A method and system for monitoring the state of a generator excitation system

Abstract

The present invention belongs to the technical field of generator excitation system monitoring. The present invention provides a generator excitation system state monitoring method and system, including: performing spatial temperature change analysis on the inside of the rotor winding, screening out winding parts in a two-point grounding state, performing stability analysis on the transition trends corresponding to the winding parts in the two-point grounding state, and if stable, constructing a transition warning model. This not only further analyzes the change trend of the winding state in different time periods by calculating the upper and lower transition trend sub-slopes and the upper and lower transition trend slopes, thereby improving the accuracy of obtaining the transition time node, but also helps operation and maintenance personnel to adjust the operating parameters of the generator excitation system in time according to the temperature change, and then monitor and analyze the generator excitation system in the one-point grounding state, and discover the problem of developing from the one-point grounding state to the two-point grounding state in advance.

Description

Technical Field

[0001] The present invention belongs to the technical field of generator excitation system monitoring, and in particular relates to a generator excitation system state monitoring method and system. Background Art

[0002] In a generator excitation system, the normal operation of the rotor winding is crucial for stable power generation. However, within the generator excitation system, there are multiple locations within the rotor winding where single-point grounding can easily develop into dual-point grounding. For example, between adjacent turns of the upper and lower winding layers within a rotor slot, and between adjacent coils on the same layer at the ends of the rotor winding, insulation aging can gradually degrade, leading to localized short circuits in the windings. This reduces the resistance of the excitation circuit and significantly increases the excitation current, exceeding the rated value. This can cause the excitation system to overheat, seriously threatening the safe and stable operation of the generator.

[0003] In the existing technology, it is difficult to accurately grasp the time node when the winding state changes from one-point grounding to two-point grounding. This makes it impossible for operation and maintenance personnel to discover the trend of increasing temperature difference in advance, and it is difficult to clarify the time range in which the state change may occur. In addition, there is a lack of effective means to evaluate the stability of the winding during the transition from one-point grounding to two-point grounding. As a result, operation and maintenance personnel are unable to timely discover potential unstable factors and cannot issue early warnings when the winding state is about to develop into a two-point grounding state, thereby increasing the risk of generator failure due to winding state change.

[0004] To this end, the present invention provides a method and system for monitoring the state of a generator excitation system. Summary of the Invention

[0005] In order to make up for the deficiencies of the prior art, at least one technical problem raised in the background technology is solved.

[0006] The technical solution adopted by the present invention to solve its technical problem is:

[0007] In a first aspect, a method for monitoring the state of a generator excitation system comprises:

[0008] Monitor and analyze the rotor winding inside the generator over multiple historical monitoring cycles to assess whether the rotor winding is in a single-point grounding state;

[0009] If the rotor winding is in a one-point grounding state, assess the risk of the rotor winding entering a two-point grounding state.

[0010] If a high-risk transition signal is generated, spatial temperature variation analysis is performed on the upper and lower windings inside the rotor winding and the rotor winding ends to screen out winding parts in a two-point grounding state;

[0011] The stability analysis of the transformation trend of the winding parts in the two-point grounding state is carried out. If it is stable, a transformation warning model is constructed.

[0012] As a further solution of the present invention, the screening process of entering a grounding node is as follows:

[0013] Divide the historical monitoring period equally into several historical monitoring nodes;

[0014] Divide the rotor winding equally from top to bottom to obtain several unit coils;

[0015] The unit coil temperature of each unit coil at each historical monitoring node is obtained and summarized and counted according to the rules for dividing the upper and lower winding layers. Multiple unit temperature sequences are output. The temperatures of adjacent unit coils in the unit temperature sequences are combined, and a convolutional neural network algorithm is used to construct multiple convolution coil temperature layers. These layers are all input into the Euclidean calculation formula to output the winding coil temperature difference.

[0016] If the winding coil temperature difference is greater than the winding coil temperature difference threshold, it indicates that the generator excitation system is in a single-point grounding state, and the analyzed historical monitoring node is marked as a single-point grounding monitoring node.

[0017] As a further solution of the present invention, the process of generating a high-risk signal for state transition is as follows:

[0018] The winding coil temperature difference corresponding to a grounded node and the winding coil temperature difference corresponding to a grounded node are sorted according to the historical monitoring node time sequence to construct a coil temperature difference change curve and divide it into multiple coil temperature difference sub-curves;

[0019] The slope of each coil temperature difference sub-curve is obtained by trigonometric method as the temperature difference sub-slope. The temperature difference sub-slopes corresponding to the temperature difference sub-curves of adjacent coils are subtracted to obtain the temperature difference sub-slope difference. The number of positive temperature difference sub-slope differences is counted and the ratio is calculated with the total number of all temperature difference sub-slope differences to output the state transition risk value.

[0020] If the state transition risk value is greater than the state transition risk threshold, a state transition high risk signal is generated.

