Mechanical part damage detection method and system based on artificial intelligence

Abstract

The invention provides a mechanical part damage detection method and system based on artificial intelligence. According to the technical scheme, the method comprises the steps that the damage detection period of a target mechanical part is set; analyzing to obtain a first external damage coefficient, a second external damage coefficient and a third external damage coefficient of the target mechanical part; obtaining a part external damage coefficient of the target mechanical part according to the first external damage coefficient, the second external damage coefficient and the third external damage coefficient; analyzing to obtain a part internal damage coefficient of the target mechanical part; a part damage coefficient of the target mechanical part is obtained according to the part external damage coefficient and the part internal damage coefficient, whether the target mechanical part is abnormal or not is judged according to the part damage coefficient and a preset part damage coefficient threshold value, and the damage abnormity degree of the target mechanical part is obtained; and obtaining the part damage degree of the target mechanical part according to the damage abnormity degree. According to the invention, the comprehensiveness and accuracy of the damage detection result of the target mechanical part are improved.

Description

Technical Field

[0001] The present application relates to the technical field of mechanical parts, and specifically to a mechanical parts damage detection method and system based on artificial intelligence. Background Art

[0002] Mechanical parts are the basic components of various mechanical equipment. Their quality and performance directly affect the operational reliability and safety of the entire mechanical equipment. In modern industrial production, mechanical parts are widely used in many fields such as automobile manufacturing, aerospace, energy, and chemical industry. For example, in automobile engines, the good condition of parts such as pistons, crankshafts, and valves is the key to ensuring the normal operation of the engine; in the aerospace field, the reliability of aircraft engine blades, landing gear and other parts is related to flight safety; therefore, damage detection and damage judgment of mechanical parts are of great significance to ensuring and maintaining the safety and reliability of mechanical parts and even mechanical equipment.

[0003] The mechanical parts damage detection methods in the related art often consider the factors causing the damage to the mechanical parts in a relatively single way, resulting in the lack of comprehensiveness and accuracy in the damage detection results of the mechanical parts. There is room for improvement. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide a mechanical parts damage detection method and system based on artificial intelligence, so as to improve the problem that the mechanical parts damage detection method in the related art often considers the factors causing the damage to the mechanical parts in a relatively single way, resulting in the lack of comprehensiveness and accuracy in the damage detection results of the mechanical parts.

[0005] In a first aspect of the present application, a mechanical parts damage detection method based on artificial intelligence is provided, comprising the following steps:

[0006] Acquire part parameter data, part external inspection data and part internal inspection data of the target mechanical part;

[0007] Setting a damage detection cycle of a target mechanical part according to the part parameter data;

[0008] The part external damage data includes part wear data, part crack data and part corrosion data of the target mechanical part; the part internal detection data includes internal signal data of the target mechanical part;

[0009] Analyze the part wear data of the target mechanical part to obtain a first external damage coefficient corresponding to the target mechanical part; analyze the part corrosion data of the target mechanical part to obtain a second external damage coefficient corresponding to the target mechanical part; analyze the part crack data of the target mechanical part to obtain a third external damage coefficient corresponding to the target mechanical part;

[0010] Obtaining a part external damage coefficient corresponding to a target mechanical part according to the first external damage coefficient, the second external damage coefficient, and the third external damage coefficient;

[0011] Generate an internal signal waveform curve of the target mechanical part according to the internal signal data of the target mechanical part, and obtain a signal curve abnormality coefficient corresponding to the target mechanical part according to the analysis of the internal signal waveform curve and a preset part standard waveform curve;

[0012] Obtaining the internal damage coefficient of the target mechanical part according to the abnormal coefficient of the signal curve;

[0013] Obtaining a part damage coefficient of a target mechanical part according to the part external damage coefficient and the part internal damage coefficient, and comparing and analyzing the part damage coefficient with a preset part damage coefficient threshold to determine whether the target mechanical part is abnormal;

[0014] When the target mechanical part is abnormal, the damage abnormality of the target mechanical part is obtained, and the part damage degree of the target mechanical part is obtained according to the damage abnormality.

[0015] Preferably, the damage detection cycle of the target mechanical part is set according to the part parameter data, specifically:

[0016] The part parameter data includes the use time of the target mechanical part and the part historical detection cycle;

[0017] The part usage time of the target mechanical part is obtained by taking and marking the part usage time YS; the part history detection period of the target mechanical part is obtained by taking and marking the part history detection period TL;

[0018] The periodic influence coefficient YX of the target mechanical part is calculated by a first calculation function YX=α×YS, where α is a preset proportional factor;

[0019] The damage detection period TJ of the target mechanical part is calculated by the second calculation function TJ=TL-TL×YX.

[0020] Preferably, the part external damage data includes part wear data, part crack data and part corrosion data of the target mechanical part, specifically:

[0021] The part wear data includes the initial volume and initial weight of the target mechanical part at the initial moment of the damage detection cycle, and the final volume and final weight of the target mechanical part at the end moment of the damage detection cycle;

[0022] The part corrosion data includes the initial color of the part surface of the target mechanical part at the initial moment of the damage detection cycle, and the detected color of the part surface of the target mechanical part at the end moment of the damage detection cycle;

[0023] The part crack data includes the total number of cracks on the part surface of the target mechanical part, as well as the crack length, crack width and crack depth of the part surface cracks.

