Modal analysis system for turbine blade production

Abstract

The invention relates to the technical field of precision machining, particularly discloses a modal analysis system for turbine blade production, aims to solve the problem of pressure control in blade surface quality evaluation and production processes, and aims to obtain a modal analysis result by collecting blade surface images and extracting surface roughness parameters including root-mean-square roughness, maximum height and arithmetic mean roughness. The invention further provides a pressure anomaly detection method, the pressure applied by equipment in the production process is monitored and analyzed in real time, the pressure is adjusted so as to ensure the surface quality of the blade, further, the correlation between the pressure and the surface roughness is analyzed, accurate pressure regulation and control are achieved through a closed-loop control strategy, and the accuracy of the pressure anomaly detection is improved. And the surface quality and the production efficiency of the turbine blade are effectively improved, and the method has important application value in the industries of aerospace, power machinery and the like.

Description

Technical Field

[0001] The invention relates to the technical field of turbine blade surface quality detection, and in particular to a modal analysis system for turbine blade production. Background Art

[0002] Turbine blades are key components that work in high temperature and high pressure environments such as aircraft engines and gas turbines. Their surface quality directly affects the performance and service life of the engine. With the rapid development of the aviation industry and the increasing demand for energy conservation and emission reduction, the manufacturing technology of turbine blades is also constantly improving. Traditionally, the surface quality inspection of turbine blades mainly relies on manual visual inspection or the use of simple measuring tools. This detection method is inefficient and easily affected by subjective factors, and it is difficult to meet the high-precision and high-efficiency requirements of modern manufacturing. In recent years, with the development of automation technology, image processing technology and artificial intelligence algorithms, automated turbine blade surface quality detection systems have gradually become possible. These systems are usually combined with advanced sensing technology and computer vision algorithms to achieve rapid identification and positioning of blade surface defects. However, most of the existing detection systems only focus on the single indicator of surface roughness and ignore other important surface characteristics, which limits their ability to comprehensively evaluate the overall quality of the blade.

[0003] The existing turbine blade surface quality detection technologies on the market mainly include optical detection method, contact detection method and non-contact detection method. Among them, the optical detection method uses optical microscopes or laser scanners and other equipment to collect blade surface images, and performs image analysis through software to obtain surface roughness indicators. Although this method can provide relatively accurate data, it still has limitations for the detection of complex curved surfaces. The contact detection method directly contacts the blade surface with a probe and records the displacement change of the probe to calculate the surface roughness. Although this method can obtain accurate surface morphology data, it is easy to damage the blade during the detection process and the detection speed is slow. The non-contact detection method uses non-contact sensors such as infrared and ultrasonic waves to obtain blade surface information. This method can avoid physical contact with the blade and reduce the risk of damage, but its accuracy is greatly affected by the performance of the sensor. Overall, although the existing turbine blade surface quality detection technology has improved the detection efficiency and accuracy to a certain extent, there are still problems such as single indicator evaluation, manual pressure adjustment, and lack of closed-loop control, which makes it difficult to meet the current high standards for blade surface quality in the aviation manufacturing industry.

[0004] In order to solve the problems existing in the above-mentioned prior art, the present invention proposes a modal analysis system and method for turbine blade production. The system adopts a variety of advanced surface quality assessment technologies, which can not only measure the common arithmetic mean roughness, but also simultaneously obtain multiple roughness indicators such as root mean square roughness and maximum height, so as to achieve a comprehensive assessment of the surface quality of the blade. In addition, the system also integrates an automated pressure monitoring and closed-loop control module, which can realize real-time monitoring and automatic adjustment of the pressure applied by the equipment, ensuring that the pressure value in the blade forming process is stable within the optimal range, greatly improving the production efficiency and product quality. By collecting the surface roughness data of blades produced at different pressure levels and performing correlation analysis on them, the pressure setting in the production process is further optimized to achieve intelligent decision-making. The modal analysis system and method for turbine blade production proposed in the present invention not only overcome the limitations of the prior art, but also can effectively improve the production quality and efficiency of turbine blades, reduce the scrap rate, and bring significant technological progress to the turbine blade manufacturing industry. Summary of the invention

