Turbine blade structure fault damage early warning system

Abstract

The invention belongs to the technical field of blade damage early warning, and provides a turbine blade structure fault damage early warning system, which is characterized in that an easily damaged area is firstly obtained, the area which is most easily damaged can be accurately found out from the whole turbine blade area, and the mode of widely detecting the whole blade in the past is changed; the method comprises the following steps of: screening easily-damaged features which really play a key role in blade damage from numerous damage features, analyzing trend change corresponding to the easily-damaged features in a historical detection period, and evaluating whether the development trend of the easily-damaged features in structures such as turbine blades is linear growth or not; if yes, a key basis is provided for predicting the residual life of the turbine blade structure, and the residual life time of the blade is calculated by using a formula based on the detected and analyzed data, so that early warning is timely performed on the fault damage of the turbine blade structure according to the easily damaged area state data detected by the ultrasonic sensor.

Description

Technical Field

[0001] The present invention belongs to the technical field of blade damage warning, and specifically relates to a structural fault damage warning system for steam turbine blades. Background Art

[0002] Steam turbine blades are one of the key components of steam turbines, and their operating status directly affects the performance and safety of steam turbines. During long-term operation, steam turbine blades are subjected to various complex factors such as high temperature, high pressure, high-speed airflow, and mechanical vibration, and are prone to damage, such as cracks and wear. If these damages cannot be detected and processed in time, it may lead to blade fracture, and thus trigger serious safety accidents.

[0003] In the prior art, the detection mainly adopts the method of regular inspection and maintenance, which cannot detect the operating status of the blades and is difficult to detect potential damages in time. In addition, traditional detection methods often conduct broad detection on the entire blade, resulting in scattered detection resources and inability to accurately locate the vulnerable damage areas, leading to low effectiveness of fault warning. Therefore, this application focuses on obtaining the vulnerable damage areas, which is conducive to accurately identifying the areas most prone to damage from the entire steam turbine blade area, changing the previous mode of broad detection of the entire blade. At the same time, from numerous damage characteristics, the vulnerable damage characteristics that truly play a key role in blade damage are screened out, and the trend changes corresponding to the vulnerable damage characteristics within the historical detection period are analyzed to evaluate whether the development trend of the vulnerable damage characteristics in structures such as steam turbine blades is linearly increasing, which provides a key basis for predicting the remaining life of the steam turbine blade structure. Based on the detected and analyzed data, the remaining life time of the blade is calculated using a formula, so as to realize early warning of the structural fault damage of the steam turbine blade in a timely manner according to the status data of the vulnerable damage areas detected by ultrasonic sensors.

[0004] Therefore, the present invention provides a structural fault damage warning system for steam turbine blades. Summary of the Invention

[0005] In order to make up for the deficiencies of the prior art and solve at least one of the technical problems proposed in the background art.

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

[0007] A structural fault damage warning system for steam turbine blades, comprising:

[0008] An area screening module: within the historical detection period, obtain the blade damage information corresponding to multiple damaged steam turbine blades from the steam turbine blade damage reports, and perform quantization processing on the blade damage information to obtain the vulnerable damage areas;

[0009] Among them, the blade damage information includes blade damage sub-areas;

[0010] Lifetime prediction module: Based on the vulnerable line growth curve, obtain the average value of the line growth slope. Obtain the current damage characteristic value corresponding to the vulnerable area through an ultrasonic sensor, and obtain the remaining lifetime of the blade according to the current damage characteristic value and the average value of the line growth slope;

[0011] Among them, the damage characteristics include regional cracks;

[0012] Characteristic analysis module: Analyze the trend changes corresponding to the vulnerable damage characteristics during the historical detection period to obtain vulnerable damage change data. Quantify the vulnerable damage change data, output to obtain a linear similarity value, and compare it with a preset linear similarity threshold to obtain a linear similarity curve. Conduct a trend growth analysis on the linear similarity curve to evaluate whether it is a vulnerable line growth curve;

[0013] Lifetime prediction module: Based on the vulnerable line growth curve, obtain the average value of the line growth slope. Obtain the current damage characteristic value corresponding to the vulnerable area through an ultrasonic sensor, and obtain the remaining lifetime of the blade according to the current damage characteristic value and the average value of the line growth slope.

