Transformer fault comprehensive diagnosis method based on multiple types of partial discharge sensors

Abstract

The invention relates to a transformer fault comprehensive diagnosis method based on multiple types of partial discharge sensors. The method comprises the following steps: acquiring partial discharge signals of a transformer through the multiple types of sensors; and according to the characteristic quantity of the partial discharge signal, carrying out transformer discharge fault analysis so as to realize comprehensive fault diagnosis of the transformer. According to the invention, by comprehensively analyzing the ultrasonic partial discharge sensor data, the high-frequency partial discharge sensor data, the ultrahigh-frequency partial discharge sensor data and the noise sensor data of the transformer, qualitative, quantitative and positioning diagnosis analysis of the discharge fault of the transformer is realized.

Description

Technical Field

[0001] The invention belongs to the technical field of online monitoring and fault diagnosis of large oil-filled power equipment such as transformers / reactors, and in particular relates to a comprehensive transformer fault diagnosis method based on multiple types of partial discharge sensors. Background Art

[0002] As the core hub equipment of the power grid, the transformer directly affects the safe operation of the power system. According to the different characteristic quantities of "sound, light, and electricity" during the transformer discharge process, transformer discharge monitoring mainly includes: ultrasonic partial discharge monitoring, high-frequency partial discharge monitoring, and ultra-high-frequency partial discharge monitoring. At present, various monitoring methods form their own systems, lacking the ability to comprehensively analyze the changes in multiple state quantities under the transformer fault state, resulting in low transformer discharge fault diagnosis accuracy, low positioning accuracy, and insufficient reliability of diagnosis results.

[0003] Therefore, there is an urgent need for a comprehensive transformer fault diagnosis method based on multi-type partial discharge sensors, so as to realize transformer discharge fault diagnosis by comprehensive analysis of transformer monitoring data, so as to improve the transformer fault warning capability and health status assessment capability. Summary of the invention

[0004] The present invention proposes a comprehensive transformer fault diagnosis method based on multiple types of partial discharge sensors. According to the different characteristic quantities of "sound, light, and electricity" in the transformer discharge process, transformer discharge monitoring mainly includes: ultrasonic partial discharge monitoring, high-frequency partial discharge monitoring, and ultra-high frequency partial discharge monitoring. By comprehensively analyzing the transformer ultrasonic partial discharge sensor data, high-frequency partial discharge sensor data, ultra-high frequency partial discharge sensor data, and noise sensor data, the qualitative, quantitative, and positional diagnosis of transformer discharge faults can be achieved, and the transformer fault warning capability and health status assessment capability can be improved.

[0005] The technical solution adopted by the present invention to achieve the above-mentioned purpose is: a comprehensive transformer fault diagnosis method based on multiple types of partial discharge sensors, comprising the following steps:

[0006] Collect partial discharge signals of transformers through various types of sensors;

[0007] According to the characteristic quantity of partial discharge signal, transformer discharge fault analysis is carried out to realize comprehensive transformer fault diagnosis.

[0008] The multiple types of sensors include ultrasonic partial discharge sensors, high frequency partial discharge sensors, ultra-high frequency partial discharge sensors, and noise sensors.

[0009] The transformer discharge fault analysis includes a qualitative analysis of the transformer discharge fault, and the specific steps are as follows:

[0010] Respectively obtain the amplitude and frequency of the partial discharge signal monitored by the high-frequency partial discharge sensor, the amplitude and frequency of the partial discharge signal monitored by the ultra-high-frequency partial discharge sensor, and the amplitude of the partial discharge monitored by the noise sensor;

[0011] The amplitude of the partial discharge signal monitored by the high-frequency partial discharge sensor is compared with several times the amplitude of the partial discharge monitored by the noise sensor, and different comparison results are used to indicate warning or non-warning.