[0024] Preferably, the part wear data of the target mechanical part is analyzed to obtain a first external damage coefficient corresponding to the target mechanical part, specifically:

[0025] The initial volume of the target mechanical part at the initial moment of the damage detection cycle is obtained by taking and marking the initial volume of the part VC; the initial weight of the target mechanical part at the initial moment of the damage detection cycle is obtained by taking and marking the initial weight of the part MC;

[0026] The final volume of the target mechanical part at the end of the damage detection cycle is obtained by taking and marking the final volume VZ of the part; the final weight of the target mechanical part at the end of the damage detection cycle is obtained by taking and marking the final weight MZ of the part;

[0027] Through the third calculation function Calculate the first external damage coefficient OWX corresponding to the target mechanical part, where a 1 、a 2 is the preset scaling factor.

[0028] Preferably, the part corrosion data of the target mechanical part is analyzed to obtain a second external damage coefficient corresponding to the target mechanical part, specifically:

[0029] Compare the detected color of the target mechanical part's surface at the end of the damage detection cycle with the initial color of the part's surface;

[0030] Marking a part surface region where the detected color of the part surface is different from the initial color of the part surface as an abnormal surface region, wherein the target mechanical part includes at least one abnormal surface region;

[0031] Obtain an abnormal surface area corresponding to the abnormal surface region, and take values ​​and mark the abnormal surface area to obtain an abnormal surface area SY; obtain a total part surface area of ​​a target mechanical part, and take values ​​and mark the total part surface area to obtain a total part surface area SZ;

[0032] Through the fourth calculation function The surface corrosion coefficient FX of the target mechanical part is calculated, where b is a preset proportional factor, j is the number of the abnormal surface area, and n is the total number of the abnormal surface areas;

[0033] Obtain the surface corrosion degree corresponding to the abnormal surface area according to the surface detection color of the part in the abnormal surface area and the initial color of the part surface, and obtain the surface corrosion degree FD by taking and marking the surface corrosion degree;

[0034] Through the fifth calculation function The second external damage coefficient SWX corresponding to the target mechanical part is calculated, where c 1 、c 2 is a preset scaling factor, j is the number of the abnormal surface area, and n is the total number of abnormal surface areas.

[0035] Preferably, the part crack data of the target mechanical part is analyzed to obtain a third external damage coefficient corresponding to the target mechanical part, specifically:

[0036] The crack length of the crack on the surface of the part is measured and marked to obtain the crack length LC; the crack width of the crack on the surface of the part is measured and marked to obtain the crack width LK; the crack depth of the crack on the surface of the part is measured and marked to obtain the crack depth LS;

[0037] By the sixth calculation function LCX=d 1 ×LC+d 2 ×LK+d 3 ×LS calculates the crack parameter coefficient LCX corresponding to the surface crack of the part, where d 1 ,d 2 ,d 3 is the preset scaling factor;

[0038] The total number of cracks on the surface of the target mechanical part is taken and marked to obtain the total number of cracks m;

[0039] Through the seventh calculation function The third external damage coefficient SWX corresponding to the target mechanical part is calculated, where i is the number of the crack on the part surface.

[0040] Preferably, the part external damage coefficient corresponding to the target mechanical part is obtained according to the first external damage coefficient, the second external damage coefficient and the third external damage coefficient, specifically:

[0041] By the eighth calculation function WSX=e 1 ×OWX+e 2 ×TWX+e 3 ×SWX calculates the external damage coefficient WSX of the target mechanical part, where e1 、e 2 、e 3 is the preset scaling factor.

[0042] Preferably, an internal signal waveform curve of the target mechanical part is generated according to the internal signal data of the target mechanical part, and a signal curve abnormality coefficient corresponding to the target mechanical part is obtained according to the analysis of the internal signal waveform curve and a preset part standard waveform curve, specifically:

[0043] Compare the internal signal waveform curve of the target mechanical part with the preset part standard waveform curve;

[0044] Mark the waveform curve of the non-overlapping part of the internal signal waveform curve and the part standard waveform curve as an abnormal waveform curve; mark the part standard waveform curve corresponding to the abnormal waveform curve as a reference waveform curve;

[0045] Obtain the peak value difference, the number of inflection points and the mean value difference of the slope between the abnormal waveform curve and the reference waveform curve;

[0046] The peak value difference between the abnormal waveform curve and the reference waveform curve is obtained and marked to obtain the peak value difference FC; the number difference of curve inflection points between the abnormal waveform curve and the reference waveform curve is obtained and marked to obtain the number difference of curve inflection points GC; the mean value difference of curve slope between the abnormal waveform curve and the reference waveform curve is obtained and marked to obtain the mean value difference of curve slope XC;

[0047] Through the ninth calculation function QYX=f 1 ×FC+f 2 ×GC+f 3 ×XC calculates the signal curve abnormality coefficient QYX corresponding to the target mechanical part, where f 1 、f 2 、f 3 is the preset scaling factor;

[0048] The internal damage coefficient of the target mechanical part is obtained according to the abnormal coefficient of the signal curve, specifically:

[0049] The internal damage coefficient NSX of the target mechanical part is calculated by the tenth calculation function NSX=δ×QYX, wherein δ is a preset proportional factor.