[0014] As a further technical solution of the present invention: The method for obtaining the vulnerable area is as follows:

[0015] Obtain the blade areas corresponding to multiple damaged steam turbine blades;

[0016] Divide the blade area into several blade sub-areas with equal areas, and perform a coincidence comparison with the blade damage sub-areas to obtain multiple coincident damage sub-areas;

[0017] Arbitrarily extract a coincident damage sub-area;

[0018] Count the number of coincident damage sub-areas, and calculate the ratio with the total number of blade sub-areas to obtain the coincident damage quantity;

[0019] Extract the coincident damage sub-area corresponding to the maximum coincident damage quantity and mark it as the vulnerable area.

[0020] As a further technical solution of the present invention: The method for obtaining the screening analysis data is as follows:

[0021] The screening analysis data includes the damage characteristic degree value;

[0022] Obtain the length corresponding to the regional crack, and calculate the ratio with the perimeter of the vulnerable area to obtain the crack length ratio;

[0023] Obtain the depth corresponding to the regional crack, and calculate the ratio with the perimeter of the vulnerable area to obtain the crack depth ratio;

[0024] Obtain the width corresponding to the regional crack, and calculate the ratio with the perimeter of the vulnerable area to obtain the crack width ratio;

[0025] Sum up the crack length ratio, crack depth ratio and crack width ratio to obtain the damage characteristic degree value.

[0026] As a further technical solution of the present invention: The acquisition method of the screening analysis data is as follows:

[0027] The screening analysis data includes the characteristic quantity ratio;

[0028] Obtain the damage characteristics corresponding to the vulnerable areas through the blade damage information, and summarize and integrate all the damage characteristics corresponding to the vulnerable areas to obtain the damage characteristic set T = {t1, t2, t3,..., t n}, where t n represents the damage characteristic corresponding to the nth vulnerable area, and n represents the number of damage characteristics corresponding to the vulnerable area;

[0029] Count the number of damages to the vulnerable areas caused by the damage characteristics, and calculate the ratio with the corresponding overlapping damage number of the vulnerable areas to obtain the characteristic quantity ratio.

[0030] As a further technical solution of the present invention: Quantify the screening analysis data to obtain a characteristic screening value, and compare it with a preset characteristic screening threshold to obtain vulnerable characteristics. The process is as follows:

[0031] Sum up the characteristic quantity ratio and the damage characteristic degree value to obtain the characteristic screening value;

[0032] If the characteristic screening value is greater than the preset characteristic screening threshold, mark the analyzed damage characteristics as vulnerable characteristics.

[0033] As a further technical solution of the present invention: The acquisition method of the vulnerable change data is as follows:

[0034] The vulnerable change data includes a trend stability value;

[0035] Divide the historical detection period into several historical detection time periods with equal interval durations, obtain the crack length corresponding to the regional crack in each historical detection time period, and substitute it into the two-dimensional coordinate system according to the time sequence of the historical detection time periods to obtain a crack change curve;

[0036] Divide the crack change curve into several crack sub-curves, and obtain the slope values corresponding to the crack sub-curves;

[0037] Substitute the slope values corresponding to the crack sub-curves into the Euclidean distance formula to obtain the trend stability value d.

[0038] As a further technical solution of the present invention: the acquisition method of the vulnerable change data is as follows:

[0039] The vulnerable change data includes the correlation tightness value;

[0040] Obtain all coordinate points on the crack change curve and integrate them according to the time series to obtain a coordinate set;

[0041] Take adjacent coordinate points as a group of coordinate analysis groups, and obtain the correlation coefficient between time and crack length through the statistical test method;

[0042] Substitute the correlation coefficient r between time and crack length into the formula: Q x = 1 - r, and calculate to obtain the correlation tightness value Q x .

