Hydrogenerator stator winding insulation peak valley adaptive filtering prediction method and system

Abstract

The invention discloses a hydro-generator stator winding insulation peak-valley adaptive filtering prediction method and system, and belongs to the technical field of power equipment optimization control. The objective of the invention is to solve the problem of accurate prediction of dynamic characteristics of stator winding insulation of a hydro-generator under peak-valley load conditions. A data acquisition module acquires a temperature signal, a vibration signal and insulation resistance of a stator winding in real time; the self-adaptive filtering processing module receives the signal, fuses the temperature signal and the vibration signal to generate a composite reference signal, and performs self-adaptive filtering processing based on the obtained composite reference signal to obtain an output signal after self-adaptive filtering processing; the insulation state prediction module performs feature extraction, calculates an insulation state index based on the extracted features, and performs trend prediction on insulation parameters by adopting an autoregression integral moving average model to obtain an insulation resistance prediction curve; and an optimization control module triggers a load adjustment scheme to generate an output control instruction based on the insulation resistance prediction curve obtained in the step S3.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of optimal control of power equipment, and particularly relates to a method and system for adaptive filtering and prediction of insulation peaks and valleys of a stator winding of a hydro-generator. BACKGROUND

[0002] The running state of the insulation system of the stator winding of the hydro-generator directly affects the efficiency and service life of the unit, and the monitoring and control of the insulation performance in the prior art mostly adopt static threshold judgment or a prediction method based on a fixed model. Under the working condition of alternating peak and valley loads, the conventional technology has the following technical defects: first, the conventional insulation monitoring system cannot automatically adjust the detection parameters according to the dynamic load change, resulting in lagging of the insulation aging feature recognition; second, the existing filtering algorithm generally adopts a preset cutoff frequency, and it is difficult to adapt to the electromagnetic interference characteristics in different peak and valley stages, causing cumulative errors in the winding state evaluation; third, the insulation parameter prediction and the operation optimization link are independent of each other, and lack of real-time feedback control mechanism, resulting in the risk of partial discharge of the unit when the load suddenly changes.

[0003] In the current technical solution, the insulation state evaluation mostly depends on periodic offline detection, and it is impossible to realize continuous tracking of the insulation performance in the running process. Although the traditional filtering methods such as Butterworth filter or Kalman filter can suppress noise in a specific frequency band, the fixed parameter characteristic thereof leads to signal distortion when the load fluctuation exceeds the design threshold, thereby affecting the accurate extraction of the insulation deterioration characteristics. The existing prediction model generally adopts linear regression or time series analysis, and it is difficult to handle the nonlinear relationship of the winding temperature gradient, dielectric loss and mechanical vibration and other multi-physical field coupling. The non-planned shutdown events caused by inaccurate insulation performance prediction are common in industry practice, and the adaptability defects of the traditional method under complex working conditions are exposed. SUMMARY

[0004] The problem to be solved by the application is to realize accurate prediction of the dynamic characteristics of the insulation of the stator winding of the hydro-generator under the peak and valley load working condition, and a method and system for adaptive filtering and prediction of insulation peaks and valleys of a stator winding of a hydro-generator are provided.

[0005] To achieve the above object, the application realizes the following technical solutions:

[0006] A method for adaptive filtering and prediction of insulation peaks and valleys of a stator winding of a hydro-generator, comprising the following steps:

[0007] S1. A data acquisition module acquires temperature signals, vibration signals and insulation resistance of the stator winding in real time;

[0008] S2. The adaptive filtering processing module receives the temperature signal and the vibration signal obtained in step S1, then fuses the temperature signal and the vibration signal to generate a composite reference signal, and performs adaptive filtering processing based on the obtained composite reference signal to obtain an output signal after adaptive filtering processing;

[0009] S3. The insulation state prediction module receives the output signal after adaptive filtering processing obtained in step S2 and the insulation resistance collected in step S1, performs feature extraction, calculates an insulation state index based on the extracted features, and performs trend prediction on the insulation parameter by using an autoregressive integrated moving average model ARIMA to obtain an insulation resistance prediction curve;

[0010] S4. The optimization control module triggers a load adjustment scheme generation output control instruction based on the insulation resistance prediction curve obtained in step S3.

