Generator stator partial discharge intelligent monitoring system and method

Abstract

The invention discloses a generator stator partial discharge intelligent monitoring system and method, and relates to the technical field of generator stator insulation state monitoring, and the system comprises an ultrasonic signal array monitoring terminal which is used for collecting ultrasonic signals in a generator stator; the multi-modal signal fusion monitoring terminal is used for receiving and processing the signals acquired by the plurality of sensors; and the background server is used for evaluating the severity of the partial discharge. The generator stator partial discharge intelligent monitoring method provided by the invention supports discharge type identification, defect positioning and insulation life prediction. High-precision detection of wide frequency bands is supported, the minimum discharge characteristic frequency of 80kHz can be identified, and the maximum discharge characteristic frequency of 200MHz can be identified. Multi-time-scale partial discharge information signal sampling is supported, the minimum acquisition step length interval is 1ns, and the maximum acquisition step length interval is 1s.

Description

Technical Field

[0001] The present invention relates to the technical field of generator stator insulation condition monitoring, and specifically to an intelligent monitoring system and method for partial discharge of a generator stator. Background Art

[0002] The traditional pulse current method (IEC 60270 standard) evaluates the insulation condition of equipment by detecting the current pulse signal generated by partial discharge, and measures the partial discharge signal of the stator through a coupling capacitor. However, the following problems still exist when this method is directly used for on-line monitoring of the stator:

[0003] Large noise interference: The operating environment of the generator is complex, and the electromagnetic noise, vibration noise overlap with the frequency band of the partial discharge signal, resulting in a low signal-to-noise ratio.

[0004] Severe signal attenuation: The stator winding structure is complex, and the discharge signal is easily attenuated and distorted when propagating to the sensor, affecting the detection sensitivity. As a result, it is difficult to cover the discharge characteristics of the entire length of the winding, and it is necessary to study the signal amplification and attenuation coupling characteristics and optimize the design of the acquisition module.

[0005] Data isolation: The traditional pulse current method relies on a single signal dimension (such as amplitude, phase). Once the signal measurement is lost or mismeasured, it will be difficult to distinguish the discharge type and defect location for the stator insulation discrimination. It is also difficult to unify the data units of the pulse current method and the on-line monitoring system directly between pC and mV, which hinders comprehensive analysis.

[0006] In summary, in view of the problems faced by the current partial discharge of the generator stator, an effective method suitable for the partial discharge state monitoring of large generator stators is needed to effectively improve the problems of large noise, limited measurement range, and no margin for data feasibility in the partial discharge monitoring of the generator stator, so as to facilitate power generation enterprises to give full play to the efficiency of digital-driven equipment energy management and better master the operating conditions of the generator. Summary of the Invention

[0007] In view of the above problems, the present invention is proposed.

[0008] Therefore, the technical problem solved by the present invention is: how to solve the common noise interference, signal attenuation and data isolation problems existing in the on-line monitoring of partial discharge in the stator by the pulse current method.

[0009] To solve the above technical problems, the present invention provides the following technical solutions: An intelligent monitoring system for partial discharge of a generator stator, comprising:

[0010] An ultrasonic signal array monitoring terminal for collecting discharge signals in the ultrasonic frequency band inside the generator stator; a multi-modal signal fusion monitoring terminal for receiving and processing the full-cycle stator conductor discharge signals with interference suppression strategies collected by multiple sensors; and,

[0011] A background server for evaluating the severity of partial discharge.

[0012] As a preferred embodiment of the intelligent monitoring system for partial discharge of a generator stator according to the present invention, wherein: the ultrasonic signal array monitoring terminal is composed of a flexible ultrasonic sensing array, a connecting cable and a terminal processor; the flexible ultrasonic sensing array is arranged on the surface of the generator stator housing, and adopts a multi-channel synchronous acquisition method, which can capture the synchronous phase by itself and obtain the ultrasonic partial discharge characteristics inside the generator.