[0043] As a further technical solution of the present invention: perform quantization processing on the vulnerable change data, output a linear similarity value, and compare it with a preset linear similarity threshold to obtain a linear similarity curve. The process is as follows:

[0044] Sum the trend stability value d and the correlation tightness value Q x to obtain a linear similarity value;

[0045] If the linear similarity value is greater than the preset linear similarity threshold, mark the crack change curve as a linear similarity curve.

[0046] As a further technical solution of the present invention: perform trend growth analysis on the linear similarity curve to evaluate whether it is a vulnerable line growth curve. The process is as follows:

[0047] Based on the linear similarity curve, obtain the slope values corresponding to all crack sub-curves;

[0048] Mark the crack sub-curve with a positive slope value as a growth sub-curve;

[0049] Mark the crack sub-curve with a negative slope value as a descending sub-curve;

[0050] Count the number of growth sub-curves and the number of descending sub-curves. If the number of growth sub-curves is greater than the number of descending sub-curves, the linear similarity curve is a vulnerable line growth curve.

[0051] As a further technical solution of the present invention: based on the vulnerable line growth curve, obtain the average line growth slope, obtain the current damage characteristic value corresponding to the vulnerable area through an ultrasonic sensor, and obtain the remaining life time of the blade according to the current damage characteristic value. The process is as follows:

[0052] Perform a sum mean calculation on the slope values corresponding to all vulnerable line growth sub-curves within the vulnerable line growth curve to obtain the average line growth slope;

[0053] Obtain the current crack length ratio, current crack depth ratio, and current crack width ratio through an ultrasonic sensor, and perform a summation mean calculation on the current crack length ratio, current crack depth ratio, and current crack width ratio to obtain the current damage characteristic value;

[0054] Substitute the mean value of the line increase slope, the current damage characteristic value, and the damage characteristic degree value into the formula to calculate the remaining life time T of the blade s 。

[0055] The beneficial effects of the present invention are as follows:

[0056] 1. The present invention obtains the blade damage information corresponding to multiple damaged steam turbine blades through the steam turbine blade damage report, and performs quantification processing to obtain the vulnerable damage area, which is beneficial to accurately find the most vulnerable area from the entire steam turbine blade area, changing the previous broad detection mode for the entire blade, concentrating the detection resources on the vulnerable area, improving the effectiveness of fault warning for the steam turbine blade area, making it easier to detect potential damage areas in a timely manner. At the same time, screening and analyzing the damage characteristics corresponding to multiple vulnerable damage areas to obtain the characteristic screening value, so as to screen out the vulnerable damage characteristics that truly play a key role in blade damage from numerous damage characteristics, and can improve the prediction efficiency of the future damage trend prediction work of the blade;

[0057] 2. The present invention analyzes the trend changes corresponding to the vulnerable damage characteristics within the historical detection period to obtain the linear similarity value, thereby reflecting the similarity of the crack change curve in terms of the stability of the change trend and the tightness of the linear correlation with time to the linear growth curve through the linear similarity value, effectively improving the judgment accuracy of whether the crack development is close to linear, reducing misjudgment. At the same time, performing trend growth analysis on the linear similarity curve to evaluate whether the development trend of vulnerable damage characteristics (such as regional cracks) in structures such as steam turbine blades is linear growth. If it is linear growth, it provides a key basis for predicting the remaining life of the steam turbine blade structure;

[0058] 3. The present invention obtains the current damage characteristic value through the vulnerable damage area of the ultrasonic sensor, and based on the current damage characteristic value, combines the damage characteristic degree value and the mean value of the line increase slope using the formula to calculate the remaining life time of the blade, thereby realizing early warning of the structural fault damage of the steam turbine blade in a timely manner according to the state data of the vulnerable damage area. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The present invention will be further described below with reference to the accompanying drawings.

[0060] Figure 1 is a schematic structural diagram of the steam turbine blade structure fault damage warning system of the present invention;

[0061] Figure 2 This is the flowchart of the steps of a method for early warning of structural fault damage of steam turbine blades according to the present invention. Detailed implementation manners

[0062] In order to make the technical means, creative features, achieved purposes and effects of the present invention easy to understand, the present invention will be further described below in conjunction with specific implementation manners.