[0021] As a further solution of the present invention, the spatial temperature variation analysis of the upper and lower windings inside the rotor winding is performed, and the target mid-coil temperature difference value is obtained as follows:

[0022] Extract the two unit coils at the ends of the rotor winding, as well as the unit coils in the upper and lower windings, and mark them as the winding top coil, winding bottom coil, and winding middle coil respectively;

[0023] Taking entering a grounding node as the time screening rule, the unit temperature sequences after entering a node and when entering a node are screened out from multiple unit temperature sequences and marked as target temperature sequences. Multiple target temperature sequences are obtained, and the unit coil temperature of the winding midpoint coil in each target temperature sequence is obtained. The unit coil temperatures in adjacent target temperature sequences where the winding midpoint coil is located are subtracted, and the absolute values ​​are taken to output the target midpoint coil temperature difference value.

[0024] As a further solution of the present invention, it is to analyze whether the upper and lower windings are in a two-point grounding state, and the process is as follows:

[0025] If the target mid-coil temperature difference is greater than the target coil temperature difference threshold, a high-amplitude temperature difference signal is displayed, and the target mid-coil corresponding to the high-amplitude temperature difference signal is marked as a large-amplitude temperature difference mid-coil. The ratio of the number of large-amplitude temperature difference mid-coils to the total number of winding mid-coils is calculated as the large-amplitude temperature difference ratio.

[0026] If the large temperature difference number ratio is greater than the large temperature difference number ratio threshold, it indicates that the upper and lower windings are in a two-point grounding state.

[0027] As a further solution of the present invention, the spatial temperature variation analysis of the rotor winding end is performed, and the process is as follows:

[0028] Using the time of entering a grounding node as the time screening rule, the unit temperature sequences after and at the time of entering a node are screened from multiple unit temperature sequences and marked as target temperature sequences. The unit coil temperature of the winding top coil in each target temperature sequence is obtained, and the unit coil temperatures of the winding top coil in adjacent target temperature sequences are subtracted and the absolute value is taken to output the target top coil temperature difference value.

[0029] The target top coil temperature difference value corresponding to the top coil of the winding in each historical monitoring period is extracted and averaged, and the target top coil temperature difference mean is output. If the target top coil temperature difference mean is greater than the target coil temperature difference threshold, it is displayed that the rotor winding end is in a two-point grounding state.

[0030] As a further solution of the present invention, a stability analysis is performed on the transformation trend corresponding to the winding parts in the two-point grounding state, and an upper and lower transformation trend curve is constructed. The process is as follows:

[0031] The historical monitoring period corresponding to the upper and lower windings being in the two-point grounding state is marked as the upper and lower two-point grounding period;

[0032] The period before the upper and lower points grounding and after entering the one-point grounding node is obtained and marked as the transition trend period;

[0033] The target median coil temperature difference value of each winding median coil in each historical monitoring period is obtained, and the up and down transition trend curve is constructed according to the time sequence within the historical monitoring cycle in which the historical monitoring period is located.

[0034] As a further solution of the present invention, a stability analysis is performed on the upper and lower transition trend curves, and the process of obtaining the transition trend stability value is as follows:

[0035] Taking the local up-down conversion trend curve between adjacent historical monitoring periods as the up-down conversion trend sub-curve, the up-down conversion trend curve is divided to obtain multiple up-down conversion trend sub-curves;

[0036] Extract the coordinates of the two endpoints corresponding to each up-down transition trend sub-curve, input them into the slope calculation formula respectively, and output the up-down transition trend sub-slope;

[0037] Perform average calculation on each up-down transition trend sub-slope and output the up-down transition trend slope;

[0038] The standard deviation of the upper and lower transition trend slopes corresponding to multiple transition trend cycles is calculated, and the transition trend stability value is output.

[0039] As a further solution of the present invention, the construction process of the transformation warning model is as follows:

[0040] If the transition trend stability value is less than or equal to the transition trend stability threshold, it means that the transition trend of the upper and lower layer windings is relatively stable during the transition from the one-point grounding state to the two-point grounding state. The upper and lower transition trend slopes corresponding to multiple transition trend cycles are averaged and calculated to output the transition warning slope.

[0041] Based on the transition warning slope, a transition warning model is constructed. The transition warning model formula is: ,in, Expressed as the transition warning slope, Expressed as a constant.

[0042] In a second aspect, a generator excitation system condition monitoring system includes the following modules:

[0043] Single-point grounding analysis module: monitors and analyzes the rotor winding inside the generator over multiple historical monitoring cycles to assess whether the rotor winding is in a single-point grounding state;

[0044] Transition risk assessment module: If the rotor winding is in a one-point grounding state, the transition risk of the rotor winding entering a two-point grounding state is assessed;

[0045] Two-point grounding analysis module: If a high-risk transition signal is generated, the module performs spatial temperature change analysis on the upper and lower windings inside the rotor winding and the rotor winding ends to screen out winding parts in a two-point grounding state;

[0046] Early warning model construction: Perform stability analysis on the transformation trend of winding parts in the two-point grounding state. If stable, construct a transformation early warning model.