[0012] The transformer discharge fault analysis includes a transformer discharge fault quantitative analysis method, and the specific steps are as follows:

[0013] 1) The difference between the maximum amplitude of the high-frequency partial discharge sensor and the amplitude of the noise sensor is used as the discharge fault intensity to quantitatively analyze the transformer discharge fault intensity;

[0014] 2) Perform the following steps for the characteristic quantity of the high-frequency partial discharge sensor:

[0015] 2.1) Calculate the change rate of each feature within M days:

[0016] ΔP x % = (P x0 -P x ) / P x0

[0017] Where ΔPx% represents the rate of change of each characteristic quantity within M days, x∈{1, 2, 3}, P1 represents the maximum discharge amplitude, P2 represents the average discharge amplitude; P3 represents the average number of pulses; P X0 Represents the historical data record of each characteristic quantity M days ago. If the transformer has not been in operation for M days, the historical data of the day when the transformer was officially put into operation can be used;

[0018] 2.2) Substitute the rate of change into the confidence function:

[0019]

[0020] In the formula, A(ΔP x %) represents the membership function value of the warning degree of each characteristic quantity of the maximum discharge amplitude, the average discharge amplitude, and the average number of pulses, x∈{1, 2, 3}; the overall threat rate of the discharge fault monitored by the transformer to the transformer is max{A(ΔP x %)|x∈{1,2,3}};

[0021] 2.3) Compare the overall threat rate of transformer discharge fault with the threshold; if it is greater than the threshold, the transformer should issue a warning of partial discharge exceeding the attention value; otherwise, combine the data value of the monitoring of dissolved gas in the transformer oil to manually determine whether to issue a warning of partial discharge exceeding the attention value.

[0022] The transformer discharge fault analysis includes a transformer discharge fault location analysis method, and the specific steps are as follows:

[0023] Step 1: Regional positioning: The high-frequency partial discharge sensor is connected to the transformer ground through the end screen of the bushing, and its partial discharge signal is Pb; the main high-frequency partial discharge sensor is set on the core or clamp grounding wire, and its partial discharge signal is Pt;

[0024] If the signal Pb measured on the casing is greater than the signal Pt measured on the body, it is considered that the partial discharge signal is likely to occur in the casing and the elevated area;

[0025] If the signal Pb measured on the casing is less than the signal Pt measured on the body, it is considered that the partial discharge signal is likely to occur in the body area;

[0026] If the signal measured on the main body Pt-△Pt<the signal measured on the casing Pb<the signal measured on the main body Pt+△Pt, △Pt represents the deviation limit, and the discharge amounts have reached the warning values, it is considered that there is a high probability that partial discharge signals exist in both the main body and the casing;

[0027] Step 2: Block positioning: When the partial discharge is determined to occur in the transformer body area through the first step, the determination range of partial discharge in the transformer body is narrowed by the block positioning method, including the following steps:

[0028] The electromagnetic wave signal received by the UHF PD sensor is selected as the starting point of the discharge time, and the time Ti when the 8 ultrasonic PD sensors receive the PD signal is identified respectively, i=1, 2, 3, 4, 5, 6, 7, 8, and the block where the sensor MIN{Ti} is located is the block with the maximum probability of PD occurrence;

[0029] Step 3: Precise positioning: According to the transformer structure, four ultrasonic partial discharge sensors with relatively short acoustic signal propagation paths are selected for precise positioning analysis. That is, for each block, sensors in the current block and the blocks directly adjacent to it are selected.

[0030] The coordinates of the four ultrasonic partial discharge sensors are set as S1 (x1, y1, z1), S2 (x2, y2, z2), S3 (x3, y3, z3), and S4 (x4, y4, z4). The speed of the sound wave is expressed as v t It indicates that the time when different ultrasonic partial discharge sensors receive the partial discharge signal is Ti, i=1, 2, 3, 5; the coordinates of the discharge position inside the transformer are (x, y, z), and the discharge position (x, y, z) inside the transformer is obtained by solving the relationship between different ultrasonic partial discharge sensors and the discharge position.

[0031] The least square method is used to solve the relationship between different ultrasonic partial discharge sensors and discharge positions.

[0032] Transformer fault comprehensive diagnosis system based on multi-type partial discharge sensors, including:

[0033] A sensor unit, including various types of sensors;

[0034] A data acquisition unit, used for acquiring a partial discharge signal of the transformer through a sensor unit;

[0035] A data processing unit, used for extracting characteristic quantities of partial discharge signals of each sensor in the sensor unit;

[0036] The data analysis unit is used to perform transformer discharge fault analysis based on the characteristic quantity of the partial discharge signal to achieve comprehensive transformer fault diagnosis.