[0063] Embodiment 1

[0064] As Figure 1-2 shown, the structural fault damage early warning system for steam turbine blades described in the embodiment of the present invention includes:

[0065] Region screening module: During the historical detection period, obtain the blade damage information corresponding to multiple damaged steam turbine blades from the steam turbine blade damage reports, perform quantization processing on the blade damage information to obtain vulnerable regions;

[0066] Among them, the blade damage information includes but is not limited to blade damage sub-regions, damage characteristics corresponding to the vulnerable regions, and blade usage time;

[0067] It should be noted that the damage characteristics causing blade region damage include but are not limited to regional cracks;

[0068] In some embodiments, use a laser three-dimensional scanning technology to scan multiple damaged steam turbine blades with a laser beam to obtain the blade regions;

[0069] Divide the blade regions into several blade sub-regions with equal areas, and perform coincidence comparison with the blade damage sub-regions to obtain multiple coincident damage sub-regions;

[0070] Arbitrarily extract one coincident damage sub-region;

[0071] Count the number of coincident damage sub-regions, and calculate the ratio with the total number of blade sub-regions to obtain the coincident damage quantity;

[0072] It can be understood that the meaning of the coincident damage quantity: the value obtained by calculating the ratio of the number of coincident damage sub-regions obtained after dividing by the laser three-dimensional scanning technology and performing coincidence comparison with the blade damage sub-regions to the total number of blade sub-regions. It reflects the proportion of the damaged sub-regions in the entire blade region, and is an index for quantifying the distribution of the blade damage degree in the overall blade region, which is beneficial to screening out vulnerable regions;

[0073] Compare the sizes of the coincident damage quantities corresponding to all coincident damage sub-regions, extract the coincident damage sub-region corresponding to the maximum coincident damage quantity, and mark it as the vulnerable region;

[0074] Feature screening module: Based on the vulnerable areas, screening and analyzing the damage features corresponding to multiple vulnerable areas to obtain screening and analysis data. Among them, the screening and analysis data includes the feature quantity ratio and the damage feature degree value. Quantify the feature quantity ratio and the damage feature degree value to obtain a feature screening value, and compare it with a preset feature screening threshold. If the feature screening value is greater than the preset feature screening threshold, mark the damage feature as a vulnerable feature;

[0075] Among them, the damage features corresponding to the vulnerable areas include, but are not limited to, regional cracks;

[0076] In some embodiments, obtain the damage features corresponding to the vulnerable areas through the blade damage information, and summarize and integrate all the damage features corresponding to the vulnerable areas to obtain a damage feature set T = {t1, t2, t3,..., t n}, where t n represents the damage feature corresponding to the nth vulnerable area, and n represents the number of damage features corresponding to the vulnerable areas;

[0077] Count the number of damages caused by the damage features to the vulnerable areas, and calculate the ratio with the overlapping damage number corresponding to the vulnerable areas to obtain the feature quantity ratio;

[0078] Arbitrarily select a damage feature from the damage feature set T;

[0079] Exemplarily, if the damage feature is a regional crack;

[0080] Obtain the length corresponding to the regional crack, and calculate the ratio with the perimeter of the vulnerable area to obtain the crack length ratio;

[0081] Obtain the depth corresponding to the regional crack, and calculate the ratio with the perimeter of the vulnerable area to obtain the crack depth ratio;

[0082] Obtain the width corresponding to the regional crack, and calculate the ratio with the perimeter of the vulnerable area to obtain the crack width ratio;

[0083] Sum up the crack length ratio, the crack depth ratio and the crack width ratio to obtain the damage feature degree value;

[0084] It should be noted that the meaning represented by the damage feature degree value is: taking the regional crack as an example, it reflects the comprehensive situation of the damage feature of the regional crack in terms of geometric dimensions such as length, depth and width relative to the perimeter of the vulnerable area, and reflects the severity of the damage feature itself to the vulnerable area of the blade;

[0085] Sum up the feature quantity ratio and the damage feature degree value to obtain the feature screening value;