[0047] The beneficial effects of the present invention are as follows:

[0048] The present invention monitors and analyzes the rotor winding inside the generator within multiple historical monitoring cycles to obtain the winding coil temperature difference, thereby reflecting the temperature gradient between multiple adjacent unit coils in the rotor winding, which is conducive to clarifying at which monitoring time node to focus on monitoring and identifying the generator rotor winding. If it is in a one-point grounding state, the rotor winding is evaluated for the risk of transitioning to a two-point grounding state to obtain a state transition risk value. Then, starting from the one-point grounding state, in the subsequent trend of the winding coil temperature difference change, the proportion of the temperature difference showing an upward trend not only helps to clearly observe the trend of the winding coil temperature difference change from a time series, discover the trend of temperature difference increase in advance, and clarify the time range in which the state transition may occur, thereby providing a basis for taking maintenance measures in advance, but also more accurately judges which positions are more likely to transition from a one-point grounding state to a two-point grounding state in the spatial dimension;

[0049] The present invention performs spatial temperature change analysis on the inside of the rotor winding, screens out winding parts in a two-point grounding state, performs stability analysis on the transition trends corresponding to the winding parts in the two-point grounding state, and constructs a transition warning model if it is stable. This not only further analyzes the changing trends of the winding states in different time periods by calculating the upper and lower transition trend sub-slopes and the upper and lower transition trend slopes, thereby improving the accuracy of obtaining the transition time nodes, but also helps operation and maintenance personnel to adjust the operating parameters of the generator excitation system in time according to the temperature changes, and then monitor and analyze the generator excitation system in the one-point grounding state, and discover the problem of the development from the one-point grounding state to the two-point grounding state in advance. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The present invention will be further described below with reference to the accompanying drawings.

[0051] Figure 1 It is a flow chart of the steps of a method for monitoring the state of a generator excitation system of the present invention;

[0052] Figure 2 It is a schematic diagram of a generator excitation system state monitoring system of the present invention. DETAILED DESCRIPTION

[0053] In order to make the technical means, creative features, objectives and effects achieved by the present invention easier to understand, the present invention is further described below in conjunction with specific implementation methods.

[0054] Example 1

[0055] In the rotor windings of the generator excitation system, there are multiple locations where a single-point grounding state can easily develop into a dual-point grounding state, causing a local short circuit in the winding, a reduction in the excitation circuit resistance, and a significant increase in the excitation current, exceeding the rated value, leading to overheating of the excitation system. For example, this includes adjacent turns of the upper and lower windings in the rotor slots, and adjacent coils on the same layer at the ends of the rotor windings. Therefore, it is necessary to monitor and analyze the generator excitation system in a single-point grounding state to promptly detect the problem of a single-point grounding state developing into a dual-point grounding state.

[0056] See also Figure 1 As shown, a method for monitoring the state of a generator excitation system according to an embodiment of the present invention includes the following steps:

[0057] Step 1: Monitor and analyze the rotor winding inside the generator over multiple historical monitoring cycles to assess whether the rotor winding is in a single-point grounding state and identify any single-point grounding nodes.

[0058] In some embodiments, the historical monitoring period is equally divided into a number of historical monitoring nodes;

[0059] Divide the rotor winding equally from top to bottom to obtain several unit coils;

[0060] The temperature sensor is used to obtain the unit coil temperature of each unit coil at each historical monitoring node, and the unit coil temperature is summarized and counted according to the rules of the upper and lower winding divisions to output multiple unit temperature sequences;

[0061] It should be noted that the unit coil temperatures in each unit temperature sequence are the unit coil temperature sets of different unit coils at the same historical monitoring node;

[0062] Combine the adjacent unit coil temperatures in the unit temperature sequence and use the convolutional neural network algorithm to construct multiple convolution coil temperature layers;

[0063] Input multiple convolution coil temperature layers into the Euclidean calculation formula and output the winding coil temperature difference ;

[0064] Specifically, the Euclid calculation formula is: ,in, Expressed as the total number of convolutional coil temperature layers, Expressed as Unit coil temperature, Expressed as Unit coil temperature;

[0065] It is understood that, since a temperature gradient exists between adjacent unit coils on the rotor winding in a generator in a single-point state, the temperature gradient between multiple adjacent unit coils in the rotor winding can be reflected by the winding coil temperature difference. The purpose is to:

[0066] Purpose 1: From a temporal perspective, the temperature gradient of adjacent coils can be used to identify when the generator rotor winding is in a single-point grounding state, and to identify the monitoring time point at which the generator rotor winding should be monitored and identified.