[0011] Further, in step S1, 12 groups of PT100 platinum resistors are arranged in a ring array at key temperature measurement points of the stator winding end and the slot part to collect temperature signals of the stator winding; the sampling frequency of the vibration signal is set to 10 kHz; the insulation resistance is measured by using a high-voltage megohm meter and a leakage current detection circuit in parallel structure, and the measurement voltage range is 0-5 kV.

[0012] Further, the specific implementation method of step S2 includes the following steps:

[0013] S2.1. The adaptive filtering processing module receives the temperature signal and the vibration signal obtained in step S1, then fuses the temperature signal and the vibration signal to generate a composite reference signal, and performs adaptive filtering processing based on the obtained composite reference signal to obtain an output signal after adaptive filtering processing;

[0014]

[0015] wherein, is the composite reference signal, is a vibration signal weighting coefficient, is a temperature change rate weight factor, is a temperature signal difference, is an amplitude sequence of the vibration signal at the n-k moment;

[0016]

[0017] wherein, is a temperature sampling value at the current moment, is a temperature sampling value at the previous moment;

[0018] S2.2. The obtained composite reference signal is subjected to adaptive filtering processing by using a transverse FIR filter to obtain an output signal after adaptive filtering processing , and the expression is:

[0019]

[0020] wherein, is the mth order time-varying filter coefficient, is the filter order, is the mth order time-varying filter coefficient, is the mth order lagged composite reference signal;

[0021] The time-varying filter coefficients are updated using a normalized LMS algorithm, and the update formula is:

[0022]

[0023] wherein, is the mth order time-varying filter coefficient, is the mth order time-varying filter coefficient, is the mth order time-varying filter coefficient, is the convergence factor, is the mth order time-varying filter coefficient, is the error signal, is the mth order time-varying filter coefficient, is the correction quantity to prevent division by zero.

[0024] Further, the specific implementation method of step S3 includes the following steps:

[0025] S3.1. The insulation state prediction module receives the output signal after adaptive filtering obtained in step S2 and the insulation resistance collected in step S1, performs feature extraction, and extracts the peak-to-peak value of the vibration signal , temperature gradient , insulation resistance change rate ;

[0026] S3.2. Based on the extracted features, the exponential smoothing model is used to calculate the insulation state index , and the expression is:

[0027]

[0028] wherein, is the smoothing coefficient, is the normalized weighting function, is the insulation state index at the previous time; S3.3. The autoregressive integrated moving average model ARIMA is used to perform trend prediction on the insulation state index time series calculated in step S3.2, and the predicted index at the future time is obtained

[0029] .

[0030] Further, when the predicted insulation index obtained in step S4 is lower than the insulation index threshold , ​, triggering the load adjustment scheme to generate output control instructions, the output control instructions including winding current limit adjustment instructions, cooling system start-stop instructions, generator output gradient control instructions:

[0031] S4.1. When >0.85, in normal state, no operation;

[0032] S4.2. When 0.7 0.85, in slight degradation state, start the cooling system start-stop instructions to reduce the winding temperature rise;

[0033] S4.3. When 0.5 0.7, in moderate degradation state, trigger the winding current limit adjustment instructions to reduce the excitation current by 5-10%;

[0034] S4.4. When 0 0.5, in serious degradation state, simultaneously trigger the output gradient control instructions and the cooling system intensive mode, and if necessary, issue an alarm signal or plan a shutdown inspection.

[0035] A hydroelectric generator stator winding insulation peak-valley adaptive filtering prediction system is realized by the hydroelectric generator stator winding insulation peak-valley adaptive filtering prediction method, and comprises a data acquisition module, an adaptive filtering processing module, an insulation state prediction module and an optimization control module.

[0036] The beneficial effects of the present application are:

[0037] The hydroelectric generator stator winding insulation peak-valley adaptive filtering prediction method can effectively reduce the winding insulation aging rate, improve the operation stability of the unit under peak-valley load conditions, reduce the maintenance downtime, and prolong the service life of the equipment. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 The flowchart of the hydroelectric generator stator winding insulation peak-valley adaptive filtering prediction method is shown. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application, that is, the described specific embodiments are only a part of the embodiments of the present application, but not all the specific embodiments. The components of the specific embodiments of the present application generally described and shown in the drawings can be arranged and designed in various different configurations, and the present application can also have other embodiments.