[0086] It can be understood that the meaning of the feature screening value is obtained by summing the feature quantity ratio and the damage feature degree value. It combines the geometric severity of the damage feature and the proportion of the number of damages caused in two dimensions of information, and is used to screen out vulnerable damage features from multiple damage features. If this value is larger, it indicates that the damage feature appears more frequently in the vulnerable damage area and the damage degree caused by the damage feature is larger. On the contrary, if this value is smaller, it indicates that the damage feature appears less frequently in the vulnerable damage area and the damage degree caused by the damage feature is smaller. Thus, the vulnerable damage features that truly play a key role in blade damage are screened out from numerous damage features, which can provide data support for the prediction work of future damage trends of the blade;

[0087] Compare the feature screening value with the preset feature screening threshold, and the process is as follows:

[0088] If the feature screening value is greater than the preset feature screening threshold, it indicates that the damage feature appears more frequently in the vulnerable damage area and the damage degree caused by the damage feature is larger. Mark the damage feature as a vulnerable damage feature;

[0089] If the feature screening value is less than or equal to the preset feature screening threshold, it indicates that the damage feature appears less frequently in the vulnerable damage area and the damage degree caused by the damage feature is smaller. Mark the damage feature as a non-vulnerable damage feature;

[0090] The specific implementation scheme of the embodiment of the present invention is: obtain the blade damage information corresponding to multiple damaged steam turbine blades through the steam turbine blade damage report, and perform quantization processing to obtain the vulnerable damage area, which is beneficial to accurately find the area most likely to be damaged from the entire steam turbine blade area, change the previous broad detection mode for the entire blade, concentrate the detection resources on the vulnerable loss area, improve the effectiveness of fault warning for the steam turbine blade area, and more easily detect potential damage areas in a timely manner. At the same time, screen and analyze the damage features corresponding to multiple vulnerable damage areas to obtain the feature screening value, so as to screen out the vulnerable damage features that truly play a key role in blade damage from numerous damage features, which can improve the prediction efficiency for the prediction work of future damage trends of the blade;

[0091] Embodiment 2

[0092] As Figure 1-2 shown, on the basis of Embodiment 1, the steam turbine blade structure fault damage warning system described in the embodiment of the present invention includes:

[0093] Feature analysis module: Based on the vulnerable features, analyze the trend changes corresponding to the vulnerable features within the historical detection period to obtain vulnerable change data. The vulnerable change data includes a trend stability value and a correlation tightness value. Quantify the trend stability value and the correlation tightness value, output a linear similarity value, compare it with a preset linear similarity threshold to obtain a linear similarity curve, and perform a trend growth analysis on the linear similarity curve to evaluate whether it is a vulnerable line growth curve;

[0094] Exemplarily, if the vulnerable feature is a regional crack;

[0095] Divide the historical detection period into several historical detection time periods with equal interval durations, and obtain the crack length corresponding to the regional crack within each historical detection time period;

[0096] Establish a two-dimensional coordinate system, with the X-axis representing time and the Y-axis representing the crack length. Substitute the crack length corresponding to the regional crack into the two-dimensional coordinate system according to the time series of the historical detection time periods to obtain a crack change curve;

[0097] Divide the crack change curve into several crack sub-curves, and obtain the corresponding slope values (the slope values are not zero) of the crack sub-curves;

[0098] It should be noted that the division method of the crack change curve is to use the line segment between adjacent crack length values on the crack change curve as the crack sub-curve, and the total number of crack sub-curves is odd;

[0099] Analyze the linear similarity of the crack change curve through the Euclidean distance method. The process is as follows:

[0100] A1. Take the slope values corresponding to adjacent crack sub-curves as a group of slope analysis groups to obtain multiple groups of slope analysis groups;

[0101] A2. Arbitrarily select a group of slope analysis groups, subtract the slope values corresponding to adjacent crack sub-curves within the slope analysis group to obtain a slope difference;

[0102] A3. Through the Euclidean calculation formula: Calculate to obtain the trend stability value d, k v 、k v+1 represent the slope values corresponding to adjacent crack sub-curves within the slope analysis group, and m represents the total number of slope values in the slope analysis group;

[0103] Obtain all the coordinate points on the crack change curve and integrate them into a coordinate set according to the time series;

[0104] Take adjacent coordinate points as a group of coordinate analysis groups, and obtain the correlation coefficient between time and crack length through the statistical test method. The process is as follows:

[0105] S1. Extract the X-axis coordinates of all coordinate points in the coordinate set, calculate their sum and average value, and obtain the average value of the X-axis coordinates.