[0037] The present invention has the following beneficial effects and advantages:

[0038] (1) The present invention realizes qualitative, quantitative and positional diagnosis and analysis of transformer discharge faults by comprehensively analyzing transformer ultrasonic partial discharge sensor data, high-frequency partial discharge sensor data, ultra-high frequency partial discharge sensor data and noise sensor data.

[0039] (2) The transformer discharge fault qualitative analysis module proposed in the present invention establishes an alarm condition coding table based on high-frequency partial discharge signals and ultra-high-frequency partial discharge signals to improve the reliability of transformer discharge fault warning.

[0040] (3) The transformer discharge fault power analysis module proposed in the present invention can not only analyze the quantitative data of discharge intensity, but also analyze the probability of threat of discharge fault to the main body, thereby improving the transformer discharge fault diagnosis and evaluation capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 Flow chart of the method of the present invention;

[0042] In the figure: 1-sensor unit, 2-data acquisition unit, 3-data processing unit, 4-data analysis unit, 5-ultrasonic partial discharge sensor, 6-high frequency partial discharge sensor, 7-ultra-high frequency partial discharge sensor, 8-noise sensor;

[0043] Figure 2 Schematic diagram of block positioning of the present invention. DETAILED DESCRIPTION

[0044] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.

[0045] A comprehensive transformer fault diagnosis method based on multiple types of partial discharge sensors can realize the qualitative, quantitative and location analysis of transformer partial discharge faults. Figure 1 shown.

[0046] (1) System composition

[0047] 1) The system composition of the diagnostic method proposed in the present invention includes: a sensor unit, a data acquisition unit, a data processing unit, and a data analysis unit.

[0048] 2) The sensor unit of the diagnostic method proposed in this patent includes:

[0049] Ultrasonic partial discharge sensor: bandwidth covers 80kHz~200kHz;

[0050] High frequency partial discharge sensor: bandwidth covers 3MHz~30MHz;

[0051] Ultra-high frequency partial discharge sensor: bandwidth covers 200MHz~1500MHz;

[0052] Noise sensor: bandwidth covers 100MHz~1000MHz.

[0053] The sensor unit is connected to the data acquisition unit. Each sensor is of existing technology.

[0054] 3) The data acquisition unit of the diagnostic method proposed in the present invention can collect at least 12 channels of ultrasonic partial discharge sensor data, at least 4 channels of high-frequency partial discharge sensor data, at least 3 channels of ultra-high frequency partial discharge sensor data, and at least 1 channel of noise sensor data. It supports 1MS / s synchronous acquisition of high-speed acquisition channels of ultrasonic partial discharge sensors, high-frequency partial discharge sensors, ultra-high frequency partial discharge sensors, and noise sensors. The data acquisition unit is connected to the data processing unit. The data of the data processing unit is connected to the data analysis unit. The data acquisition unit and the data processing unit are both prior art.

[0055] 4) The data processing unit of the diagnostic method proposed in the present invention realizes the extraction of characteristic quantities such as amplitude, phase, frequency, etc. of ultrasonic partial discharge sensors, high-frequency partial discharge sensors, ultra-high-frequency partial discharge sensors, and noise sensors.

[0056] 5) The data analysis unit of the diagnostic method proposed in the present invention includes: a qualitative analysis module, a quantitative analysis module, and a positioning analysis module.

[0057] (2) Qualitative analysis method of transformer discharge fault

[0058] The transformer discharge fault qualitative analysis method proposed in this patent adopts an alarm condition coding table based on high-frequency partial discharge signals and ultra-high-frequency partial discharge signals, in which the amplitude and frequency of high-frequency partial discharge signals and ultra-high-frequency partial discharge signals are selected as diagnostic feature quantities. The alarm condition coding table is shown in Table 1.

[0059] Table 1 Alarm condition coding table

[0060]

[0061] (3) Quantitative analysis method of transformer discharge fault

[0062] The quantitative analysis method for transformer discharge fault proposed in this patent adopts a calculation method based on high-frequency partial discharge signals and noise sensor signals to realize transformer discharge fault intensity analysis and discharge fault threat probability.