[0067] Objective 2: From a spatial perspective, quantifying the temperature gradient of adjacent coils through temperature differences is helpful in identifying local overheating characteristics in advance when the generator rotor winding is in a single-point grounding state, and determining whether the location is a potential risk point for double-point grounding.

[0068] The winding coil temperature difference is compared with the winding coil temperature difference threshold. The process is as follows:

[0069] If the winding coil temperature difference is less than or equal to the winding coil temperature difference threshold, it means that within the analyzed historical monitoring node, the coil temperature difference at different adjacent positions in the generator rotor winding is small and is in a normal state;

[0070] If the winding coil temperature difference is greater than the winding coil temperature difference threshold, it means that within the analyzed historical monitoring node, the coil temperatures at different adjacent positions in the generator rotor winding are quite different, and the generator is in a single-point grounding state. The analyzed historical monitoring node is marked as a single-point grounding monitoring node.

[0071] All the single-point grounding monitoring nodes are sorted according to the node time sequence in the historical monitoring cycle, and the single-point grounding monitoring node corresponding to the first node time sequence is selected as the entry single-point grounding node;

[0072] Step 2: If the rotor winding is in a one-point grounding state, the risk of transitioning to a two-point grounding state is assessed to obtain a risk assessment result.

[0073] In some embodiments, based on the entry into a grounding node, the winding coil temperature difference corresponding to the grounding node and the winding coil temperature difference corresponding to the grounding node are sorted according to the historical monitoring node time sequence to construct a coil temperature difference change curve;

[0074] Among them, in the coil temperature difference change curve, the X axis represents time and the Y axis represents the winding coil temperature difference;

[0075] The local temperature difference change curve between adjacent coordinate points in the coil temperature difference change curve is used as a coil temperature difference sub-curve to obtain multiple coil temperature difference sub-curves;

[0076] The slope of each coil temperature difference sub-curve is obtained using the trigonometric method as the temperature difference sub-slope. The specific process is as follows:

[0077] A1, extract the coordinates of the two endpoints and the midpoint on the coil temperature difference sub-curve, where the two endpoint coordinates include the starting point coordinate and the end point coordinate;

[0078] A2, according to the coordinates from the starting point midpoint coordinates The order of the end point coordinates is connected in sequence to construct a fitting triangle;

[0079] A3. Use the coordinate point distance formula to obtain the lengths of the three sides of the fitted triangle, and obtain the corresponding height of the fitted triangle at the starting and ending sides.

[0080] Specifically, the starting point coordinates The distance between the midpoint coordinates is: , calculate the starting distance ,in,( , ) represents the starting point coordinates, ( , ) is expressed as the midpoint coordinate;

[0081] midpoint coordinates The distance between the end point coordinates is: , calculate the mid-to-final distance ,in,( , ) is the end point coordinate, ( , ) is expressed as the midpoint coordinate;

[0082] Starting point coordinates The distance between the end point coordinates is: , calculate the starting and ending distances ,in,( , ) represents the starting point coordinates, ( , ) is represented as the end point coordinate;

[0083] The formula for fitting the triangle height is: , calculate the height of the fitted triangle ;

[0084] If the fitted triangle height If it approaches 0, it means that the offset between the midpoint coordinate and the line connecting the two end points of the coil temperature difference curve is small, and the temperature difference slope is ;

[0085] If the fitted triangle height If it does not approach 0, it means that the midpoint coordinate is offset from the line connecting the two end points of the coil temperature difference curve by a large amount, and the temperature difference slope is ;

[0086] Subtract the temperature difference sub-slopes corresponding to the temperature difference sub-curves of adjacent coils to obtain the temperature difference sub-slope difference;

[0087] Count the number of positive temperature difference sub-slope differences, and calculate the ratio with the total number of all temperature difference sub-slope differences to output the state transition risk value;

[0088] It can be understood that the state transition risk value means: it is calculated by counting the number of positive temperature difference sub-slope differences corresponding to the adjacent coil temperature difference sub-curves and calculating the ratio with the total number of all temperature difference sub-slope differences. It reflects the proportion of the temperature difference that shows an upward trend in the subsequent temperature difference change trend of the winding coil starting from the grounding state.

[0089] Its role is to:

[0090] Function 1: In the time dimension, it can clearly observe the changing trend of the winding coil temperature difference from the time series, discover the trend of temperature difference increase in advance, and clarify the time range when the state transition may occur, thereby providing a basis for taking maintenance measures in advance, effectively reducing the probability of generator failure caused by untimely processing of state transitions, and improving the operating reliability and stability of the generator;

[0091] Function 2: In the spatial dimension, it can accurately locate areas with obvious temperature difference rising trends, that is, potential risk points, so as to more accurately determine which locations are more likely to transition from a one-point grounding state to a two-point grounding state in the spatial dimension;

[0092] The state transition risk value is compared with the state transition risk threshold as follows:

[0093] If the state transition risk value is greater than the state transition risk threshold, it indicates that there is a high trend of transitioning from a one-point grounding state to a two-point grounding state, the state transition risk is high, and a state transition high risk signal is generated;

[0094] If the state transition risk value is less than or equal to the state transition risk threshold, it means that there is a low trend of transitioning from the one-point grounding state to the two-point grounding state, the state transition risk is low, and a state transition low risk signal is generated;

[0095] The specific scheme of this embodiment is as follows: within multiple historical monitoring cycles, the rotor winding inside the generator is monitored and analyzed to obtain the winding coil temperature difference, thereby reflecting the temperature gradient between multiple adjacent unit coils in the rotor winding, which is conducive to clarifying at which monitoring time node to focus on monitoring and identifying the generator rotor winding. If it is in a one-point grounding state, the rotor winding is evaluated for the risk of transitioning to a two-point grounding state to obtain a state transition risk value. Then, starting from the one-point grounding state, the proportion of the temperature difference showing an upward trend in the subsequent winding coil temperature difference change trend is determined. This not only helps to clearly observe the changing trend of the winding coil temperature difference from a time series, discover the trend of temperature difference increase in advance, and clarify the time range in which the state transition may occur, thereby providing a basis for taking maintenance measures in advance, but also more accurately judges which positions are more likely to transition from a one-point grounding state to a two-point grounding state in the spatial dimension.

[0096] Example 2

[0097] See also Figure 1 As shown, the generator excitation system state monitoring method according to the embodiment of the present invention further includes the following steps:

[0098] Step 3: If the state transition signals a high-risk signal, analyze the spatial temperature variation inside the rotor winding to screen out winding parts in a two-point grounding state;

[0099] It should be noted that the rotor winding includes upper and lower windings and rotor winding ends;

[0100] The rotor winding ends are the unit coils at the two ends of the rotor winding, and the upper and lower layer windings are the remaining unit coils except the unit coils at the two ends of the rotor winding.

[0101] In some embodiments, two unit coils at the end of the rotor winding are extracted and respectively identified as a winding top coil and a winding bottom coil;

[0102] Extract the unit coils in the upper and lower windings, and represent them as the winding mid-position coils;

[0103] It is further explained that the top coil of the winding, the bottom coil of the winding and the middle coil of the winding are all unit coils;

[0104] For example, take the winding center coil as an example;

[0105] Taking the time of entering a grounding node as the time screening rule, the unit temperature sequence after entering a node (including the time of entering a node) is screened out from multiple unit temperature sequences and marked as the target temperature sequence, thereby obtaining multiple target temperature sequences;

[0106] It should be noted that since the target temperature sequence is selected from multiple unit temperature sequences, and each unit temperature sequence corresponds to the unit coil temperature of the mid-position coil of different windings at the same historical monitoring node, similarly, the target temperature sequence also corresponds to the unit coil temperature set of the mid-position coils of windings at multiple different positions at the same historical monitoring node;

[0107] Therefore, the adjacent target temperature sequence refers to the set of unit coil temperatures of the winding midpoint coils at multiple different positions at adjacent historical monitoring nodes;

[0108] Randomly select a winding mid-position coil for temperature change analysis. The process is as follows:

[0109] Obtain the unit coil temperature of the winding midpoint coil in each target temperature sequence, and subtract the unit coil temperatures in adjacent target temperature sequences where the winding midpoint coil is located, take the absolute value, and output the target midpoint coil temperature difference value;

[0110] Extract the target median coil temperature difference value corresponding to each winding median coil in the same historical monitoring period and compare it with the target coil temperature difference threshold. The process is as follows:

[0111] If the target mid-line temperature difference value is less than or equal to the target mid-line temperature difference threshold, it means that the mid-line temperature difference amplitude of the winding is small during the analyzed historical monitoring period, indicating a low-amplitude temperature difference signal;

[0112] If the target mid-coil temperature difference value is greater than the target coil temperature difference threshold, it means that the mid-coil temperature difference amplitude of the winding is large during the analyzed historical monitoring period, and a high-amplitude temperature difference signal is displayed. The target mid-coil corresponding to the high-amplitude temperature difference signal is marked as a large-amplitude temperature difference mid-coil.

[0113] The ratio of the number of coils with large temperature difference to the total number of coils with large temperature difference is calculated as the ratio of the number of coils with large temperature difference.

