: and and ; in: = is the positive electrode weighted average current, is the negative electrode weighted average current, is taken 4~6, is the sensor maximum error.

[0009] Further, in step S5, the positive electrode calibration current weighted integral change rate and the negative electrode calibration current weighted integral change rate are respectively calculated. If and the continuous 2 sliding windows satisfy, and the weighted linear regression slope , is the slope standard deviation, then it is determined that the positive electrode insulation gradually decreases; If and the continuous 2 sliding windows satisfy, and the weighted linear regression slope , is the slope standard deviation, then it is determined that the negative electrode insulation gradually decreases; In step S6, the triggering mechanism of the graded alarm is: Positive electrode ground threshold alarm: is triggered; Negative electrode ground threshold alarm: is triggered; Positive electrode ground trend alarm: and lasts for 60s is triggered; Negative electrode ground trend alarm: , and lasts for 60s is triggered; Bipolar alarm: when the bipolar ground condition is met and lasts for 3s, it is triggered.

[0010] Compared with the prior art, the present application has the following advantages: I. Strong anti-interference ability, through the processing of the weighted integral formula, the weight of recent data can be increased, the latest trend change can be more sensitively reacted, the interference of old data is reduced, the trend caused by recent data is highlighted, and the influence of instantaneous noise is smoothed.

[0011] II. Trend identification is sensitive, and the insulation gradually decreasing trend is identified 15~20s earlier than the traditional method.

[0012] III. Accurate polarity judgment, combined with the weighted integral ratio and error compensation, the single / bipolar ground judgment accuracy reaches 99.5% (the prior art is prone to misjudgment due to low current).

[0013] IV. High calculation precision, through the sensor calibration formula and weighted average processing, the insulation resistance calculation error is controlled within 3% (the prior art is >10%). ​

[0014] Five, compatibility, applicable to 110V, 220V, etc. Different voltage level of DC system, no need to adjust the hardware structure. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 The flowchart of the DC system insulation state monitoring method based on weighted integration of the application. DETAILED DESCRIPTION

[0016] In order to clearly illustrate the technical features of the present scheme, the present scheme will be described below through specific embodiments.

[0017] As shown in Figure 1 , a DC system insulation state monitoring method based on weighted integration, comprising the following steps: S1, through the positive MTR sensor module and negative MTR sensor module respectively real-time acquisition DC system positive ground original current And negative ground original current .

[0018] S2, the original current is calibrated, the zero drift and gain error are eliminated, and the calibrated positive ground current And negative ground current ; The calibration step is realized by the following formula: ; (2) Among them, The gain calibration coefficient of the positive MTR sensor module, The gain calibration coefficient of the negative MTR sensor module; The zero point calibration coefficient of the positive MTR sensor module, The zero point calibration coefficient of the negative MTR sensor module.

[0019] The calibration parameters are obtained by the following way: MTR sensor module gain calibration: input 1mA standard current, data processing module calculates , ;MTR sensor module zero point calibration: respectively, the positive and negative MTR sensor input end short circuit, collect 200 groups of original data, take the average as .

[0020] S3, the calibrated current is processed by sliding window weighted integration, the window length is =30s, 1s interval, 1 sampling point, the data processing module performs sliding window weighted integration on the calibrated current, and the weight function is a function decaying with time ; wherein: , , satisfying ; calculating the positive electrode weighted integral and the negative electrode weighted integral .

[0021] Weighted integration: sliding window integration of current signal based on time-decaying weight function, recent data weight high, long-term data weight low, used to balance real-time and anti-interference.

[0022] Sliding window: continuous acquisition of time series data segment (30s in the invention, containing 30 1s interval data points).

[0023] Weighted integration calculation is realized by the following formula: ; ; wherein, is the weighted integral of the positive electrode calibration current in a window period, reflecting the weighted positive electrode calibration current cumulative amount, strengthening the recent positive electrode current trend, and the unit is ; is the weighted integral of the negative electrode calibration current in a window period, reflecting the weighted negative electrode calibration current cumulative amount, strengthening the recent negative electrode current trend, and the unit is ; is the positive electrode calibration current before seconds, when =0, the calibrated positive electrode ground current has the highest weight; is the negative electrode calibration current before seconds, when =0, the calibrated negative electrode ground current has the highest weight.