[0050] Preferably, the part damage coefficient of the target mechanical part is obtained according to the part external damage coefficient and the part internal damage coefficient, and the part damage coefficient is compared and analyzed with a preset part damage coefficient threshold to determine whether the target mechanical part is abnormal, and when the target mechanical part is abnormal, the damage abnormality degree of the target mechanical part is obtained, and the part damage degree of the target mechanical part is obtained according to the damage abnormality degree, specifically:

[0051] By the eleventh calculation function LSX=g 1 ×WSX+g 2 ×NSX calculates the target mechanical parts damage coefficient LSX, where g 1 , g 2 Preset scaling factors;

[0052] If the part damage coefficient LSX of the target mechanical part is less than the preset part damage coefficient threshold, it is judged that the target mechanical part is normal;

[0053] If the part damage coefficient LSX of the target mechanical part is greater than or equal to the preset part damage coefficient threshold, the target mechanical part is judged to be abnormal, and the preset part damage coefficient threshold is taken and marked to obtain the part damage coefficient threshold SXY;

[0054] The damage abnormality degree SYD of the target mechanical part is calculated by the twelfth calculation function SYD=LSX-SXY;

[0055] The part damage degree of the target mechanical part is positively correlated with the damage abnormality degree SYD of the target mechanical part.

[0056] In a second aspect of the present application, a mechanical parts damage detection system based on artificial intelligence is provided, the system is applied to a mechanical parts damage detection method based on artificial intelligence, the system comprising:

[0057] Data acquisition module: acquires part parameter data, part external detection data and part internal detection data of target mechanical parts;

[0058] Setting a damage detection cycle of a target mechanical part according to the part parameter data;

[0059] The part external damage data includes part wear data, part crack data and part corrosion data of the target mechanical part; the part internal detection data includes internal signal data of the target mechanical part;

[0060] The first analysis module: analyzes the part wear data of the target mechanical part to obtain a first external damage coefficient corresponding to the target mechanical part; analyzes the part corrosion data of the target mechanical part to obtain a second external damage coefficient corresponding to the target mechanical part; analyzes the part crack data of the target mechanical part to obtain a third external damage coefficient corresponding to the target mechanical part;

[0061] A second analysis module: obtaining a part external damage coefficient corresponding to a target mechanical part according to the first external damage coefficient, the second external damage coefficient and the third external damage coefficient;

[0062] A third analysis module: generating an internal signal waveform curve of the target mechanical part according to the internal signal data of the target mechanical part, and obtaining a signal curve abnormality coefficient corresponding to the target mechanical part according to the analysis of the internal signal waveform curve and a preset part standard waveform curve;

[0063] Fourth analysis module: obtaining the internal damage coefficient of the target mechanical part according to the abnormal coefficient of the signal curve;

[0064] Damage detection module: obtains the part damage coefficient of the target mechanical part according to the external damage coefficient of the part and the internal damage coefficient of the part, and compares and analyzes the part damage coefficient with a preset part damage coefficient threshold to determine whether the target mechanical part is abnormal;

[0065] When the target mechanical part is abnormal, the damage abnormality of the target mechanical part is obtained, and the part damage degree of the target mechanical part is obtained according to the damage abnormality.

[0066] In summary, the beneficial effects of the present application are as follows: the present application obtains a first external damage coefficient corresponding to the target mechanical part by analyzing the part wear data of the target mechanical part; obtains a second external damage coefficient corresponding to the target mechanical part by analyzing the part corrosion data of the target mechanical part; obtains a third external damage coefficient corresponding to the target mechanical part by analyzing the part crack data of the target mechanical part; and obtains the part external damage coefficient corresponding to the target mechanical part based on the first external damage coefficient, the second external damage coefficient and the third external damage coefficient; in addition, generates an internal signal waveform curve of the target mechanical part based on the internal signal data of the target mechanical part, and generates an internal signal waveform curve based on the internal signal waveform curve The signal curve abnormality coefficient corresponding to the target mechanical part is obtained by analyzing the signal curve abnormality coefficient with the preset part standard waveform curve; and the part internal damage coefficient of the target mechanical part is obtained based on the signal curve abnormality coefficient; finally, the part damage coefficient of the target mechanical part is obtained based on the part external damage coefficient and the part internal damage coefficient, and the part damage coefficient is compared with the preset part damage coefficient threshold to determine whether the target mechanical part is abnormal, and when the target mechanical part is abnormal, the damage abnormality degree of the target mechanical part is obtained, and the part damage degree of the target mechanical part is obtained based on the damage abnormality degree; the present application comprehensively considers the factors causing damage to the target mechanical part, and improves the comprehensiveness and accuracy of the damage detection results of the target mechanical part. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions of the embodiments of the present application, some of the drawings in the embodiments of the present application will be briefly described below. It should be understood that the following drawings only show some embodiments of the present application and therefore should not be considered as limiting the scope of the present application.

[0068] Figure 1 A schematic diagram of a process flow of a mechanical parts damage detection method based on artificial intelligence provided in an embodiment of the present application;

[0069] Figure 2 A schematic structural diagram of a mechanical parts damage detection system based on artificial intelligence provided in an embodiment of the present application. DETAILED DESCRIPTION

[0070] Below is a combination of the embodiments and Figure 1 and Figure 2 The present application is further described in detail, but the implementation methods of the present application are not limited thereto.

[0071] Reference Figure 1 As shown, it is a flow chart of a mechanical parts damage detection method based on artificial intelligence provided in an embodiment of the present application.