[0114] It should be noted that the historical monitoring period is the duration between two adjacent historical monitoring nodes;

[0115] If the large temperature difference ratio is greater than the large temperature difference ratio threshold, it means that within the same historical monitoring period, the temperature changes corresponding to multiple mid-point coils in the upper and lower windings are large, indicating that the upper and lower windings are in a two-point grounding state;

[0116] If the large temperature difference ratio is less than or equal to the large temperature difference ratio threshold, it means that within the same historical monitoring period, the temperature changes corresponding to the middle coils of multiple windings in the upper and lower windings are small, and it does not indicate that the upper and lower windings are in a two-point grounding state;

[0117] Similarly, take the top coil of the winding as an example;

[0118] Using the time of entering a grounding node as the time screening rule, the unit temperature sequence after entering a node (including entering a node) is screened out from multiple unit temperature sequences and marked as the target temperature sequence;

[0119] It should be noted that since the target temperature sequence is selected from multiple unit temperature sequences, and the unit temperature sequence corresponds to the unit coil temperature of the top coil or the bottom coil of the winding at the same historical monitoring node, similarly, the target temperature sequence also corresponds to the unit coil temperature set of the top coil or the bottom coil of the winding at the same historical monitoring node;

[0120] The temperature change analysis of the top coil of the winding is carried out as follows:

[0121] Obtain the unit coil temperature of the winding top coil in each target temperature sequence, and subtract the unit coil temperatures in adjacent target temperature sequences of the winding top coil, take the absolute value, and output the target top coil temperature difference value;

[0122] Extract the target top coil temperature difference value corresponding to the top coil of the winding in each historical monitoring period, perform averaging calculation, and output the target top coil temperature difference mean value;

[0123] The target top coil temperature difference mean is compared with the target coil temperature difference threshold value. The process is as follows:

[0124] If the target top coil temperature difference mean is greater than the target coil temperature difference threshold, it means that the temperature deviation corresponding to the top coil of the winding is large during different historical monitoring periods, indicating that the rotor winding end is in a two-point grounding state;

[0125] If the target top coil temperature difference mean is less than or equal to the target coil temperature difference threshold, it means that the temperature deviation corresponding to the top coil of the winding is small during different historical monitoring periods, and the rotor winding end is not displayed as being in a two-point grounding state;

[0126] Step 4: Perform stability analysis on the transition trend corresponding to the winding parts in the two-point grounding state. If it is stable, build a transition warning model;

[0127] In some embodiments, stability analysis is performed on the transition trends corresponding to the upper and lower windings and the rotor winding ends, respectively, and the process is as follows:

[0128] For example, take the upper and lower windings as an example;

[0129] The historical monitoring period corresponding to the upper and lower windings being in the two-point grounding state is marked as the upper and lower two-point grounding period;

[0130] The period before the upper and lower points grounding and after entering the one-point grounding node is obtained and marked as the transition trend period;

[0131] The transition trend cycle consists of multiple consecutive historical monitoring periods, and the transition trend cycle belongs to the historical monitoring period, and there will be at least one or more transition trend cycles;

[0132] Obtain the target median coil temperature difference value of each winding median coil in each historical monitoring period, and construct an up and down transition trend curve according to the time sequence of the historical monitoring period within the historical monitoring cycle;

[0133] Among them, the X-axis of the up-down transition trend curve is time, and the Y-axis is temperature;

[0134] Taking the local up-down conversion trend curve between adjacent historical monitoring periods as the up-down conversion trend sub-curve, the up-down conversion trend curve is divided to obtain multiple up-down conversion trend sub-curves;

[0135] Extract the coordinates of the two endpoints corresponding to each up-down transition trend sub-curve, input them into the slope calculation formula respectively, and output the up-down transition trend sub-slope;

[0136] Perform average calculation on each up-down transition trend sub-slope and output the up-down transition trend slope;

[0137] Calculate the standard deviation of the upper and lower transition trend slopes corresponding to multiple transition trend cycles, and output the transition trend stability value;

[0138] If the transition trend stability value is greater than the transition trend stability threshold, it means that the transition trend is relatively unstable during the transition from the one-point grounding state to the two-point grounding state in the upper and lower windings;

[0139] If the transition trend stability value is less than or equal to the transition trend stability threshold, it means that the transition trend of the upper and lower layer windings is relatively stable during the transition from the one-point grounding state to the two-point grounding state. The upper and lower transition trend slopes corresponding to multiple transition trend cycles are averaged and calculated to output the transition warning slope.

[0140] Based on the transition warning slope, a transition warning model is constructed. The transition warning model formula is: ,in, Expressed as the transition warning slope, Expressed as a constant;

[0141] It should be noted that the purpose of building a transformation warning model is to:

[0142] Objective 1: To monitor and analyze the generator excitation system in a single-point grounding state. By deeply analyzing the trend stability of winding parts during the transition from a single-point grounding state to a two-point grounding state, problems that develop from a single-point grounding state to a two-point grounding state can be detected in advance.

[0143] Objective 2: From the time dimension, by calculating the up and down transition trend sub-slope and the up and down transition trend slope, we can further analyze the changing trend of the winding state in different time periods and improve the accuracy of obtaining the transition time node;

[0144] Objective 3: From a temperature perspective, combining temperature changes with time changes to quantify the relationship between temperature changes and winding state transitions will help operators adjust the operating parameters of the generator excitation system in a timely manner based on temperature changes.