[0013] As a preferred embodiment of the intelligent monitoring system for partial discharge of a generator stator according to the present invention, wherein: the multi-modal signal fusion monitoring terminal is composed of a classical coupling capacitor sensor, a vibration sensor and an optical fiber temperature sensor for multi-modal signal fusion acquisition; the broadband coupling capacitor is arranged at the stator outgoing line end in a double-ended differential installation manner, and the vibration sensor and the optical fiber temperature sensor are arranged on the surface of the stator insulation coating, and the coupling efficiency of the discharge signal collected by the optimized coupling capacitor is monitored by using non-electric quantity signals.

[0014] As a preferred embodiment of the intelligent monitoring system for partial discharge of a generator stator according to the present invention, wherein: the background server performs feature fusion on the full-frequency band pulse current signals, vibration spectra and temperature changes collected by the two terminals, and identifies the discharge mode through a convolutional neural network; based on the Markov chain model, combined with the historical discharge amount and pulse repetition rate, it predicts the remaining insulation life and generates a maintenance priority recommendation.

[0015] Another object of the present invention is to provide an intelligent monitoring method for partial discharge of a generator stator, which combines multi-modal signal fusion and ultrasonic signal array frequency division acquisition to achieve accurate, wide-band and multi-time scale extraction of partial discharge signals of the generator and insulation state evaluation.

[0016] To solve the above technical problems, the present invention provides the following technical solution: An intelligent monitoring method for partial discharge of a generator stator, comprising: using an ultrasonic sensor array arranged on the surface of the generator stator housing to collect ultrasonic signals inside the generator stator, and finally converging through multiple channels to an ultrasonic signal array monitoring terminal; installing two sets of capacitive couplers at the busbar rings at the outgoing ends of each phase winding of the generator stator, and maintaining an electrical distance of more than 2 m between the two sets of capacitive couplers to isolate grid-side noise; configuring multiple sets of multi-modal sensors for each unit, the multi-modal sensors including vibration sensing and temperature sensing, and connecting them to a multi-modal signal fusion monitoring terminal through coaxial cables; conditioning the signals collected by the ultrasonic signal array monitoring terminal and the multi-modal signal fusion monitoring terminal; converting the conditioned signals into digital signals and storing them; extracting features from the stored signals, performing edge computing, and extracting partial discharge features from the digital signals; using the sensor spatial arrangement position and time difference positioning algorithm to locate the partial discharge source; transmitting the positioning results of the ultrasonic signal array monitoring terminal and the multi-modal signal fusion monitoring terminal to the background server to complete the tasks of discharge type identification, defect location, and insulation life prediction through multi-dimensional analysis.

[0017] As a preferred embodiment of the intelligent monitoring method for partial discharge of the generator stator described in the present invention, wherein: the ultrasonic signal array monitoring terminal adopts an active filter circuit, and the circuit consists of an operational amplifier, a non-polar capacitor, and a resistor element to condition the signal of the flexible ultrasonic sensor; the output signal of the flexible ultrasonic sensor is processed by a low-pass filter and a high-pass filter, and the component parameters of the low-pass filter are set as R1 = 300 Ω, C1 = 1 nF; the component parameters of the high-pass filter are set as R2 = 1 kΩ, C2 = 2 nF; the operational amplifier circuit is used to amplify the signal amplitude, and the amplification factor is set to 2 times to obtain the filtered and amplified ultrasonic signal.