[0063] 1) The calculation formula for discharge fault intensity is as follows:

[0064] P=P 高 -P 噪 ——Formula 1

[0065] Where: P: discharge fault intensity, P 高： High frequency partial discharge sensor maximum amplitude, P 噪： Noise sensor amplitude

[0066] The difference between the maximum amplitude of the high-frequency partial discharge sensor and the amplitude of the noise sensor is used to quantitatively analyze the transformer discharge fault intensity.

[0067] 2) The calculation formula for the change rate of each characteristic within 30 days is as follows:

[0068] ΔP x % = (P x0 -P x ) / P x0 ——Formula 2

[0069] Where: ΔPx%: the rate of change of each characteristic quantity within 30 days (x∈{1, 2, 3}), P1: maximum discharge amplitude, P2: average discharge amplitude; P3: average number of pulses; P X0： The historical data of each characteristic quantity is recorded 30 days ago. If the transformer has not been in operation for 30 days, the historical data of the day when the transformer was officially put into operation can be used.

[0070] The confidence function expression is as follows:

[0071]

[0072] Where: A(ΔP x %): Membership function value of the characteristic parameter attribute “warning”, (x∈{1,2,3});

[0073] The overall threat probability of transformer discharge fault is max{A(ΔP x %)|x∈{1,2,3}}.

[0074] The overall threat rate of transformer discharge fault is compared with the threshold; if it is greater than the threshold, the transformer should issue a warning of partial discharge exceeding the attention value; otherwise, it is manually judged whether to issue a warning of partial discharge exceeding the attention value in combination with the monitoring data value of dissolved gas in transformer oil.

[0075] (4) Transformer discharge fault location analysis method

[0076] The transformer discharge fault location analysis method proposed in this patent locates the partial discharge position based on the acoustic and electrical three-step positioning method of ultrasonic partial discharge sensor signal, ultra-high frequency partial discharge signal, and high frequency sensor signal.

[0077] Step 1: Regional Positioning

[0078] The bushing is located above the transformer, and its high-frequency partial discharge sensor is connected to the transformer ground through the bushing end screen. Its partial discharge signal is set as Pb, and the high-frequency partial discharge sensor of the main body is on the core or clamp grounding wire. Its partial discharge signal is set as Pt. If the partial discharge occurs only in one part, for example, the partial discharge occurs at the bushing, the partial discharge signal is received by the partial discharge sensor connected to the main body through the transformer body, and its signal will be attenuated. Conversely, the partial discharge occurring in the main body is coupled to the bushing, which will also produce a certain attenuation. Based on these characteristics, the following method can be proposed for regional positioning.

[0079] If the signal Pb measured on the casing is greater than the signal Pt measured on the body, it is considered that the partial discharge signal is likely to occur in the casing and the elevated area.

[0080] If the signal Pb measured by the casing is less than the signal Pt measured by the body, it is considered that the partial discharge signal is likely to occur in the body area.

[0081] If the signal measured on the main body Pt-△Pt<the signal measured on the casing Pb<the signal measured on the main body Pt+△Pt, △Pt represents the deviation limit, and the discharge amounts have reached the warning values, it is considered that there is a high probability that partial discharge signals exist in both the main body and the casing;

[0082] Step 2: Block positioning

[0083] If the first step determines that the partial discharge occurs in the transformer body area, the block positioning method is used to narrow the determination range of the partial discharge in the transformer body and improve the positioning accuracy. Since the propagation speed of electromagnetic wave signals is much faster than that of acoustic signals and has strong anti-interference ability, the electromagnetic wave signal received by the ultra-high frequency partial discharge sensor is selected as the starting time of the discharge, and the time Ti when the 8 ultrasonic partial discharge sensors receive the partial discharge signal is identified respectively, i = 1, 2, 3, 4, 5, 6, 7, 8, and the block where the sensor MIN{Ti} is located is the block with the maximum probability of partial discharge.