[0072] A mechanical parts damage detection method based on artificial intelligence comprises the following steps:

[0073] Acquire part parameter data, part external inspection data and part internal inspection data of the target mechanical part;

[0074] Setting the damage detection cycle of the target mechanical parts according to the part parameter data;

[0075] The external damage data of the parts include the wear data, crack data and corrosion data of the target mechanical parts; the internal detection data of the parts include the internal signal data of the target mechanical parts;

[0076] Analyze the part wear data of the target mechanical part to obtain a first external damage coefficient corresponding to the target mechanical part; analyze the part corrosion data of the target mechanical part to obtain a second external damage coefficient corresponding to the target mechanical part; analyze the part crack data of the target mechanical part to obtain a third external damage coefficient corresponding to the target mechanical part;

[0077] Obtaining a part external damage coefficient corresponding to a target mechanical part according to the first external damage coefficient, the second external damage coefficient, and the third external damage coefficient;

[0078] Generate an internal signal waveform curve of the target mechanical part according to the internal signal data of the target mechanical part, and obtain a signal curve abnormality coefficient corresponding to the target mechanical part according to the analysis of the internal signal waveform curve and a preset part standard waveform curve;

[0079] The internal damage coefficient of the target mechanical part is obtained according to the abnormal coefficient of the signal curve;

[0080] The part damage coefficient of the target mechanical part is obtained according to the external damage coefficient of the part and the internal damage coefficient of the part, and the part damage coefficient is compared and analyzed with the preset part damage coefficient threshold to determine whether the target mechanical part is abnormal;

[0081] When the target mechanical part is abnormal, the damage abnormality of the target mechanical part is obtained, and the part damage degree of the target mechanical part is obtained according to the damage abnormality.

[0082] The damage detection cycle of the target mechanical parts is set according to the part parameter data, specifically:

[0083] The part parameter data includes the use time of the target mechanical part and the historical inspection cycle of the part;

[0084] The part usage time of the target mechanical part is obtained by taking and marking the part usage time YS; the part history detection period of the target mechanical part is obtained by taking and marking the part history detection period TL;

[0085] The periodic influence coefficient YX of the target mechanical part is calculated by a first calculation function YX=α×YS, where α is a preset proportional factor;

[0086] The damage detection period TJ of the target mechanical part is calculated by the second calculation function TJ=TL-TL×YX.

[0087] In some embodiments, if the target mechanical part has been used for 30 days, the part usage time YS is 30; if the preset proportional factor α is 0.005, the period influence coefficient YX of the target mechanical part calculated by the first calculation function YX=α×YS is 0.15;

[0088] If the historical inspection period of the target mechanical part is 2 days, the historical inspection period TL takes a value of 2. At this time, the damage inspection period TJ of the target mechanical part is calculated by the second calculation function TJ=TL-TL×YX to be 1.85 days.

[0089] The external damage data of parts include the wear data, crack data and corrosion data of target mechanical parts, specifically:

[0090] The parts wear data include the initial volume and initial weight of the target mechanical parts at the initial moment of the damage detection cycle, and the final volume and final weight of the target mechanical parts at the end moment of the damage detection cycle;

[0091] The part corrosion data includes the initial color of the part surface of the target mechanical part at the initial moment of the damage detection cycle, and the detected color of the part surface of the target mechanical part at the end moment of the damage detection cycle;

[0092] The part crack data includes the total number of cracks on the part surface of the target mechanical part, as well as the crack length, crack width and crack depth of the part surface cracks.

[0093] In some embodiments, part wear data, part corrosion data, and part crack data can all be directly acquired by using measuring tools or other means.

[0094] The wear data of the target mechanical parts are analyzed to obtain the first external damage coefficient corresponding to the target mechanical parts, which is:

[0095] The initial volume of the target mechanical part at the initial moment of the damage detection cycle is obtained by taking and marking the initial volume of the part VC; the initial weight of the target mechanical part at the initial moment of the damage detection cycle is obtained by taking and marking the initial weight of the part MC;

[0096] The final volume of the target mechanical part at the end of the damage detection cycle is obtained by taking and marking the final volume VZ of the part; the final weight of the target mechanical part at the end of the damage detection cycle is obtained by taking and marking the final weight MZ of the part;

[0097] Through the third calculation function Calculate the first external damage coefficient OWX corresponding to the target mechanical part, where a 1 、a 2 is the preset scale factor.

[0098] In some embodiments, if the initial volume of the target mechanical part at the initial moment of the damage detection cycle is 1 square meter, the initial volume VC of the part is 1; if the initial weight of the target mechanical part at the initial moment of the damage detection cycle is 20 kg, the initial weight MC of the part is 20;

[0099] If the final volume of the target mechanical part at the end of the damage detection cycle is 0.9 square meters, the final volume VZ of the part is 0.9; if the final weight of the target mechanical part at the end of the damage detection cycle is 18 kg, the final weight MZ of the part is 18;

[0100] If the damage detection period TJ of the target mechanical part is 1, if the preset proportional factor a 1 、a 2 is 100, 10, then through the third calculation function The first external damage coefficient OWX corresponding to the target mechanical part is calculated to be 30.

[0101] The corrosion data of the target mechanical parts are analyzed to obtain the second external damage coefficient corresponding to the target mechanical parts, which is:

[0102] Compare the detected color of the target mechanical part's surface at the end of the damage detection cycle with the initial color of the part's surface;

[0103] Marking a part surface region where the detected color of the part surface is different from the initial color of the part surface as an abnormal surface region, and the target mechanical part contains at least one abnormal surface region;

[0104] Obtain the abnormal surface area corresponding to the abnormal surface region, and take and mark the abnormal surface area to obtain the abnormal surface area SY; obtain the total surface area of ​​the target mechanical part, and take and mark the total surface area of ​​the part to obtain the total surface area SZ;

[0105] Through the fourth calculation function The surface corrosion coefficient FX of the target mechanical part is calculated, where b is a preset proportional factor, j is the number of the abnormal surface area, and n is the total number of the abnormal surface areas;

[0106] The surface corrosion degree corresponding to the abnormal surface area is obtained according to the surface detection color of the part in the abnormal surface area and the initial color of the part surface, and the surface corrosion degree is taken and marked to obtain the surface corrosion degree FD;

[0107] Through the fifth calculation function The second external damage coefficient SWX corresponding to the target mechanical part is calculated, where c 1 、c 2 is a preset scaling factor, j is the number of the abnormal surface area, and n is the total number of abnormal surface areas.