[0018] As a preferred embodiment of the intelligent monitoring method for partial discharge of the generator stator according to the present invention, the signal conditioning includes the use of an active filter circuit and a passive filter circuit. The active filter circuit consists of an operational amplifier, non-polar capacitors, and resistor elements to condition the power frequency 50 Hz synchronous electrical signal through a first signal conditioning circuit. The passive circuit consists of non-polar capacitors and resistor elements to condition the high-frequency partial discharge signal through a second signal conditioning circuit. As a preferred embodiment of the intelligent monitoring method for partial discharge of the generator stator according to the present invention, in the first signal conditioning circuit, the component parameters of the low-pass filter are set as R21 = 2 kΩ and C21 = 0.8 μF; the component parameters of the high-pass filter are set as R22 = 5 kΩ and C22 = 1 μF. The operational amplifier circuit is used to amplify the signal amplitude, and the amplification factor is set to 2 times to obtain the filtered and amplified ultrasonic signal. In the second signal conditioning circuit, the component parameters of the low-pass filter are set as R31 = 200 Ω and C31 = 4 pF; the component parameters of the high-pass filter are set as R32 = 1 kΩ and C32 = 4 pF.

[0019] As a preferred embodiment of the intelligent monitoring method for partial discharge of the generator stator according to the present invention, the extraction of partial discharge characteristics from the digital signal includes constructing a signal matrix for the signal array conditioned by the ultrasonic signal array monitoring terminal, expressed as

[0020]

[0021] where V α0 (t) represents the voltage value of the conditioned output signal at the 0° of the radial plane and near the excitation side of the generator stator at time t. The subsequent row vectors represent the conditioned values at 60° to 300° at time t. The subscripts β and γ respectively represent the position identifiers of the middle section and near the turbine side of the stator.

[0022] Define matrix I, expressed as

[0023] I = [0 e j60° e j120° -e j180° -e j240° -e j300°

[0024] where e is the natural constant and j represents the imaginary unit.

[0025] Multiply V o (t) by matrix I to obtain a calculation sequence over a period of time, and perform 0° calibration of the relative phase according to a preset rule. Monitor V o ​After (t)×I, for the zero elements among the obtained 3×N-dimensional elements, mark the time when the point at this position is at the 0° position of the relative phase. Then, combined with the displacement deviation of the three-phase ABC of the generator stator bar in the radial plane space, it is the true power frequency 0° phase point, and then the phase reference values of the complete power frequency signal are generated in sequence.

[0026] As a preferred scheme of the generator stator partial discharge intelligent monitoring method described in the present invention, wherein: the execution of the 0° calibration of the relative phase according to the preset rules includes reading and temporarily storing the V o (t) signal; entering the loop calculation of V o (t)×I to start searching for zero elements; judging if the sum of the elements of V o ×I is zero, then return the current position mark m and time mark i values, otherwise judge the next row of elements; at the i-th moment, after all m nodes are searched, then proceed to the (i + 1)-th point until all searches are completed; after the search is completed, if the vector sum of the elements at the m-th position at the i-th moment is zero, then it is considered that the acquisition at this position at this moment is the phase 0° point, and according to the power frequency and sampling frequency, it is easy to generate the synchronous voltage phase to provide the time domain coordinate reference system for the ultrasonic signal.

[0027] Advantages of the present invention: The generator stator partial discharge intelligent monitoring method provided by the present invention supports multi-dimensional analysis, supports discharge type identification, defect location, and insulation life prediction. It supports high-precision detection in a wide frequency band, and can identify the minimum discharge characteristic frequency of 80 kHz and the maximum discharge characteristic frequency of 200 MHz. It supports multi-time scale sampling of partial discharge information signals, with the minimum acquisition step interval of 1 ns and the maximum acquisition step interval of 1 s. It discloses a conditioning circuit applicable to ultrasonic array signals and high-frequency coupling capacitor signals, providing a reference for related technical applications. The monitoring scheme has strong engineering applicability and is compatible with newly installed and retrofitted units. Description of the Drawings

[0028] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings without creative efforts based on these drawings.

[0029] Figure 1 It is the overall schematic diagram of the generator stator partial discharge intelligent monitoring system and method based on multi-modal signal fusion and ultrasonic signal array provided by an embodiment of the present invention.