[0106] S2. Extract the Y-axis coordinates of all coordinate points in the coordinate set, calculate their sum and average value, and obtain the average value of the Y-axis coordinates.

[0107] S3. Through the correlation coefficient calculation formula: Calculate the correlation coefficient r between time and crack length, where j represents the total number of coordinate points in the coordinate set, p represents the p-th coordinate point in the coordinate set, represents the average value of the X-axis coordinates, represents the average value of the Y-axis coordinates, X p represents the X-axis coordinate corresponding to the p-th coordinate point in the coordinate set, Y p represents the Y-axis coordinate corresponding to the p-th coordinate point in the coordinate set.

[0108] Substitute the correlation coefficient r between time and crack length into the formula: Q x = 1 - r, and calculate the associated tightness value Q x ;

[0109] Sum the trend stability value d and the associated tightness value Q x to calculate the linear similarity value;

[0110] It can be explained that the meaning represented by the linear similarity value is: obtained by summing the trend stability value and the associated tightness value, comprehensively reflecting the similarity of the crack change curve to the linear growth curve in terms of the stability of the change trend and the linear association tightness with time;

[0111] Compare the linear similarity value with the preset linear similarity threshold, and the process is as follows:

[0112] If the linear similarity value is greater than the preset linear similarity threshold, mark the crack change curve as a linearly similar curve;

[0113] If the linear similarity value is less than or equal to the preset linear similarity threshold, mark the crack change curve as a non-linearly similar curve;

[0114] Based on the linearly similar curve, obtain the slope values corresponding to all crack sub-curves;

[0115] Mark the crack sub-curves with positive slope values as growth sub-curves;

[0116] Mark the crack sub-curves with negative slope values as decline sub-curves;

[0117] Count the number of increasing sub - curves and decreasing sub - curves, and compare them. The process is as follows:

[0118] If the number of increasing sub - curves is greater than the number of decreasing sub - curves, the linear similar curve is a vulnerable line - increasing curve;

[0119] If the number of increasing sub - curves is less than the number of decreasing sub - curves, the linear similar curve is a vulnerable non - line - increasing curve;

[0120] The specific implementation scheme of the embodiment of the present invention is as follows: Analyze the trend changes corresponding to the vulnerable features within the historical detection period to obtain a linear similarity value, so as to reflect the similarity between the crack change curve and the linear growth curve in terms of the stability of the change trend and the tightness of the linear correlation with time through the linear similarity value, effectively improving the judgment accuracy of whether the crack development is close to linear and reducing misjudgment. At the same time, conduct a trend growth analysis on the linear similar curve to evaluate whether the development trend of vulnerable features (such as regional cracks) in structures such as steam turbine blades is linear growth. If it is linear growth, it provides a key basis for predicting the remaining life of the steam turbine blade structure.

[0121] Embodiment 3

[0122] As Figure 1-2 shown, based on Embodiment 1 and Embodiment 2, the steam turbine blade structure fault damage warning system described in the embodiment of the present invention includes:

[0123] Life prediction module: Based on the vulnerable line - increasing curve, obtain the average value of the line - increasing slope, obtain the current damage feature value corresponding to the vulnerable area through an ultrasonic sensor, and obtain the remaining life time of the blade according to the current damage feature value and the average value of the line - increasing slope;

[0124] It should be noted that the ultrasonic sensor emits ultrasonic waves to the blade and receives the reflected waves when the steam turbine blade stops working, and detects the size information of the cracks on the blade surface by analyzing the characteristics such as the amplitude and time of the echo. Among them, the size information of the cracks includes the current crack length, the current crack depth, and the current crack width;