[0108] In some embodiments, obtaining the surface corrosion degree corresponding to the abnormal surface area based on the detected color of the part surface in the abnormal surface area and the initial color of the part surface means obtaining the surface corrosion degree corresponding to the abnormal surface area based on the color change degree between the detected color of the part surface and the initial color of the part surface. The specific value of the surface corrosion degree can be obtained based on a color change degree-surface corrosion degree comparison table, wherein the color change degree-surface corrosion degree comparison table can be preset based on the historical detection data of the target mechanical part;

[0109] If the abnormal surface area corresponding to the abnormal surface region is 0.5 square meters, the abnormal surface area SY is taken as 0.5; if the total surface area of ​​the target mechanical part is 10 square meters, the total surface area SZ is taken as 10; if the surface corrosion degree corresponding to the abnormal surface region is 3 according to the color change degree-surface corrosion degree comparison table, the surface corrosion degree FD is taken as 3.

[0110] The part crack data of the target mechanical part is analyzed to obtain the third external damage coefficient corresponding to the target mechanical part, which is:

[0111] The crack length of the crack on the surface of the part is measured and marked to obtain the crack length LC; the crack width of the crack on the surface of the part is measured and marked to obtain the crack width LK; the crack depth of the crack on the surface of the part is measured and marked to obtain the crack depth LS;

[0112] By the sixth calculation function LCX=d 1 ×LC+d 2 ×LK+d 3 ×LS calculates the crack parameter coefficient LCX corresponding to the surface crack of the part, where d 1 ,d 2 ,d 3 is the preset scaling factor;

[0113] The total number of cracks on the surface of the target mechanical part is taken and marked to obtain the total number of cracks m;

[0114] Through the seventh calculation function The third external damage coefficient SWX corresponding to the target mechanical part is calculated, where i is the number of the crack on the part surface.

[0115] In some embodiments, if the crack length of a crack on the surface of a part is 0.02m, the crack length LC is 0.02; if the crack width of the crack on the surface of the part is 0.01m, the crack width LK is 0.01; if the crack depth of the crack on the surface of the part is 0.01m, the crack depth LS is 0.01. If the preset proportional factor d 1 ,d 2 ,d 3 is 100, 100, 100, then through the sixth calculation function LCX = d 1 ×LC+d 2 ×LK+d 3 ×LS calculation shows that the crack parameter coefficient LCX corresponding to the surface crack of the part is 4.

[0116] The external damage coefficient of the target mechanical part is obtained according to the first external damage coefficient, the second external damage coefficient and the third external damage coefficient, which is:

[0117] By the eighth calculation function WSX=e 1 ×OWX+e 2 ×TWX+e 3 ×SWX calculates the external damage coefficient WSX of the target mechanical part, where e 1 、e 2 、e 3 is the preset scaling factor.

[0118] In some embodiments, if the first external damage coefficient OWX corresponding to the target mechanical part is 30, the second external damage coefficient TWX corresponding to the target mechanical part is 30, and the third external damage coefficient SWX corresponding to the target mechanical part is 20, if the preset proportionality factor e 1 、e 2 、e 3 is 1, 0.8, 1, then through the eighth calculation function

[0119] WSX=e 1 ×OWX+e 2 ×TWX+e 3 ×SWX calculates the external damage coefficient WSX of the target mechanical part to be 74.

[0120] The internal signal waveform curve of the target mechanical part is generated according to the internal signal data of the target mechanical part, and the signal curve abnormality coefficient corresponding to the target mechanical part is obtained according to the analysis of the internal signal waveform curve and the preset part standard waveform curve, which is specifically:

[0121] Compare the internal signal waveform curve of the target mechanical part with the preset part standard waveform curve;

[0122] Mark the waveform curve of the non-overlapping part of the internal signal waveform curve and the part standard waveform curve as an abnormal waveform curve; mark the part standard waveform curve corresponding to the abnormal waveform curve as a reference waveform curve;

[0123] Obtain the peak value difference, the number of inflection points and the mean value difference of the slope between the abnormal waveform curve and the reference waveform curve;

[0124] The peak value difference between the abnormal waveform curve and the reference waveform curve is obtained and marked to obtain the peak value difference FC; the number difference of curve inflection points between the abnormal waveform curve and the reference waveform curve is obtained and marked to obtain the number difference of curve inflection points GC; the mean value difference of curve slope between the abnormal waveform curve and the reference waveform curve is obtained and marked to obtain the mean value difference of curve slope XC;

[0125] Through the ninth calculation function QYX=f 1 ×FC+f 2 ×GC+f 3 ×XC calculates the signal curve abnormality coefficient QYX corresponding to the target mechanical part, where f 1 、f 2 、f 3 is the preset scaling factor;

[0126] The internal damage coefficient of the target mechanical part is obtained according to the abnormal coefficient of the signal curve, which is:

[0127] The internal damage coefficient NSX of the target mechanical part is calculated by the tenth calculation function NSX=δ×QYX, wherein δ is a preset proportional factor.