[0030] Figure 2 It is the appearance schematic and laying method schematic diagram of the ultrasonic signal array sensor provided by an embodiment of the present invention.

[0031] Figure 3 The signal conditioning circuit diagram of the ultrasonic signal array sensor provided by an embodiment of the present invention.

[0032] Figure 4 The signal conditioning circuit diagram of the coupled capacitor sensing signal provided by an embodiment of the present invention.

[0033] Figure 5 The simulation diagram of the implementation effect result of the coupled capacitor sensing signal conditioning circuit provided by an embodiment of the present invention. Detailed implementation manners

[0034] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will describe the specific implementation manners of the present invention in detail with reference to the accompanying drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0035] Many specific details are set forth in the following description to facilitate a full understanding of the present invention. However, the present invention may be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0036] Embodiment 1

[0037] Refer to Figure 1 - Figure 2 , which is an embodiment of the present invention in system deployment. A plurality of flexible ultrasonic sensors are arranged on the surface of the generator stator housing to form an ultrasonic array. In this example, one point is arranged at intervals of 60° in the spatial angle from the axis in the radial plane of the generator, and radial plane sensing is respectively arranged at the front, middle, and rear of the generator in the axial plane, with a total of 18 ultrasonic partial discharge sensors / machine. The signals of the ultrasonic sensors are connected to the monitoring terminal 1. Capacitive couplers (80 pF, 40 - 200 MHz) are installed at the busbars of each phase winding of the generator stator, and 12 groups of sensors are configured for each unit, and are connected to the monitoring terminal 2 through coaxial cables. Vibration sensors are arranged at the stator slot wedges, and the optical fiber temperature measurement unit is embedded in the bar insulation layer, and the data is transmitted to the terminal 2 through the CAN bus.

[0038] The terminal 1 and the terminal 2 are arranged as shown in Figure 1 . The signal processing results after edge computing are transmitted to the background server through an Ethernet cable, and then the processed results are handed over to the human-machine interaction interface for display.

[0039] The characteristics of terminal 1 and terminal 2 are that terminal 1 processes partial discharge signals within 500 kHz, and the general DSP and operational amplifier circuits can meet the acquisition requirements; terminal 2 processes partial discharge signals within 200 MHz, and the minimum hardware system of FPGA + multi-CPU with high processing capabilities and a high-speed acquisition board can meet the computing requirements.

[0040] Embodiment 2

[0041] Refer to Figure 3 - Figure 5 , which is an embodiment of the present invention in terms of signal conditioning. For the output signal of the flexible ultrasonic sensor, it is designed according to the signal conditioning circuit shown in Figure 3 . Since the amplitude ratio of the ultrasonic signal is low and the acquisition rate is not high, active filtering is used to improve the signal processing efficiency. This type of signal conditioning circuit consists of operational amplifiers, non-polar capacitors, and resistor elements. Among them, V in is the measurement signal of a single flexible ultrasonic sensor, and V o represents the conditioned ultrasonic signal of V in . The low-pass filter is placed at the first stage of signal conditioning, and the key component parameters meet R1 = 300 Ω and C1 = 1 nF; the high-pass filter is placed at the second stage of signal conditioning, and the key component parameters meet R2 = 1 kΩ and C2 = 2 nF. Both operational amplifiers are fixedly used with 1 kΩ resistors to form an operational amplifier circuit to complete the conditioning of doubling the signal. The signal conditioning of terminal 1 does not have a dedicated synchronous voltage acquisition circuit, but uses the spatial matrix signal of the multi-ultrasonic array to adaptively complete phase calibration, optimizing and reducing the computing burden of the terminal and the size of the device.

[0042] For the signal array conditioned by monitoring terminal 1, the following signal matrix can be easily written:

[0043]

[0044] In the formula, V α0 (t) represents the voltage value of the conditioned output signal at the radial plane of 0° and the near-excitation side of the generator stator at time t, and the row vector successively represents the conditioned values at 60° to 300° at time t; the subscripts β and γ respectively represent the position identifiers of the middle section of the stator and the near-turbine side.