[0040] Therefore, the detailed description of the specific embodiments of the present application provided below in the drawings is not intended to limit the scope of the claimed present application, but only represents selected specific embodiments of the present application. Based on the specific embodiments of the present application, all other specific embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0041] In order to further understand the invention content, characteristics and effects of the present application, the following specific embodiments are exemplified, and the drawings are combined Figure 1 The detailed description is as follows:

[0042] Example 1:

[0043] A kind of hydroelectric generator stator winding insulation peak and valley adaptive filtering prediction method, comprising the following steps:

[0044] S1. Data acquisition module real-time acquisition temperature signal, vibration signal and insulation resistance of stator winding;

[0045] Further, in step S1, 12 groups of PT100 platinum resistors are arranged in a ring array, arranged at the key temperature measuring points of the end and slot part of the stator winding to collect the temperature signal of the stator winding;The sampling frequency of the vibration signal is set to 10 kHz;High-voltage megohmmeter and leakage current detection circuit are connected in parallel to measure the insulation resistance, and the measurement voltage range is 0-5kV.

[0046] S2. The adaptive filtering processing module receives the temperature signal and vibration signal obtained in step S1, then fuses the temperature signal and vibration signal to generate a composite reference signal, and performs adaptive filtering processing based on the obtained composite reference signal to obtain an output signal after adaptive filtering processing;

[0047] Further, the specific implementation method of step S2 includes the following steps:

[0048] S2.1. The adaptive filtering processing module receives the temperature signal and vibration signal obtained in step S1, then fuses the temperature signal and vibration signal to generate a composite reference signal, and the expression is:

[0049]

[0050] wherein, is a composite reference signal, is a vibration signal weighting coefficient, is a temperature change rate weight factor, is a temperature signal difference, is a vibration signal amplitude sequence at the n-k moment;

[0051]

[0052] wherein, is a temperature sampling value at the current moment, is a temperature sampling value at the last moment;

[0053] S2.2. Adopting a transverse FIR filter to perform adaptive filtering processing on the obtained composite reference signal to obtain an output signal after adaptive filtering processing , the expression is:

[0054]

[0055] wherein, is a time-varying filter coefficient of the mth order, is a filter order, is a time-varying filter coefficient of the mth order, is a lagged composite reference signal of the mth order;

[0056] The time-varying filter coefficient is updated by using a normalized LMS algorithm, and the update formula is:

[0057]

[0058] wherein, is an updated time-varying filter coefficient of the mth order, , is a convergence factor, , is an error signal, , is a correction amount for preventing zero division.

[0059] S3. The insulation state prediction module receives the output signal after adaptive filtering processing obtained in step S2, performs feature extraction, calculates an insulation state index based on the extracted features, and performs trend prediction on the insulation parameter by using an autoregressive integral moving average model ARIMA to obtain an insulation resistance prediction curve;

[0060] Further, the specific implementation method of step S3 includes the following steps:

[0061] ​S3.1. The insulation state prediction module receives the output signal after adaptive filtering obtained in step S2 and the insulation resistance collected in step S1, performs feature extraction, and extracts the peak-to-peak value of the vibration signal , temperature gradient , insulation resistance change rate ;

[0062] S3.2. Based on the extracted features, an exponential smoothing model is used to calculate the insulation state index , and the expression is:

[0063]

[0064] wherein , is a smoothing coefficient, is a normalized weighting function, is the insulation state index at the previous moment;

[0065] S3.3. An autoregressive integrated moving average model ARIMA is used to perform trend prediction on the insulation state index time series obtained in step S3.2, and the predicted index at the future moment is obtained .

[0066] S4. The optimization control module triggers the load adjustment scheme generation output control instruction based on the insulation resistance prediction curve obtained in step S3.

[0067] Further, when the predicted insulation index obtained in step S4 is lower than the insulation index threshold , , the load adjustment scheme generation output control instruction is triggered, and the output control instruction includes winding current limit adjustment instruction, cooling system start-stop instruction, and generator output gradient control instruction:

[0068] S4.1. When > 0.85, it is in a normal state, and no operation is performed.

[0069] S4.2. When 0.7 < 0.85, it is in a slight degradation state, and the cooling system start-stop instruction is started to reduce the winding temperature rise.