[0128] In some embodiments, if the difference in the peak values ​​between the abnormal waveform curve and the reference waveform curve is 10, the peak value difference FC is 10; if the difference in the number of inflection points between the abnormal waveform curve and the reference waveform curve is 3, the difference in the number of inflection points GC is 3; if the difference in the mean values ​​of the slopes between the abnormal waveform curve and the reference waveform curve is 0.2, the mean value difference XC is 0.2; if the preset proportional factor f is 1 、f 2 、f 3 is 1, 1, 10, then through the ninth calculation function QYX = f1 ×FC+f 2 ×GC+f 3 ×XC calculates the signal curve abnormality coefficient QYX corresponding to the target mechanical part to be 15; at this time, if the preset proportional factor δ is 2, the internal damage coefficient NSX of the target mechanical part is calculated by the tenth calculation function NSX=δ×QYX to be 30;

[0129] It should be noted that the internal signal data of the target mechanical parts can be obtained by ultrasonic testing through ultrasonic instruments such as ultrasonic flaw detectors, and the internal signal waveform curve is generated based on the internal signal data of the target mechanical parts obtained from the beginning to the end of the ultrasonic testing; in addition, the preset part standard waveform curve is generated based on the internal signal data of the undamaged target mechanical parts; in addition, the benchmark waveform curve is that part of the part standard waveform curve to which the abnormal waveform curve should correspond.

[0130] The part damage coefficient of the target mechanical part is obtained according to the external damage coefficient of the part and the internal damage coefficient of the part, and the part damage coefficient is compared and analyzed with the preset part damage coefficient threshold to determine whether the target mechanical part is abnormal, and when the target mechanical part is abnormal, the damage abnormality degree of the target mechanical part is obtained, and the part damage degree of the target mechanical part is obtained according to the damage abnormality degree, which is specifically:

[0131] By the eleventh calculation function LSX=g 1 ×WSX+g 2 ×NSX calculates the target mechanical parts damage coefficient LSX, where g 1 , g 2 Preset scaling factors;

[0132] If the part damage coefficient LSX of the target mechanical part is less than the preset part damage coefficient threshold, the target mechanical part is judged to be normal;

[0133] If the part damage coefficient LSX of the target mechanical part is greater than or equal to the preset part damage coefficient threshold, the target mechanical part is judged to be abnormal, and the preset part damage coefficient threshold is taken and marked to obtain the part damage coefficient threshold SXY;

[0134] The damage abnormality degree SYD of the target mechanical part is calculated by the twelfth calculation function SYD=LSX-SXY;

[0135] The damage degree of the target mechanical part is positively correlated with the damage abnormality degree SYD of the target mechanical part.

[0136] In some embodiments, if the external damage coefficient WSX of the target mechanical part is 30, and the internal damage coefficient NSX of the target mechanical part is 20, the preset proportionality factor g 1 , g 2 is 1, 1.2, then through the eleventh calculation function LSX = g 1 ×WSX+g 2 ×NSX calculates the target mechanical part's part damage coefficient LSX to be 54;

[0137] If the preset part damage coefficient threshold is 50, the part damage coefficient threshold SXY is 50, and the damage abnormality SYD of the target mechanical part is calculated by the twelfth calculation function SYD=LSX-SXY to be 4;

[0138] The greater the damage abnormality degree SYD of the target mechanical part, the more serious the degree of damage to the target mechanical part.

[0139] Reference Figure 2 FIG. 1 is a schematic diagram of a mechanical parts damage detection system based on artificial intelligence provided in an embodiment of the present application, the system comprising:

[0140] Data acquisition module: acquires part parameter data, part external detection data and part internal detection data of target mechanical parts;

[0141] Setting the damage detection cycle of the target mechanical parts according to the part parameter data;

[0142] The external damage data of the parts include the wear data, crack data and corrosion data of the target mechanical parts; the internal detection data of the parts include the internal signal data of the target mechanical parts;

[0143] The first analysis module: analyzes the part wear data of the target mechanical part to obtain the first external damage coefficient corresponding to the target mechanical part; analyzes the part corrosion data of the target mechanical part to obtain the second external damage coefficient corresponding to the target mechanical part; analyzes the part crack data of the target mechanical part to obtain the third external damage coefficient corresponding to the target mechanical part;

[0144] The second analysis module: obtains the part external damage coefficient corresponding to the target mechanical part according to the first external damage coefficient, the second external damage coefficient and the third external damage coefficient;

[0145] The third analysis module generates an internal signal waveform curve of the target mechanical part according to the internal signal data of the target mechanical part, and obtains a signal curve abnormality coefficient corresponding to the target mechanical part according to the analysis of the internal signal waveform curve and a preset part standard waveform curve;

[0146] The fourth analysis module: obtaining the internal damage coefficient of the target mechanical part according to the abnormal coefficient of the signal curve;

[0147] Damage detection module: obtains the damage coefficient of the target mechanical part based on the external damage coefficient and the internal damage coefficient of the part, and compares and analyzes the damage coefficient with the preset damage coefficient threshold to determine whether the target mechanical part is abnormal;

[0148] When the target mechanical part is abnormal, the damage abnormality of the target mechanical part is obtained, and the part damage degree of the target mechanical part is obtained according to the damage abnormality.