[0045] In this example, matrix I is defined as:

[0046] I = [0 e j60° e j120° -e j180° -e j240° -e j300° (2)

[0047] In the formula, e is the natural constant, and the signals at different positions are unified by the spatial angle for easy processing and calculation; j represents the imaginary unit.

[0048] Multiply V o (t) by the matrix I to obtain a calculation sequence of N sampling points at a certain time. The 0° calibration of the relative phase can be carried out according to the following rules:

[0049] S1: Read and temporarily store V o (t) signal for a certain time. For example, there are 100 points (N = 100) within this time window;

[0050] S2: Enter the loop to calculate V o (t)×I and start searching for zero elements. First, enter the i-th moment. First, calculate the 6 sensors arranged in the front section of the generator. The description of the front section position is recorded as m = 1 (the middle section position m = 2, the rear section m = 3), that is, judge and retrieve the sum of the elements in the first row of the matrix V o ;

[0051] S3: Judge if the sum of the elements of V o ×I is zero, then return the current position mark m and time mark i value. Otherwise, judge the next row of elements.

[0052] S4: After all m nodes are searched at the i-th moment, then repeat S2 - S3 for the (i + 1)-th point until all searches are completed.

[0053] S5: After the search is completed, if the element vector sum at the i-th moment at the m-th position is zero, then it is considered that the acquisition at this position at this moment is the phase 0° point. According to the power frequency and sampling frequency, it is easy to generate a synchronous voltage phase to provide a time domain coordinate reference system for the ultrasonic signal.

[0054] It should be noted that a certain time refers to a certain time of N sampling points.

[0055] Monitor V o (t)×I. After obtaining the zero elements in the 3×N-dimensional elements, calibrate that the point at this moment and this position is at the 0° position of the relative phase. Then, combined with the displacement deviation from the three-phase ABC of the generator stator bar in the radial plane space, it is the true power frequency 0° phase point. Then, generate the phase reference value of the complete power frequency signal in sequence for use in the result analysis of the ultrasonic array signal in partial discharge identification.

[0056] For the vibration sensor three-axis acceleration signal and the fiber optic temperature measurement unit temperature signal collected in the terminal 2, there are no specific conditioning parameters required and can be configured according to needs. For the coupling capacitor sensing signal, design according to Figure 4 the shown signal conditioning circuit. This type of signal conditioning circuit consists of an operational amplifier, a non-polar capacitor, and a resistor element. Among them, V in is the coupling capacitor sensing signal, and V o1 represents the V inThe separated power frequency signal of the synchronous voltage. The low-pass filter in this branch is placed at the first stage of the branch signal conditioning. The key components of the low-pass filter refer to the RC filter composed of a resistor component and a capacitor component. The parameters of the resistor and capacitor involved satisfy R21 = 2kΩ and C21 = 0.8μF. The high-pass filter is placed at the second stage of the branch signal conditioning. The key components of the high-pass filter refer to the RC filter composed of a resistor component and a capacitor component. The parameters of the resistor and capacitor involved satisfy R22 = 5kΩ and C22 = 1μF. Both operational amplifiers are fixedly used with 1kΩ resistors to form an operational amplifier loop to complete the amplification and conditioning of the signal by 2 times. V o2 Indicates for V in The separated partial discharge signal. The low-pass filter in this branch is placed at the first stage of the branch signal conditioning. The parameters of the key components of the resistor and capacitor involved satisfy R31 = 200Ω and C31 = 4pF. The high-pass filter is placed at the second stage of the branch signal conditioning. The parameters of the key components of the resistor and capacitor involved satisfy R32 = 1kΩ and C32 = 4pF. The amplification and conditioning of the operational amplifier are not adopted for the signal of this branch. The reason is that when the signal frequency is too high, the processing ability of the operational amplifier is saturated, which is likely to cause the loss of a single-point signal in the time domain and affect the partial discharge recognition result.