[0070] S4.3. When 0.5 < 0.7, it is in a moderate degradation state, and the winding current limit adjustment instruction is triggered to reduce the excitation current by 5–10%.

[0071] S4.4. When 0 0.5 indicates a severely degraded state, triggering both the output gradient control command and the cooling system enhancement mode. If necessary, an alarm signal will be issued or a planned shutdown inspection will be initiated.

[0072] Furthermore, the load adjustment scheme also includes time windows where the current operating period falls within a period of more than 30% difference between peak and off-peak electricity prices; and unit load volatility. Lasting 10 minutes;

[0073] Furthermore, when An emergency shutdown protocol is triggered when the sampling period is below 0.7 for three consecutive sampling periods.

[0074] Example 2:

[0075] A peak-valley adaptive filtering prediction system for stator winding insulation of a hydro-generator is implemented based on the peak-valley adaptive filtering prediction method for stator winding insulation of a hydro-generator described in Example 1. The system includes a data acquisition module, an adaptive filtering processing module, an insulation state prediction module, and an optimization control module, which are connected in sequence.

[0076] Furthermore, each module employs a dual redundancy design, with anomaly detection and automatic switching mechanisms implemented in critical signal channels. The filter coefficient update cycle is synchronized with the data acquisition module's sampling cycle, and the control command output delay is less than 50ms. All algorithm parameters can be calibrated and mode switched online via the user interaction module.

[0077] Furthermore, the adaptive filtering module calculates the frequency domain characteristics of the interference signal in real time using Fast Fourier Transform; and updates the filter coefficient matrix based on the frequency domain analysis results. ,in Indicates the first The weight parameters of the first-order filter are determined; the filter parameters are updated using a recursive least squares algorithm, the recursive formula of which is:

[0078]

[0079] in, Here is the Kalman gain matrix. For the first The prediction error at any given time is set to an integer multiple of the generator rotor rotation period; the digital filter bank of the adaptive filtering module in S2 includes: a fundamental frequency tracking bandpass filter, whose center frequency changes synchronously with the generator rated frequency; a high-frequency harmonic suppression filter, whose cutoff frequency is dynamically adjusted according to the spectral characteristics of the partial discharge signal; and a low-frequency drift compensation filter, which uses a zero-phase filtering method to eliminate baseline drift.

[0080] Further, the insulation state prediction module also performs peak-valley feature analysis, and constructs a probability distribution function of the peak-valley feature based on a kernel density estimation method:

[0081]

[0082] wherein is a Gaussian kernel function, is a bandwidth parameter, is a historical insulation parameter sampling value;

[0083] According to the power grid dispatching instruction and the insulation prediction result, the winding operation parameter is dynamically adjusted, and a target function thereof is:

[0084]

[0085] wherein, is a loss power, is a voltage fluctuation amount, is a partial discharge intensity, is a dynamic weight coefficient;

[0086] Further, the optimization control module dynamically reduces the generator output power according to the insulation state prediction value at the power grid load peak period, and a reduction amplitude calculation formula is:

[0087]

[0088] wherein, is a winding temperature rise prediction value, is a dielectric loss tangent measured value, and is an adjustment coefficient; the adjustment coefficient and is dynamically adjusted through a fuzzy control algorithm, and a membership function thereof adopts a triangular distribution:

[0089]

[0090]

[0091] wherein, respectively represent preset threshold interval boundaries of the characteristic parameters.

[0092] It has to be noted that the terms "first", "second", and the like in connection with an entity or action refer to this entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without further constraints, exclude the presence of additional elements of the process, method, article, or apparatus.

[0093] While the application has been described with reference to specific implementations thereof, it should be understood that various modifications and substitutions can be made by those skilled in the art without departing from the scope of the present application. In particular, any one of the features of the present application disclosed above can be utilized independently of any other and the scope of the application should not be limited by the specific embodiments disclosed herein, but should be given the widest coverage possible in its true scope.