[0005] The purpose of the present invention is to provide a modal analysis system for turbine blade production. The core technical principle of the system is to comprehensively use multi-index surface quality evaluation technology, automated pressure monitoring and closed-loop control system, and data analysis and intelligent decision-making modules. Specifically, in terms of surface quality evaluation, the system uses multiple indicators such as arithmetic mean roughness, root mean square roughness and maximum height. The surface profile data of the turbine blade is obtained through precision sensors, and relevant indicators are calculated to comprehensively evaluate the surface quality, including microscopic and macroscopic characteristics. In order to ensure the consistency and stability of pressure during the production process, the present invention integrates a high-precision pressure sensor and dynamically adjusts the pressure applied by the mold to the blade through a closed-loop control system. Ensure that it remains within the optimal range. In addition, the system automatically records the surface quality indicators and corresponding production conditions (such as pressure) during each inspection process, and stores these data in the database. By analyzing the collected data with statistical methods and machine learning algorithms, the system can identify the changing patterns of surface quality under different production conditions, and make optimization suggestions accordingly, such as recommending the most suitable production pressure range to obtain the best surface quality. In summary, the present invention can achieve precise control and efficient detection of the surface quality of turbine blades through the application of the above technical principles, effectively improve the production efficiency and quality level of turbine blades, which is of great significance to improving the competitiveness of the turbine blade manufacturing industry. To solve the above-mentioned background problems.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A modal analysis system for turbine blade production, comprising an image acquisition and processing module, a pressure monitoring and analysis module, a correlation analysis module and a closed-loop control module;

[0008] The image acquisition and processing module is used to acquire blade surface images;

[0009] Processing the leaf surface image to extract the morphological features of the leaf surface;

[0010] Among them, the morphological characteristics include the root mean square roughness, maximum height and arithmetic mean roughness of the leaf surface;

[0011] The morphological features of the blade surface are processed to obtain the blade surface roughness score, and the qualified and unqualified signals of the blade surface are generated;

[0012] The pressure monitoring and analysis module obtains the pressure value applied by the equipment based on the unqualified signal on the blade surface;

[0013] Process the pressure value to generate normal and abnormal pressure signals;

[0014] The correlation analysis module analyzes the correlation between the blade surface roughness and the pressure applied by the equipment based on the blade surface unqualified signal and the pressure abnormality signal;

[0015] The closed-loop control module performs closed-loop control based on the correlation between the blade surface roughness and the pressure applied by the equipment.

[0016] As a further solution of the present invention: the process of obtaining the morphological characteristics of the blade surface is:

[0017] During the production process of turbine blades, blade surface images of turbine blades are collected;

[0018] Processing the blade surface image, mapping all pixel points in the blade surface image to a plane coordinate system;

[0019] Obtain the grayscale value I of each pixel, process the grayscale values ​​of all pixels according to the mean calculation formula, and obtain the average grayscale value Ij of the leaf surface image;

[0020] By formula The root mean square roughness R of the blade surface is calculated q ;

[0021] Where M is the width of the leaf surface image, N is the height of the leaf surface image, I(x, y) is the gray value of the pixel (x, y), and Ij is the average gray value of the leaf surface image;

[0022] Obtain the maximum grayscale value Imax and the minimum grayscale value Imin in the leaf surface image;

[0023] The maximum grayscale value Imax in the leaf surface image and the minimum grayscale value Imin in the leaf surface image are calculated by difference, and the maximum vertical distance between the highest point and the lowest point of the leaf surface image is obtained, which is recorded as the maximum height R z ;

[0024] By formula The arithmetic mean of the absolute values ​​of all pixels in the blade surface image is calculated and recorded as the arithmetic mean roughness Ra.

[0025] As a further solution of the present invention: the specific process of obtaining the blade surface roughness score is:

[0026] By formula CCD = w1 × R q +w2×R z +w3×R a The blade surface roughness score CCD is calculated;

[0027] Where w1 is R q The weight coefficient of R z The weight coefficient of R a The weight coefficient of .

[0028] As a further solution of the present invention, the process of generating qualified and unqualified signals of blade surfaces is as follows:

[0029] comparing the blade surface roughness score to a blade surface roughness score threshold;

[0030] If the blade surface roughness score is less than or equal to the blade surface roughness score threshold, a blade surface qualified signal is generated;

[0031] If the blade surface roughness score is greater than the blade surface roughness score threshold, a blade surface unqualified signal is generated.