[0024] S4, judging the grounding type based on the weighted integral ratio and the weighted average current, including positive electrode grounding, negative electrode grounding and bipolar grounding; The grounding type judgment includes: (1) Positive electrode grounding determination: and ; wherein, = The positive electrode weighted average current (the use of weighted integral average can avoid the collected data being pulled by long-term data, and more truly reflect the current level of the positive electrode after calibration), in order to simplify the calculation, the normalization processing can be carried out here, =1; to ensure that the positive electrode calibration current is much larger than the negative electrode calibration current; Take 4~6, The maximum error of the sensor covers 99% of the error range.

[0025] The weighted integral of the positive electrode calibration current S_(+,θ) (t) is obviously larger than the weighted integral of the negative electrode calibration current; and in order to exclude noise interference, the positive electrode weighted average current I ̅_(+,θ) is required to exceed 3 times the error range of the positive electrode MTR sensor module.

[0026] (2) Negative ground determination: And ; Among them, = The negative electrode weighted average current (the use of weighted integral average can avoid the collected data being pulled by long-term data, and more truly reflect the current level of the negative electrode after calibration), in order to simplify the calculation, the normalization processing can be carried out here, =1; to ensure that the negative electrode calibration current is much larger than the negative electrode calibration current, Take 4~6; The maximum error of the sensor covers 99% of the error range.

[0027] The weighted integral of the positive electrode calibration current S_(+,θ) (t) is obviously larger than the weighted integral of the negative electrode calibration current; and in order to exclude noise interference, the positive electrode weighted average current I ̅_(+,θ) is required to exceed 3 times the error range of the negative electrode MTR sensor module.

[0028] (3) Bipolar ground determination: And And ; Among them, = The positive electrode weighted average current, = The negative electrode weighted average current, Take 4~6, The maximum error of the sensor.

[0029] S5, judge whether the insulation is gradually decreased by the weighted integral change rate and linear regression significance test; The positive electrode calibration current weighted integral change rate And the negative electrode calibration current weighted integral change rate ; like And the conditions are met for two consecutive sliding windows, while the weighted linear regression slope is... , If the slope standard deviation is given, it is determined that the positive electrode insulation gradually decreases. like And the conditions are met for two consecutive sliding windows, while the weighted linear regression slope is... , If the slope standard deviation is given, it is determined that the insulation of the negative electrode gradually decreases. S6. Trigger graded alarms based on grounding type and severity.

[0030] The triggering mechanism for tiered alarms is as follows: (1) Positive grounding threshold alarm: When the positive electrode weighted average current I̅_(+,θ) > 1.1mA within a window period, an alarm is immediately triggered. Assuming the DC system is 220V, what is the safety threshold for the DC return current insulation resistance to ground? To reduce misjudgments, simplify calculations, and allow for margin, we take... =1mA, This represents the maximum error of the sensor.

[0031] (2) Negative grounding threshold alarm: Time; that is, within a window period when the negative electrode weighted average current When this happens, an alarm will be triggered immediately.

[0032] in, Assuming the DC system is 220V, what is the safety threshold for the DC return current insulation resistance to ground? To reduce misjudgments, simplify calculations, and allow for margin, we take... =1mA, This represents the maximum error of the sensor.

[0033] (3) Positive grounding trend alarm: It is triggered after 60 seconds.

[0034] When the positive grounding determination condition is met: and Then, the data processing module determines that a positive ground fault has occurred in the DC circuit, and then the data processing module compares... ,and And significance judgment Then, that is, within a window period, when the weighted integral rate of change of the positive electrode calibration current. Furthermore, the weighted average current at the positive terminal continues to increase within the last 5 seconds. Meeting the above conditions indicates that the insulation of the positive terminal of the DC circuit is gradually decreasing and this condition persists for two window periods. Then, an alarm was triggered.

[0035] wherein, I is the positive calibration current, and I is the negative calibration current. , assuming the DC system is 220V, according to the typical positive ground fault current variation characteristics , the weighted integral is more sensitive to recent changes, to improve sensitivity, the positive calibration current weighted integral change rate threshold is reduced by 20% to reduce false positives.

[0036] (4) Negative ground trend alarm: , and lasts for 60s.

[0037] When the negative ground determination condition is met: and After that, the data processing module determines that the DC loop has a negative ground, and then the data processing module compares , and , and the significance is judged , that is, when the negative calibration current weighted integral change rate and the negative weighted average current continues to increase in the last 5 seconds, the above conditions are met, indicating that the DC loop negative insulation is gradually decreasing, and the alarm is triggered after lasting for two window periods .