[0125] In some embodiments, sum up and calculate the average value of the slope values corresponding to all vulnerable line - increasing sub - curves within the vulnerable line - increasing curve to obtain the average value of the line - increasing slope;

[0126] Obtain the current crack length ratio, the current crack depth ratio, and the current crack width ratio through the ultrasonic sensor, and perform a sum and average calculation on the crack length ratio, the current crack depth ratio, and the current crack width ratio to obtain the current damage feature value;

[0127] Through the formula: Calculate to obtain the remaining life time T of the blade s, where L y is expressed as the damage feature degree value, L d is expressed as the current damage feature value, K j is expressed as the average value of the line increase slope;

[0128] The remaining life time T of the blade s means: a value that reflects the length of time that the steam turbine blade can continue to be used under the consideration of the current damage condition and the development trend of the damage feature. It provides a quantitative index for evaluating the remaining service life of the steam turbine blade, and helps the operation and maintenance personnel predict the time for maintaining the steam turbine blade or the replacement time of the blade;

[0129] The specific implementation manner of the embodiment of the present invention is: obtaining the current damage feature value corresponding to the vulnerable area through an ultrasonic sensor, and according to the current damage feature value, using the formula to combine the damage feature degree value and the average value of the line increase slope to calculate the remaining life time of the blade, so as to realize the early warning of the structural fault damage of the steam turbine blade according to the state data of the vulnerable area.

[0130] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification only illustrate the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. Steam turbine blade structure failure damage early warning system, characterized by: include: Area screening module: within the historical detection period, obtain blade damage information corresponding to multiple damaged turbine blades from the turbine blade damage report, quantify the blade damage information, and obtain the vulnerable area; Wherein, the blade damage information includes blade damage sub-areas; Feature screening module: Screen and analyze the damage features corresponding to multiple vulnerable areas to obtain screening analysis data, quantify the screening analysis data to obtain feature screening values, and compare them with the preset feature screening thresholds to obtain vulnerable features; Among them, damage characteristics include regional cracks; Feature analysis module: Analyze the trend changes corresponding to the vulnerable features in the historical detection cycle to obtain vulnerable change data, quantify the vulnerable change data, output the linear similarity value, and compare it with the preset linear similarity threshold to obtain the linear similarity curve, perform trend growth analysis on the linear similarity curve, and evaluate whether it is a vulnerable line growth curve; Life prediction module: Based on the vulnerable line increase curve, the mean line increase slope is obtained, the current damage characteristic value corresponding to the vulnerable area is obtained through the ultrasonic sensor, and the remaining life of the blade is obtained based on the current damage characteristic value and the mean line increase slope.

2. The turbine blade structure failure damage early warning system according to claim 1 is characterized in that: The vulnerable area is obtained as follows: Obtain blade regions corresponding to a plurality of damaged steam turbine blades; The leaf region is divided into a number of leaf sub-regions with equal areas, and the sub-regions are overlapped and compared with the leaf damaged sub-regions to obtain a number of overlapped damaged sub-regions; Arbitrarily select a coincident damage sub-region; The number of overlapping damage sub-areas is counted, and the ratio is calculated with the total number of leaf sub-areas to obtain the number of overlapping damages; The overlapping damage sub-region corresponding to the maximum number of overlapping damages is extracted and marked as the vulnerable area.

3. The turbine blade structure failure damage early warning system according to claim 2 is characterized in that: The method of obtaining screening analysis data is as follows: The screening analysis data include damage characteristic degree values; Obtain the length corresponding to the regional crack and calculate the ratio with the perimeter of the vulnerable area to obtain the crack length ratio; The corresponding depth of the regional crack is obtained, and the ratio is calculated with the perimeter of the vulnerable area to obtain the crack depth ratio; Obtain the width corresponding to the regional crack and calculate the ratio with the perimeter of the vulnerable area to obtain the crack width ratio; The crack length ratio, crack depth ratio and crack width ratio are summed up and calculated to obtain the damage characteristic degree value.