[0149] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A mechanical parts damage detection method based on artificial intelligence, characterized in that: The following steps are involved: Acquire part parameter data, part external inspection data and part internal inspection data of the target mechanical part; Setting a damage detection cycle of a target mechanical part according to the part parameter data; The part external damage data includes part wear data, part crack data and part corrosion data of the target mechanical part; the part internal detection data includes internal signal data of the target mechanical part; Analyze the part wear data of the target mechanical part to obtain a first external damage coefficient corresponding to the target mechanical part; analyze the part corrosion data of the target mechanical part to obtain a second external damage coefficient corresponding to the target mechanical part; analyze the part crack data of the target mechanical part to obtain a third external damage coefficient corresponding to the target mechanical part; Obtaining a part external damage coefficient corresponding to a target mechanical part according to the first external damage coefficient, the second external damage coefficient, and the third external damage coefficient; Generate an internal signal waveform curve of the target mechanical part according to the internal signal data of the target mechanical part, and obtain a signal curve abnormality coefficient corresponding to the target mechanical part according to the analysis of the internal signal waveform curve and a preset part standard waveform curve; Obtaining the internal damage coefficient of the target mechanical part according to the abnormal coefficient of the signal curve; Obtaining a part damage coefficient of a target mechanical part according to the part external damage coefficient and the part internal damage coefficient, and comparing and analyzing the part damage coefficient with a preset part damage coefficient threshold to determine whether the target mechanical part is abnormal; When the target mechanical part is abnormal, the damage abnormality of the target mechanical part is obtained, and the part damage degree of the target mechanical part is obtained according to the damage abnormality.

2. The method for detecting mechanical parts damage based on artificial intelligence according to claim 1, characterized in that: The damage detection cycle of the target mechanical part is set according to the part parameter data, specifically: The part parameter data includes the use time of the target mechanical part and the part historical detection cycle; The used time of the target mechanical part is obtained by taking and marking the used time of the part to obtain the used time YS of the part; The part historical detection period of the target mechanical part is obtained by taking values ​​and marking the part historical detection period TL; The periodic influence coefficient YX of the target mechanical part is calculated by a first calculation function YX=α×YS, where α is a preset proportional factor; The damage detection period TJ of the target mechanical part is calculated by the second calculation function TJ=TL-TL×YX.

3. The method for detecting mechanical parts damage based on artificial intelligence according to claim 2, characterized in that: The part external damage data includes part wear data, part crack data and part corrosion data of the target mechanical part, specifically: The part wear data includes the initial volume and initial weight of the target mechanical part at the initial moment of the damage detection cycle, and the final volume and final weight of the target mechanical part at the end moment of the damage detection cycle; The part corrosion data includes the initial color of the part surface of the target mechanical part at the initial moment of the damage detection cycle, and the detected color of the part surface of the target mechanical part at the end moment of the damage detection cycle; The part crack data includes the total number of cracks on the part surface of the target mechanical part, as well as the crack length, crack width and crack depth of the part surface cracks.

4. The method for detecting mechanical parts damage based on artificial intelligence according to claim 3 is characterized in that: The part wear data of the target mechanical part is analyzed to obtain a first external damage coefficient corresponding to the target mechanical part, which is specifically: The initial volume of the target mechanical part at the initial moment of the damage detection cycle is obtained by taking and marking the initial volume of the part VC; the initial weight of the target mechanical part at the initial moment of the damage detection cycle is obtained by taking and marking the initial weight of the part MC; The final volume of the target mechanical part at the end of the damage detection cycle is obtained by taking and marking the final volume of the part VZ; The final weight of the target mechanical part at the end of the damage detection cycle is obtained by taking and marking the final weight of the part MZ; Through the third calculation function The first external damage coefficient OWX corresponding to the target mechanical part is calculated, where a1 and a2 are preset proportional factors.

5. The method for detecting mechanical parts damage based on artificial intelligence according to claim 4 is characterized in that: The part corrosion data of the target mechanical part is analyzed to obtain a second external damage coefficient corresponding to the target mechanical part, which is specifically: Compare the detected color of the target mechanical part's surface at the end of the damage detection cycle with the initial color of the part's surface; Marking a part surface region where the detected color of the part surface is different from the initial color of the part surface as an abnormal surface region, wherein the target mechanical part includes at least one abnormal surface region; Obtaining an abnormal surface area corresponding to the abnormal surface region, and taking a value and marking the abnormal surface area to obtain an abnormal surface area SY; Obtaining the total surface area of ​​the target mechanical part, and taking and marking the total surface area of ​​the part to obtain the total surface area SZ of the part; Through the fourth calculation function The surface corrosion coefficient FX of the target mechanical part is calculated, where b is a preset proportional factor, j is the number of the abnormal surface area, and n is the total number of the abnormal surface areas; Obtain the surface corrosion degree corresponding to the abnormal surface area according to the surface detection color of the part in the abnormal surface area and the initial color of the part surface, and obtain the surface corrosion degree FD by taking and marking the surface corrosion degree; Through the fifth calculation function The second external damage coefficient SWX corresponding to the target mechanical part is calculated, wherein c1 and c2 are preset proportional factors, j is the number of the abnormal surface area, and n is the total number of the abnormal surface areas.

6. The method for detecting mechanical parts damage based on artificial intelligence according to claim 5, characterized in that: The part crack data of the target mechanical part is analyzed to obtain a third external damage coefficient corresponding to the target mechanical part, which is specifically: The crack length of the crack on the surface of the part is measured and marked to obtain the crack length LC; the crack width of the crack on the surface of the part is measured and marked to obtain the crack width LK; the crack depth of the crack on the surface of the part is measured and marked to obtain the crack depth LS; The crack parameter coefficient LCX corresponding to the surface crack of the part is calculated by a sixth calculation function LCX=d1×LC+d2×LK+d3×LS, wherein d1, d2, and d3 are preset proportional factors; The total number of cracks on the surface of the target mechanical part is taken and marked to obtain the total number of cracks m; Through the seventh calculation function The third external damage coefficient SWX corresponding to the target mechanical part is calculated, where i is the number of the crack on the part surface.