[0057] To illustrate the advancement of the signal conditioning of the present invention, a periodic signal containing 20MHz, 90MHz, and 300MHz is physically simulated in a high-precision signal generator and oscilloscope, and a simulated 100MHz fluctuating partial discharge signal is superimposed. After recording this signal, the effect of this signal conditioning circuit is tested in the simulation link. Figure 5 (a) shows the collected original signal, Figure 5 (b) shows the Figure 4 V in the circuit shown o2 The simulation curve of the conditioned signal after the branch circuit conditioning and separation. The results show that the partial discharge signal conditioned by the conditioning circuit of the present invention can effectively extract requirements such as pulse amplitude, rise time, and frequency band screening.

[0058] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not restrictive. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. An intelligent monitoring system for partial discharge of a generator stator, characterized in that: include: Ultrasonic signal array monitoring terminal, used to collect the discharge signal of the ultrasonic frequency band inside the generator stator; A multi-modal signal fusion monitoring terminal for receiving and processing stator conductor discharge full-cycle signals with interference suppression strategies collected by multiple sensors; and The backend server is used to evaluate the severity of partial discharge.

2. The generator stator partial discharge intelligent monitoring system according to claim 1, characterized in that: The ultrasonic signal array monitoring terminal is composed of a flexible ultrasonic sensor array, a connecting cable and a terminal processor; The flexible ultrasonic sensor array is arranged on the surface of the generator stator shell and adopts a multi-channel synchronous acquisition method to capture the synchronous phase by itself and obtain the discharge characteristics of the ultrasonic section inside the generator.

3. The generator stator partial discharge intelligent monitoring system according to claim 2, characterized in that: The multimodal signal fusion monitoring terminal is composed of a classic coupling capacitance sensor, a vibration sensor, and an optical fiber temperature sensor to form a multimodal signal fusion acquisition; The broadband coupling capacitor is arranged at the stator output terminal in a double-ended differential installation manner, and the vibration sensor and the optical fiber temperature sensor are arranged on the surface of the stator insulation coating. The coupling efficiency of the discharge signal collected by the coupling capacitor is optimized by monitoring the non-electrical signal.

4. The generator stator partial discharge intelligent monitoring system according to claim 3, characterized in that: The backend server performs feature fusion on the full-band pulse current signals, vibration spectrum, and temperature changes collected by the two terminals, and identifies the discharge mode through a convolutional neural network; Based on the Markov chain model, the remaining insulation life is predicted by combining the historical discharge amount and pulse repetition rate, and maintenance priority recommendations are generated.

5. A method for intelligent monitoring of partial discharge of a generator stator, characterized in that: include: The ultrasonic sensor array arranged on the surface of the generator stator shell is used to collect the ultrasonic signal inside the generator stator, and the multi-channel signals are finally converged to the ultrasonic signal array monitoring terminal; Two sets of capacitor couplers are installed at the collector ring at the outlet end of each phase winding of the generator stator, and an electrical distance of more than 2m is reserved between the two sets of capacitor couplers to isolate the grid-side noise; each unit is equipped with multiple sets of multi-modal sensors, including vibration sensors and temperature sensors, which are connected to the multi-modal signal fusion monitoring terminal through coaxial cables; Perform signal conditioning on the collected signals of the ultrasonic signal array monitoring terminal and the multi-modal signal fusion monitoring terminal; Convert the conditioned signal into a digital signal and store it; Perform feature extraction on the stored signal, perform edge computing, and extract partial discharge features from the digital signal; The local discharge source is located by using the spatial arrangement of sensors and the time difference positioning algorithm; The positioning results of the ultrasonic signal array monitoring terminal and the multimodal signal fusion monitoring terminal are transmitted to the background server for multi-dimensional analysis to complete the tasks of discharge type identification, defect location and insulation life prediction.