4. The turbine blade structure failure damage early warning system according to claim 3 is characterized in that: The method of obtaining screening analysis data is as follows: Screening analysis data included feature quantity ratios; The damage features corresponding to the vulnerable areas are obtained through the blade damage information, and the damage features corresponding to all vulnerable areas are summarized and integrated to obtain the damage feature set T = t1, t2, t3, ..., t n , where t n It is represented by the damage feature corresponding to the nth vulnerable area, and n is represented by the number of damage features corresponding to the vulnerable area; The number of damages in the vulnerable area caused by the damage characteristics is counted, and the ratio is calculated with the number of overlapping damages corresponding to the vulnerable area to obtain the characteristic quantity ratio.

5. The steam turbine blade structure failure damage early warning system according to claim 4 is characterized in that: The screening analysis data is quantified to obtain the feature screening value, which is then compared with the preset feature screening threshold to obtain the vulnerable feature. The process is as follows: The feature quantity ratio and the damage feature degree value are summed up to obtain the feature screening value; If the feature screening value is greater than the preset feature screening threshold, the analyzed damage feature is marked as a vulnerable feature.

6. The turbine blade structure failure damage early warning system according to claim 5 is characterized in that: The method for obtaining the damage-prone change data is as follows: Vulnerability change data include trend stability values; The historical detection cycle is divided into several historical detection periods with equal intervals, the crack length corresponding to the regional crack in each historical detection period is obtained, and the crack change curve is obtained by substituting the time series of the historical detection period into the two-dimensional coordinate system; Divide the crack variation curve into a number of crack sub-curves, and obtain the slope values ​​corresponding to the crack sub-curves; Substitute the slope value corresponding to the crack sub-curve into the Euclidean distance formula to obtain the trend stability value d.

7. The turbine blade structure failure damage early warning system according to claim 6 is characterized in that: The method for obtaining the damage-prone change data is as follows: The vulnerable change data includes closely related values; Obtain all coordinate points on the crack change curve and integrate them according to the time series to obtain a coordinate set; The adjacent coordinate points are taken as a set of coordinate analysis groups, and the correlation coefficient between time and crack length is obtained by statistical test method; Substituting the correlation coefficient r between time and crack length into the formula: Q x =1-r, calculate the close correlation value Q x .

8. The turbine blade structure failure damage early warning system according to claim 7 is characterized in that: The damage-prone change data is quantified and output to obtain a linear similarity value, which is then compared with a preset linear similarity threshold to obtain a linear similarity curve. The process is as follows: The trend stability value d and the close correlation value Q x Perform sum calculation to obtain the linear similarity value; If the linear similarity value is greater than a preset linear similarity threshold, the crack change curve is marked as a linear similarity curve.

9. The steam turbine blade structure failure damage early warning system according to claim 8, characterized in that: Perform trend growth analysis on the linear similarity curve to assess whether it is a vulnerable line growth curve. The process is as follows: Based on the linear similarity curve, the slope values ​​corresponding to all crack sub-curves are obtained; The crack sub-curve with positive slope value is marked as the growth sub-curve; The crack sub-curve with negative slope value is labeled as descending sub-curve; The number of increasing sub-curves and the number of decreasing sub-curves are counted. If the number of increasing sub-curves is greater than the number of decreasing sub-curves, the linear similarity curve is a vulnerable linear increasing curve.

10. The steam turbine blade structure failure damage early warning system according to claim 9, characterized in that: Based on the vulnerable line increase curve, the mean value of the line increase slope is obtained, the current damage characteristic value corresponding to the vulnerable area is obtained through the ultrasonic sensor, and the remaining life of the blade is obtained according to the current damage characteristic value. The process is as follows: The slope values ​​corresponding to all vulnerable line increase sub-curves in the vulnerable line increase curve are summed and averaged to obtain the line increase slope average; The current crack length ratio, the current crack depth ratio and the current crack width ratio are obtained by using an ultrasonic sensor, and the current crack length ratio, the current crack depth ratio and the current crack width ratio are summed and averaged to obtain the current damage characteristic value; Substitute the mean value of the line increase slope, the current damage characteristic value and the damage characteristic degree value into the formula to calculate the remaining life time T of the blade. s .

Images