[0084] The ultrasonic partial discharge sensor is fixed on the outer surface of the transformer oil tank by magnetic attraction, and four sensors are arranged on the two surfaces of the high-voltage side and the low-voltage side of the transformer. The schematic diagram of the sensor layout and the simplified block numbering of the transformer body rectangle are shown in the figure below.

[0085] Step 3: Accurate positioning

[0086] After determining the blocks, four ultrasonic partial discharge sensors with relatively short acoustic signal propagation paths were selected for precise positioning analysis based on the transformer structure. Figure 2 shown.

[0087] Sensor numbers selected for block 1: 1, 2, 3, 5

[0088] Sensor numbers selected for block 2: 1, 2, 4, 6

[0089] The sensor numbers selected in block 3 are: 1, 2, 3, 7

[0090] Sensor numbers selected in block 4: 2, 3, 4, 8

[0091] The sensor numbers selected in block 5 are: 1, 5, 6, 7

[0092] Sensor numbers selected for block 6: 5, 6, 7, 8

[0093] Sensor numbers selected for block 7: 3, 5, 7, 8

[0094] The sensor numbers selected in block 8 are: 4, 6, 7, 8

[0095] Taking the determination of partial discharge occurring in block 1 as an example, the coordinates of ultrasonic partial discharge sensor No. 1 are set to S1 (x1, y1, z1), the coordinates of ultrasonic partial discharge sensor No. 2 are set to S2 (x2, y2, z2), the coordinates of ultrasonic partial discharge sensor No. 3 are set to S3 (x3, y3, z3), and the coordinates of ultrasonic partial discharge sensor No. 5 are set to S4 (x4, y4, z4). The speed of the sound wave is expressed as v t Indicates that the time when different ultrasonic partial discharge sensors receive partial discharge signals is Ti, i = 1, 2, 3, 5. If the coordinates of the discharge position inside the transformer are (x, y, z), then the relationship between different ultrasonic partial discharge sensors and the discharge position is as described in Formula 4:

[0096]

[0097] By solving equation 4, the coordinates of the partial discharge position can be accurately calculated.

[0098] When partial discharge occurs in other blocks, the coordinates of the ultrasonic partial discharge sensor with the corresponding number can be selected and substituted into Formula 4 for solution and calculation to obtain the precise coordinates of the partial discharge position.

Claims

1. A comprehensive transformer fault diagnosis method based on multiple types of partial discharge sensors, characterized in that: The following steps are involved: Collect partial discharge signals of transformers through various types of sensors; According to the characteristic quantity of partial discharge signal, transformer discharge fault analysis is carried out to realize comprehensive transformer fault diagnosis.

2. The transformer fault comprehensive diagnosis method based on multiple types of partial discharge sensors according to claim 1 is characterized in that: The multiple types of sensors include ultrasonic partial discharge sensors, high frequency partial discharge sensors, ultra-high frequency partial discharge sensors, and noise sensors.

3. The transformer fault comprehensive diagnosis method based on multiple types of partial discharge sensors according to claim 1 is characterized in that: The transformer discharge fault analysis includes a qualitative analysis of the transformer discharge fault, and the specific steps are as follows: Respectively obtain the amplitude and frequency of the partial discharge signal monitored by the high-frequency partial discharge sensor, the amplitude and frequency of the partial discharge signal monitored by the ultra-high-frequency partial discharge sensor, and the amplitude of the partial discharge monitored by the noise sensor; The amplitude of the partial discharge signal monitored by the high-frequency partial discharge sensor is compared with several times the amplitude of the partial discharge monitored by the noise sensor, and different comparison results are used to indicate warning or non-warning.