7. The method for detecting mechanical parts damage based on artificial intelligence according to claim 6, characterized in that: The external damage coefficient of the target mechanical part is obtained according to the first external damage coefficient, the second external damage coefficient and the third external damage coefficient, specifically: The external damage coefficient WSX of the target mechanical part is calculated by the eighth calculation function WSX=e1×OWX+e2×TWX+e3×SWX, wherein e1, e2, and e3 are preset proportional factors.

8. The method for detecting mechanical parts damage based on artificial intelligence according to claim 7, characterized in that: The internal signal waveform curve of the target mechanical part is generated according to the internal signal data of the target mechanical part, and the signal curve abnormality coefficient corresponding to the target mechanical part is obtained according to the analysis of the internal signal waveform curve and the preset part standard waveform curve, which is specifically: Compare the internal signal waveform curve of the target mechanical part with the preset part standard waveform curve; Mark the waveform curve of the non-overlapping part of the internal signal waveform curve and the part standard waveform curve as an abnormal waveform curve; mark the part standard waveform curve corresponding to the abnormal waveform curve as a reference waveform curve; Obtain the peak value difference, the number of inflection points and the mean value difference of the slope between the abnormal waveform curve and the reference waveform curve; The peak value difference between the abnormal waveform curve and the reference waveform curve is obtained and marked to obtain the peak value difference FC; the number difference of curve inflection points between the abnormal waveform curve and the reference waveform curve is obtained and marked to obtain the number difference of curve inflection points GC; the mean value difference of curve slope between the abnormal waveform curve and the reference waveform curve is obtained and marked to obtain the mean value difference of curve slope XC; The signal curve abnormality coefficient QYX corresponding to the target mechanical part is calculated by the ninth calculation function QYX=f1×FC+f2×GC+f3×XC, where f1, f2, and f3 are preset proportional factors; The internal damage coefficient of the target mechanical part is obtained according to the abnormal coefficient of the signal curve, specifically: The internal damage coefficient NSX of the target mechanical part is calculated by the tenth calculation function NSX=δ×QYX, wherein δ is a preset proportional factor.

9. The method for detecting mechanical parts damage based on artificial intelligence according to claim 8, characterized in that: The part damage coefficient of the target mechanical part is obtained according to the part external damage coefficient and the part internal damage coefficient, and the part damage coefficient is compared and analyzed with a preset part damage coefficient threshold to determine whether the target mechanical part is abnormal, and when the target mechanical part is abnormal, the damage abnormality degree of the target mechanical part is obtained, and the part damage degree of the target mechanical part is obtained according to the damage abnormality degree, specifically: The component damage coefficient LSX of the target mechanical component is calculated by the eleventh calculation function LSX=g1×WSX+g2×NSX, wherein the proportional factors of g1 and g2 are preset; If the part damage coefficient LSX of the target mechanical part is less than the preset part damage coefficient threshold, it is judged that the target mechanical part is normal; If the part damage coefficient LSX of the target mechanical part is greater than or equal to the preset part damage coefficient threshold, the target mechanical part is judged to be abnormal, and the preset part damage coefficient threshold is taken and marked to obtain the part damage coefficient threshold SXY; The damage abnormality degree SYD of the target mechanical part is calculated by the twelfth calculation function SYD=LSX-SXY; The part damage degree of the target mechanical part is positively correlated with the damage abnormality degree SYD of the target mechanical part.

10. A mechanical parts damage detection system based on artificial intelligence, the system is used to implement the mechanical parts damage detection method based on artificial intelligence according to any one of claims 1 to 9, characterized in that: The system includes: Data acquisition module: acquires part parameter data, part external detection data and part internal detection data of target mechanical parts; Setting a damage detection cycle of a target mechanical part according to the part parameter data; The part external damage data includes part wear data, part crack data and part corrosion data of the target mechanical part; the part internal detection data includes internal signal data of the target mechanical part; The first analysis module: analyzes the part wear data of the target mechanical part to obtain a first external damage coefficient corresponding to the target mechanical part; analyzes the part corrosion data of the target mechanical part to obtain a second external damage coefficient corresponding to the target mechanical part; analyzes the part crack data of the target mechanical part to obtain a third external damage coefficient corresponding to the target mechanical part; A second analysis module: obtaining a part external damage coefficient corresponding to a target mechanical part according to the first external damage coefficient, the second external damage coefficient and the third external damage coefficient; A third analysis module: generating an internal signal waveform curve of the target mechanical part according to the internal signal data of the target mechanical part, and obtaining a signal curve abnormality coefficient corresponding to the target mechanical part according to the analysis of the internal signal waveform curve and a preset part standard waveform curve; Fourth analysis module: obtaining the internal damage coefficient of the target mechanical part according to the abnormal coefficient of the signal curve; Damage detection module: obtains the part damage coefficient of the target mechanical part according to the external damage coefficient of the part and the internal damage coefficient of the part, and compares and analyzes the part damage coefficient with a preset part damage coefficient threshold to determine whether the target mechanical part is abnormal; When the target mechanical part is abnormal, the damage abnormality of the target mechanical part is obtained, and the part damage degree of the target mechanical part is obtained according to the damage abnormality.

Images