[0038] wherein, I is the positive calibration current, and I is the negative calibration current. , assuming the DC system is 220V, according to the typical negative ground fault current variation characteristics , the weighted integral is more sensitive to recent changes, to improve sensitivity, the negative calibration current weighted integral change rate threshold is reduced by 20% to reduce false positives.

[0039] (5) Bipolar alarm: triggered when the bipolar ground condition is met and lasts for 3s.

[0040] In this embodiment, the positive MTR sensor and the negative MTR sensor are both closed-loop Hall MTR sensors, with a range of ±10mA, an accuracy of 0.1% FS, and an output of 0~5V analog signal.

[0041] The alarm uses a three-level sound and light alarm circuit and an RS485 remote communication interface. The first-level sound and light alarm circuit includes a yellow LED + 200Hz buzzer, which is used for single-pole trend alarm; the second-level sound and light alarm circuit includes an orange LED + 500Hz buzzer, which is used for single-pole threshold alarm; the third-level alarm circuit includes a red LED + 1000Hz buzzer, which is used for bipolar alarm.

[0042] Of course, the above description is not limited to the above examples, and the technical features not described in the application can be implemented by or using the prior art, which will not be described here; the above examples and drawings are only used to illustrate the technical solutions of the application and are not a limitation on the application, and the application has been described in detail with reference to the preferred embodiments, and those skilled in the art should understand that changes, modifications, additions or substitutions made by those skilled in the art within the essential scope of the application do not deviate from the purpose of the application, and should also belong to the protection scope of the claims of the application.

Claims

1. A method for monitoring the state of insulation of a DC system based on weighted integration, characterized in that: Comprising the following steps: S1, collecting the direct current system positive pole to ground original current and the negative pole to ground original current through the positive pole MTR sensor module and the negative pole MTR sensor module respectively in real time and the negative pole to ground original current ; S2, calibrating the original current to eliminate zero drift and gain error, and obtaining the calibrated positive electrode-to-ground current and negative electrode-to-ground current ; S3. Perform sliding window weighted integration on the calibrated current, with a window duration of [value missing]. =30s, the weighting function is a function that decays over time. ;in: , ,satisfy ; computing positive and negative weighted integrals and negative weighted integrals ; S4, judging the grounding type based on the weighted integral ratio and the weighted average current, including positive grounding, negative grounding and bipolar grounding; S5, judging whether the insulation is gradually declining through the weighted integral change rate and the linear regression significance test; S6, triggering the hierarchical alarm according to the grounding type and the severity.

2. The method of claim 1, wherein: In step S2, the calibration step is realized by the following formula: ； （2）； wherein, GAINPOS is a gain calibration coefficient for the positive MTR sensor module, GAINNEG is a gain calibration coefficient for the negative MTR sensor module; OFFSETPOS is an offset calibration coefficient for the positive MTR sensor module, OFFSETNEG is an offset calibration coefficient for the negative MTR sensor module.

3. The method of claim 1, wherein: In step S3, the weighted integral calculation is realized by the following formula: ； ； wherein, is the positive electrode calibration current before the 1stsecond, when =0, the calibrated positive electrode current to ground is the highest weight; is the negative electrode calibration current before the 1stsecond, when =0, the calibrated negative electrode current to ground is the highest weight.

4. The method of claim 1, wherein: In step S4, the grounding type judgment includes: Positive electrode ground determination: And ; Negative ground determination: And ; Bipolar ground determination: and and ; wherein: = positive electrode average current, = positive electrode weighted average current, = negative electrode average current, = negative electrode weighted average current, = 4 ~ 6, = sensor maximum error.

5. The method of claim 1, wherein: In step S5, the positive electrode calibration current weighted integral change rate and the negative electrode calibration current weighted integral change rate are calculated, respectively. If and the continuous 2 sliding windows satisfy, and the weighted linear regression slope , is the slope standard deviation, then it is determined that the positive electrode insulation gradually decreases; If and the continuous 2 sliding windows satisfy, and the weighted linear regression slope , is the slope standard deviation, then it is determined that the negative electrode insulation gradually decreases.

6. The method of claim 1, wherein: In step S6, the triggering mechanism of the hierarchical alarm is: Positive electrode ground threshold alarm: Time triggered; Negative ground threshold alarm: Time triggered; Positive ground trend alarm: and triggers at 60s duration; Negative ground trend alarm: , and triggers when it lasts 60s; Bipolar alarm: triggered when the bipolar grounding condition is met and lasts for 3s.

Images