4. The transformer fault comprehensive diagnosis method based on multiple types of partial discharge sensors according to claim 1 is characterized in that: The transformer discharge fault analysis includes a transformer discharge fault quantitative analysis method, and the specific steps are as follows: 1) The difference between the maximum amplitude of the high-frequency partial discharge sensor and the amplitude of the noise sensor is used as the discharge fault intensity to quantitatively analyze the transformer discharge fault intensity; 2) Perform the following steps for the characteristic quantity of the high-frequency partial discharge sensor: 2.1) Calculate the change rate of each feature within M days: ΔP x %=(P x0 -P x ) / P x0 Where ΔPx% represents the rate of change of each characteristic quantity within M days, x∈{1, 2, 3}, P1 represents the maximum discharge amplitude, P2 represents the average discharge amplitude; P3 represents the average number of pulses; P X0 Represents the historical data record of each characteristic quantity M days ago. If the transformer has not been in operation for M days, the historical data of the day when the transformer was officially put into operation can be used; 2.2) Substitute the rate of change into the confidence function: In the formula, A(ΔP x %) represents the membership function value of the warning degree of each characteristic quantity of maximum discharge amplitude, average discharge amplitude, and average number of pulses, x∈{1, 2, 3}; The overall threat rate of the transformer to the transformer is max{A(ΔP x %)|x∈{1,2,3}}; 2.3) Compare the overall threat rate of transformer discharge fault with the threshold; if it is greater than the threshold, the transformer should issue a warning that the partial discharge exceeds the attention value; Otherwise, it is manually determined whether to issue a warning that the partial discharge exceeds the warning value based on the monitoring data of the dissolved gas in the transformer oil.

5. The transformer fault comprehensive diagnosis method based on multiple types of partial discharge sensors according to claim 1 is characterized in that: The transformer discharge fault analysis includes a transformer discharge fault location analysis method, and the specific steps are as follows: Step 1: Regional positioning: The high-frequency partial discharge sensor is connected to the transformer ground through the end screen of the bushing, and its partial discharge signal is Pb; the main high-frequency partial discharge sensor is set on the core or clamp grounding wire, and its partial discharge signal is Pt; If the signal Pb measured on the casing is greater than the signal Pt measured on the body, it is considered that the partial discharge signal is likely to occur in the casing and the elevated area; If the signal Pb measured on the casing is less than the signal Pt measured on the body, it is considered that the partial discharge signal is likely to occur in the body area; If the signal measured on the main body Pt-△Pt<the signal measured on the casing Pb<the signal measured on the main body Pt+△Pt, △Pt represents the deviation limit, and the discharge amounts have reached the warning values, it is considered that there is a high probability that partial discharge signals exist in both the main body and the casing; Step 2: Block positioning: When the partial discharge is determined to occur in the transformer body area through the first step, the determination range of the partial discharge in the transformer body is narrowed by the block positioning method, including the following steps: The electromagnetic wave signal received by the UHF PD sensor is selected as the starting point of the discharge time, and the time Ti when the 8 ultrasonic PD sensors receive the PD signal is identified respectively, i=1, 2, 3, 4, 5, 6, 7, 8, and the block where the sensor MIN{Ti} is located is the block with the maximum probability of PD occurrence; Step 3: Precise positioning: According to the transformer structure, four ultrasonic partial discharge sensors with relatively short acoustic signal propagation paths are selected for precise positioning analysis. That is, for each block, sensors in the current block and the blocks directly adjacent to it are selected. The coordinates of the four ultrasonic partial discharge sensors are set as S1 (x1, y1, z1), S2 (x2, y2, z2), S3 (x3, y3, z3), and S4 (x4, y4, z4). The speed of the sound wave is expressed as v t It indicates that the time when different ultrasonic partial discharge sensors receive the partial discharge signal is Ti, i=1, 2, 3, 5; the coordinates of the discharge position inside the transformer are (x, y, z), and the discharge position (x, y, z) inside the transformer is obtained by solving the relationship between different ultrasonic partial discharge sensors and the discharge position.

6. The transformer fault comprehensive diagnosis method based on multiple types of partial discharge sensors according to claim 1 is characterized in that: The least square method is used to solve the relationship between different ultrasonic partial discharge sensors and discharge positions.

7. A transformer fault comprehensive diagnosis system based on multiple types of partial discharge sensors, characterized in that: include: A sensor unit, including various types of sensors; A data acquisition unit, used for acquiring a partial discharge signal of the transformer through a sensor unit; A data processing unit, used for extracting characteristic quantities of partial discharge signals of each sensor in the sensor unit; The data analysis unit is used to perform transformer discharge fault analysis based on the characteristic quantity of the partial discharge signal to achieve comprehensive transformer fault diagnosis.

Images