[0145] The specific solution of this embodiment is: perform spatial temperature change analysis on the inside of the rotor winding, screen out winding parts in a two-point grounding state, perform stability analysis on the transition trend corresponding to the winding parts in the two-point grounding state, and if stable, construct a transition warning model. This not only further analyzes the change trend of the winding state in different time periods by calculating the upper and lower transition trend sub-slopes and the upper and lower transition trend slopes, thereby improving the accuracy of obtaining the transition time node, but also helps operation and maintenance personnel to adjust the operating parameters of the generator excitation system in time according to the temperature change, and then monitor and analyze the generator excitation system in a one-point grounding state, and discover the problem of developing from a one-point grounding state to a two-point grounding state in advance.

[0146] Example 3

[0147] Based on the same inventive concept as the generator excitation system state monitoring method in the above embodiment, Figure 2 As shown, the present application provides a generator excitation system state monitoring system, wherein the system specifically includes:

[0148] Single-point grounding analysis module: monitors and analyzes the rotor winding inside the generator over multiple historical monitoring cycles to assess whether the rotor winding is in a single-point grounding state;

[0149] Transition risk assessment module: If the rotor winding is in a one-point grounding state, the transition risk of the rotor winding entering a two-point grounding state is assessed;

[0150] Two-point grounding analysis module: If a high-risk transition signal is generated, the module performs spatial temperature change analysis on the upper and lower windings inside the rotor winding and the rotor winding ends to screen out winding parts in a two-point grounding state;

[0151] Early warning model construction: Perform stability analysis on the transformation trend of winding parts in the two-point grounding state. If stable, construct a transformation early warning model.

[0152] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the foregoing embodiments. The foregoing embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring the state of a generator excitation system, characterized in that: include: Monitor and analyze the rotor winding inside the generator over multiple historical monitoring cycles to assess whether the rotor winding is in a single-point grounding state; If the rotor winding is in a one-point grounding state, assess the risk of the rotor winding entering a two-point grounding state. The process of generating a high-risk signal for state transition is as follows: The winding coil temperature difference corresponding to a grounded node and the winding coil temperature difference corresponding to a grounded node are sorted according to the historical monitoring node time sequence to construct a coil temperature difference change curve and divide it into multiple coil temperature difference sub-curves; The slope of each coil temperature difference sub-curve is obtained by trigonometric method as the temperature difference sub-slope. The temperature difference sub-slopes corresponding to the temperature difference sub-curves of adjacent coils are subtracted to obtain the temperature difference sub-slope difference. The number of positive temperature difference sub-slope differences is counted and the ratio is calculated with the total number of all temperature difference sub-slope differences to output the state transition risk value. If the state transition risk value is greater than the state transition risk threshold, a state transition high risk signal is generated; If a high-risk transition signal is generated, spatial temperature variation analysis is performed on the upper and lower windings inside the rotor winding and the rotor winding ends to screen out winding parts in a two-point grounding state; The stability analysis of the transformation trend of the winding parts in the two-point grounding state is carried out. If it is stable, a transformation warning model is constructed.

2. A generator excitation system state monitoring method according to claim 1, characterized in that: The screening process for entering a ground node is as follows: Divide the historical monitoring period equally into several historical monitoring nodes; Divide the rotor winding equally from top to bottom to obtain several unit coils; The unit coil temperature of each unit coil at each historical monitoring node is obtained and summarized and counted according to the rules for dividing the upper and lower winding layers. Multiple unit temperature sequences are output. The temperatures of adjacent unit coils in the unit temperature sequences are combined, and a convolutional neural network algorithm is used to construct multiple convolution coil temperature layers. These layers are all input into the Euclidean calculation formula to output the winding coil temperature difference. If the winding coil temperature difference is greater than the winding coil temperature difference threshold, it indicates that the generator excitation system is in a single-point grounding state, and the analyzed historical monitoring node is marked as a single-point grounding monitoring node.

3. The method for monitoring the state of a generator excitation system according to claim 1, characterized in that: The spatial temperature variation analysis of the upper and lower windings inside the rotor winding is performed. The process of obtaining the target mid-coil temperature difference is as follows: Extract the two unit coils at the ends of the rotor winding, as well as the unit coils in the upper and lower windings, and mark them as the winding top coil, winding bottom coil, and winding middle coil respectively; Taking entering a grounding node as the time screening rule, the unit temperature sequences after entering a node and when entering a node are screened out from multiple unit temperature sequences and marked as target temperature sequences. Multiple target temperature sequences are obtained, and the unit coil temperature of the winding midpoint coil in each target temperature sequence is obtained. The unit coil temperatures in adjacent target temperature sequences where the winding midpoint coil is located are subtracted, and the absolute values ​​are taken to output the target midpoint coil temperature difference value.