Claims

1. A method for predicting the crest and trough of the insulation of a hydroelectric generator stator winding, characterized by, It comprises the following steps: S1. The data acquisition module collects the temperature signal, vibration signal and insulation resistance of the stator winding in real time; S2. The adaptive filtering processing module receives the temperature signal and vibration signal obtained in step S1, then fuses the temperature signal and vibration signal to generate a composite reference signal, performs adaptive filtering processing based on the obtained composite reference signal, and obtains an output signal after adaptive filtering processing; S3. The insulation state prediction module receives the output signal after adaptive filtering processing obtained in step S2 and the insulation resistance collected in step S1, performs feature extraction, calculates an insulation state index based on the extracted features, and uses an autoregressive integrated moving average model ARIMA to perform trend prediction on the insulation parameter to obtain an insulation resistance prediction curve; S4. The optimization control module triggers a load adjustment scheme generation output control instruction based on the insulation resistance prediction curve obtained in step S3.

2. A method for adaptive filtering and forecasting of peak and valley of stator winding insulation of a hydro-generator according to claim 1, characterized in that, In step S1, 12 groups of PT100 platinum resistors are arranged in a ring array at key temperature measuring points of the stator winding end and slot part to collect the temperature signal of the stator winding; the sampling frequency of the vibration signal is set to 10 kHz; the insulation resistance is measured by using a high-voltage megohm meter and a leakage current detection circuit in parallel structure, and the measurement voltage range is 0-5 kV.

3. A method of predicting the peak-to-valley of the insulation of a stator winding of a hydro-generator according to claim 2, characterised in that, The specific implementation method of step S2 comprises the following steps: S2.

1. The adaptive filtering processing module receives the temperature signal and vibration signal obtained in step S1, then fuses to generate a composite reference signal, and the expression is: wherein, is a composite reference signal, is a vibration signal weighting coefficient, is a temperature change rate weight factor, is a difference of temperature signals, is a vibration signal amplitude sequence at the n-k moment. wherein, is a temperature sample value of the current time, is a temperature sample value of the previous time; S2.

2. Adopting a transverse FIR filter to perform adaptive filtering processing on the obtained composite reference signal to obtain an output signal after adaptive filtering processing The expression is: in, For the first Time-varying filter coefficients, Let the filter order be . The m-th order hysteresis composite reference signal; The time-varying filter coefficient is updated by using a normalized LMS algorithm, and the update formula is: wherein is an updated first order time-varying filter coefficient, , is a convergence factor, , is an error signal, , is a correction to prevent division by zero.

4. The method of claim 3, wherein the method is characterized by: The specific implementation method of step S3 comprises the following steps: S3.

1. The insulation state prediction module receives the output signal after adaptive filtering obtained in step S2 and the insulation resistance collected in step S1, performs feature extraction, and extracts the peak-to-peak value of the vibration signal , temperature gradient , insulation resistance change rate ; S3.

2. Based on the extracted features, an exponential smoothing model is used to calculate the insulation condition index , the expression is: wherein, , is a smoothing coefficient, is a normalized weighting function, is the insulation condition index of the previous time instant; S3.

3. Adopting autoregressive integrated moving average model ARIMA to make trend prediction on the insulation condition index time series calculated in step S3.2, to obtain the predicted index at future time .

5. A method of predicting the peak-to-valley of the insulation of a stator winding of a hydro-generator according to claim 4, characterised in that, the predicted insulation index obtained in step S4 below the insulation index threshold , , the load adjustment scheme triggers generation of output control instructions, including winding current limit adjustment instructions, cooling system start-stop instructions, generator power output gradient control instructions: S4.

1. When > 0.85, normal state, no operation; S4.

2. When 0.7 < x < 1.0, then When 0.85, in a slight degradation state, start the cooling system start-stop instruction to reduce the winding temperature rise; S4.

3. When 0.5 < x < 1.5, then 0.7, in a moderate degradation state, trigger the winding current limit adjustment instruction to reduce the excitation current by 5-10%; S4.

4. When 0 < t < T 0.5, in a serious degradation state, trigger the force gradient control instruction and the cooling system strengthening mode, and issue an alarm signal or plan to shut down for inspection if necessary.

6. A hydroelectric generator stator winding insulation peak and valley adaptive filtering prediction system, relying on a hydroelectric generator stator winding insulation peak and valley adaptive filtering prediction method according to any one of claims 1-5, characterized in that, It comprises a data acquisition module, an adaptive filtering processing module, an insulation state prediction module and an optimization control module, and the data acquisition module, the adaptive filtering processing module, the insulation state prediction module and the optimization control module are connected in sequence.

Images