[0032] As a further solution of the present invention: based on the blade surface unqualified signal, the process of processing the pressure value is:

[0033] During the blade forming process, the monitoring period is divided into a number of time subunits of equal length;

[0034] Obtain the pressure value applied by the equipment when each time sub-unit is in the middle of the blade forming process, and record it as the unit pressure value;

[0035] All unit pressure values ​​are processed to obtain a unit pressure value group;

[0036] The unit pressure values ​​of the unit pressure value group are processed according to the mean value calculation formula to obtain the average value of the unit pressure values;

[0037] Based on the average value of the unit pressure value, the unit pressure value group is processed according to the standard deviation calculation formula to obtain the standard deviation of the unit pressure value;

[0038] The unit pressure value and the average value of the unit pressure value are calculated to obtain the deviation value of the unit pressure value;

[0039] The deviation value of the unit pressure value is calculated by ratio with the standard deviation of the unit pressure value to obtain the standard ratio of the unit pressure value.

[0040] As a further solution of the present invention: based on the standard ratio of the unit pressure values, the standard ratio of the unit pressure values ​​is compared with a standard ratio threshold of the unit pressure values;

[0041] If the standard ratio of the unit pressure value is greater than the standard ratio threshold of the unit pressure value, a pressure abnormality signal is generated;

[0042] If the standard ratio of the unit pressure value is less than or equal to the standard ratio threshold of the unit pressure value, a pressure normal signal is generated.

[0043] As a further solution of the present invention, based on the blade surface unqualified signal and the pressure abnormality signal, the specific process of analyzing the correlation between the blade surface roughness and the pressure applied by the equipment is as follows:

[0044] Get the root mean square roughness R of each blade produced at different pressure levels qs ;

[0045] Get the root mean square roughness R of each blade qs The pressure value P applied by the corresponding equipment;

[0046] The root mean square roughness R of each blade qs One-to-one correspondence is made with the pressure value P applied by the equipment to obtain the data pair (R qs , P);

[0047] All data pairs are integrated to obtain the data set (R qs,i , P i );

[0048] According to the mean calculation formula, the root mean square roughness of all blades is processed to obtain the mean root mean square roughness Rj of the blades. qs ;

[0049] The pressure values ​​applied by all devices are processed according to the mean calculation formula to obtain the mean value Pj of the pressure values ​​applied by the devices;

[0050] By formula The correlation xs between the blade surface roughness and the pressure applied by the equipment is calculated;

[0051] Wherein, i is the number of data pairs, i=1,2,...,n.

[0052] As a further solution of the present invention: based on the correlation xs between the surface roughness of the blade and the pressure applied by the device, the absolute value of the correlation value xs is compared with a correlation threshold;

[0053] If |xs| is greater than or equal to the correlation threshold, it indicates that there is a strong linear relationship between the blade surface roughness and the pressure applied by the equipment, and a correlation signal is generated;

[0054] If |xs| is less than the correlation threshold, there is no obvious linear relationship between the blade surface roughness and the pressure applied by the device, and a non-correlation signal is generated.

[0055] As a further solution of the present invention: the specific process of regulating the pressure applied by the device through the closed-loop control strategy is:

[0056] Closed-loop control is performed based on the correlation between the blade surface roughness and the pressure applied by the equipment;

[0057] By formula The pressure value control amount YLZ from the real-time pressure value to the target pressure value is calculated;

[0058] Among them, CZ is the error value between the target pressure value and the real-time pressure value, k1, k2, k3 are the coefficients of the formula, It is an integral term calculated based on the accumulation of the error value CZ over time. CZ(τ) is the error value between the actual pressure and the target pressure at time τ. dτ is a small increment of the time variable τ. t is the current time. is the differential term calculated based on the rate of change of the error value CZ, is the rate of change of the error value CZ with time t;

[0059] Based on the pressure value control amount, adjust the pressure applied by the device.

[0060] As a further solution of the present invention: after regulating the pressure applied by the device, the surface of the blade is continuously monitored during the production process;

[0061] If the surface of the produced blade is still abnormal, repeat the above steps.

[0062] Beneficial effects of the present invention:

[0063] (1) In view of the shortcomings of the existing turbine blade surface quality detection technology, the present invention proposes a system based on modal analysis. The system can accurately identify and locate tiny defects on the surface of turbine blades by introducing advanced vibration signal processing technology and machine learning algorithms. Compared with traditional detection methods, the present invention has significant advantages. On the one hand, it greatly improves the detection efficiency and can complete a comprehensive inspection of a large number of blades in a short time. On the other hand, due to the use of high-precision sensors and complex algorithm models, the detection accuracy is greatly improved, thereby effectively avoiding the safety hazards caused by missed detection or false detection. In addition, the present invention also pays special attention to the intelligence of the system. By integrating intelligent technology, it realizes automated data analysis and decision support functions, reduces the workload of operators, and can continuously optimize the detection strategy based on historical data, further enhancing the stability and reliability of the system.