4. A generator excitation system state monitoring method according to claim 3, characterized in that: Analyze whether the upper and lower windings are in a two-point grounding state. The process is as follows: If the target mid-coil temperature difference is greater than the target coil temperature difference threshold, a high-amplitude temperature difference signal is displayed, and the target mid-coil corresponding to the high-amplitude temperature difference signal is marked as a large-amplitude temperature difference mid-coil. The ratio of the number of large-amplitude temperature difference mid-coils to the total number of winding mid-coils is calculated as the large-amplitude temperature difference ratio. If the large temperature difference number ratio is greater than the large temperature difference number ratio threshold, it indicates that the upper and lower windings are in a two-point grounding state.

5. The method for monitoring the state of a generator excitation system according to claim 1, characterized in that: The spatial temperature variation analysis of the rotor winding end is carried out as follows: Using the time of entering a grounding node as the time screening rule, the unit temperature sequences after and at the time of entering a node are screened from multiple unit temperature sequences and marked as target temperature sequences. The unit coil temperature of the winding top coil in each target temperature sequence is obtained, and the unit coil temperatures of the winding top coil in adjacent target temperature sequences are subtracted and the absolute value is taken to output the target top coil temperature difference value. The target top coil temperature difference value corresponding to the top coil of the winding in each historical monitoring period is extracted and averaged, and the target top coil temperature difference mean is output. If the target top coil temperature difference mean is greater than the target coil temperature difference threshold, it is displayed that the rotor winding end is in a two-point grounding state.

6. The method for monitoring the state of a generator excitation system according to claim 1, characterized in that: Perform stability analysis on the transformation trend corresponding to the winding parts in the two-point grounding state and construct the upper and lower transformation trend curves. The process is as follows: The historical monitoring period corresponding to the upper and lower windings being in the two-point grounding state is marked as the upper and lower two-point grounding period; The period before the upper and lower points grounding and after entering the one-point grounding node is obtained and marked as the transition trend period; The target median coil temperature difference value of each winding median coil in each historical monitoring period is obtained, and the up and down transition trend curve is constructed according to the time sequence within the historical monitoring cycle in which the historical monitoring period is located.

7. The method for monitoring the state of a generator excitation system according to claim 1, characterized in that: Perform stability analysis on the upper and lower transition trend curves, and the process of obtaining the transition trend stability value is as follows: Taking the local up-down conversion trend curve between adjacent historical monitoring periods as the up-down conversion trend sub-curve, the up-down conversion trend curve is divided to obtain multiple up-down conversion trend sub-curves; Extract the coordinates of the two endpoints corresponding to each up-down transition trend sub-curve, input them into the slope calculation formula respectively, and output the up-down transition trend sub-slope; Perform average calculation on each up-down transition trend sub-slope and output the up-down transition trend slope; The standard deviation of the upper and lower transition trend slopes corresponding to multiple transition trend cycles is calculated, and the transition trend stability value is output.

8. The method for monitoring the state of a generator excitation system according to claim 1, characterized in that: The construction process of the transition warning model is as follows: If the transition trend stability value is less than or equal to the transition trend stability threshold, it means that the transition trend of the upper and lower layer windings is relatively stable during the transition from the one-point grounding state to the two-point grounding state. The upper and lower transition trend slopes corresponding to multiple transition trend cycles are averaged and calculated to output the transition warning slope. Based on the transition warning slope, a transition warning model is constructed. The transition warning model formula is: ,in, Expressed as the transition warning slope, Expressed as a constant.

9. A generator excitation system state monitoring system, characterized by: Includes the following modules: Single-point grounding analysis module: monitors and analyzes the rotor winding inside the generator over multiple historical monitoring cycles to assess whether the rotor winding is in a single-point grounding state; Transition risk assessment module: If the rotor winding is in a one-point grounding state, the transition risk of the rotor winding entering a two-point grounding state is assessed; The process of generating a high-risk signal for state transition is as follows: The winding coil temperature difference corresponding to a grounded node and the winding coil temperature difference corresponding to a grounded node are sorted according to the historical monitoring node time sequence to construct a coil temperature difference change curve and divide it into multiple coil temperature difference sub-curves; The slope of each coil temperature difference sub-curve is obtained by trigonometric method as the temperature difference sub-slope. The temperature difference sub-slopes corresponding to the temperature difference sub-curves of adjacent coils are subtracted to obtain the temperature difference sub-slope difference. The number of positive temperature difference sub-slope differences is counted and the ratio is calculated with the total number of all temperature difference sub-slope differences to output the state transition risk value. If the state transition risk value is greater than the state transition risk threshold, a state transition high risk signal is generated; Two-point grounding analysis module: If a high-risk transition signal is generated, the module performs spatial temperature change analysis on the upper and lower windings inside the rotor winding and the rotor winding ends to screen out winding parts in a two-point grounding state; Early warning model construction: Perform stability analysis on the transformation trend of winding parts in the two-point grounding state. If stable, construct a transformation early warning model.

Images