6. The method for intelligent monitoring of partial discharge of a generator stator according to claim 5, characterized in that: The ultrasonic signal array monitoring terminal adopts an active filtering circuit, which is composed of an operational amplifier, a non-polar capacitor and a resistor element to form a conditioning circuit to condition the flexible ultrasonic sensor signal; The output signal of the flexible ultrasonic sensor is processed by a low-pass filter and a high-pass filter, wherein the component parameters of the low-pass filter are set to R1=300Ω, C1=1nF; the component parameters of the high-pass filter are set to R2=1kΩ, C2=2nF; The operational amplifier loop is used to amplify the signal amplitude, and the amplification factor is set to 2 times to obtain a filtered and amplified ultrasonic signal.

7. The method for intelligent monitoring of partial discharge of a generator stator according to claim 6, characterized in that: The signal conditioning includes using active filter circuits and passive filter circuits; The active filter circuit is composed of an operational amplifier, a non-polar capacitor and a resistor element. The first signal conditioning circuit conditions the industrial frequency 50Hz synchronous electrical signal. The passive circuit is composed of a non-polar capacitor and a resistor element. The second signal conditioning circuit conditions the high-frequency partial discharge signal.

8. The method for intelligent monitoring of partial discharge of a generator stator according to claim 7, characterized in that: In the first signal conditioning circuit, the component parameters of the low-pass filter are set to R21=2kΩ, C21=0.8μF; the component parameters of the high-pass filter are set to R22=5kΩ, C22=1μF; the operational amplifier loop is used to amplify the signal amplitude, and the amplification factor is set to 2 times to obtain a filtered and amplified ultrasonic signal; In the second signal conditioning circuit, the component parameters of the low-pass filter are set to R31=200Ω, C31=4pF; the component parameters of the high-pass filter are set to R32=1kΩ, C32=4pF.

9. The method for intelligent monitoring of partial discharge of a generator stator according to claim 8, characterized in that: The extracting of local discharge features from digital signals includes constructing a signal matrix for the signal array conditioned by the ultrasonic signal array monitoring terminal, which is expressed as: Among them, V α0 (t) represents the voltage value of the conditioned output signal at the radial plane 0° and the generator stator near the excitation side at time t, and the row vectors behind it represent the conditioned values ​​at 60° to 300° at time t; the subscripts β and γ represent the position marks of the middle section of the stator and the side near the steam turbine, respectively; Define the matrix I, expressed as, I=[0 and j60° And j120° -And j180° -And j240° -And j300° ] Among them, e is a natural constant, j represents an imaginary unit; V o (t) is multiplied by the matrix I to obtain a calculation sequence for a period of time, and the 0° calibration of the relative phase is performed according to the preset rules; Monitor V o (t)×I, the zero elements in the obtained 3×N-dimensional elements are calibrated at this moment. The point at this position is at the 0° position of the relative phase. Combined with the displacement deviation of the three-phase ABC of the generator stator bar in the radial plane space, it is the real power frequency 0° phase point, and then the phase reference value of the complete power frequency signal is generated in sequence.

10. The method for intelligent monitoring of partial discharge of a generator stator according to claim 9, characterized in that: The 0° calibration of the relative phase according to the preset rule includes reading and temporarily storing V o (t) signal; Enter the loop to calculate V o (t)×I starts searching for zero elements; If there is V o If the sum of the elements of ×I is zero, return the current position mark m and time mark i value, otherwise determine the next row element; At the i-th moment, after all m nodes are searched, the i+1th point is searched until all the nodes are searched; After the search is completed, at the mth position, the sum of the element vectors at the ith moment is zero, then the acquisition of this position at this moment is considered to be a phase 0° point. According to the power frequency and the sampling frequency, it is easy to generate a synchronous voltage phase to provide a time domain coordinate reference system for the ultrasonic signal.

Images