[0064] (2) In the specific application process, the technical solution of the present invention fully considers the complexity and variability of the actual industrial environment. By conducting in-depth research on the modal characteristics of different parts of the blade, the present invention can adjust the detection parameters according to the characteristics of different types of turbine blades to ensure that the optimal state can be maintained throughout the entire detection process. At the same time, in order to better meet the needs of on-site operations, the present invention also adds portable design elements, making the entire system more flexible and easy to use. It can be fixedly installed at the end of the production line as a final inspection device, or it can be moved to any position of the production line at any time for online inspection, which greatly improves the versatility and practicality of the equipment. In addition, the present invention also takes into account the cost control issue. While ensuring high performance, low-cost but reliable components are used as much as possible, which reduces the overall purchase and maintenance costs, and helps promote the popularization and application of this technology in a wider range.

[0065] (3) From a long-term perspective, the present invention is of great significance for improving the technical level and product quality of the entire turbine manufacturing industry. With the continuous improvement of the performance requirements for turbine blades in the fields of aviation, energy, etc., how to ensure the surface quality of blades has become a major bottleneck restricting the development of the industry. The present invention not only solves this problem through innovative modal analysis technology, but also provides a new development direction for future turbine blade manufacturing. In particular, with the development of new generation information technologies such as big data and cloud computing, the intelligent detection solution proposed in the present invention is expected to be deeply integrated with these advanced technologies to form a more powerful comprehensive solution, which will not only help enterprises achieve digital transformation, but also effectively promote the collaborative innovation of the entire upstream and downstream of the industrial chain, and accelerate the pace of the industry moving towards high-end and intelligent directions. In summary, the present invention not only fills the current market gap, but also lays a solid foundation for future technological progress. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] The present invention will be further described below in conjunction with the accompanying drawings.

[0067] Figure 1 It is a flowchart of the specific steps of a modal analysis system for turbine blade production of the present invention;

[0068] Figure 2 is a flow chart of blade surface roughness score in the present invention;

[0069] Figure 3 It is a flow chart of a modal analysis system for turbine blade production according to the present invention. DETAILED DESCRIPTION

[0070] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0071] Example 1

[0072] See also Figure 1 and Figure 3 As shown, the present invention is a modal analysis system for turbine blade production, comprising an image acquisition and processing module, a pressure monitoring and analysis module, a correlation analysis module and a closed-loop control module;

[0073] The image acquisition and processing module is used to acquire blade surface images during the turbine blade production process;

[0074] Process the blade surface image to obtain the blade surface roughness score and generate qualified and unqualified signals of the blade surface;

[0075] The pressure monitoring and analysis module obtains the pressure value applied by the equipment based on the unqualified signal on the blade surface;

[0076] Process the pressure value to generate normal and abnormal pressure signals;

[0077] The correlation analysis module analyzes the correlation between the blade surface roughness and the pressure applied by the equipment based on the blade surface unqualified signal and the pressure abnormality signal;

[0078] The closed-loop control module performs closed-loop control based on the correlation between the surface roughness of the blade and the pressure applied by the equipment;

[0079] The surface quality of the blades is continuously monitored during the production process and the above steps are repeated for adjustment if necessary.

[0080] Example 2

[0081] See also Figure 2 As shown, during the production process of turbine blades, blade surface images of turbine blades are collected;

[0082] The blade surface image is processed, and a corner of the blade surface image is used as the origin, and all pixel points in the blade surface image are mapped to a plane coordinate system, and the center of each pixel point is a coordinate point;

[0083] Obtain the gray value I of each pixel point of the leaf surface image;

[0084] The grayscale values ​​of all pixels are processed according to the mean calculation formula to obtain the average grayscale value Ij of the leaf surface image;

[0085] By formula The root mean square roughness R of the blade surface is calculated q ;

[0086] Where M is the width of the leaf surface image, N is the height of the leaf surface image, I(x, y) is the gray value of the pixel (x, y), and Ij is the average gray value of the leaf surface image;

[0087] It should be noted that the root mean square roughness R q It is more sensitive to subtle changes in the surface, can capture subtle features of the surface more accurately, is more in line with the concept of statistical root mean square error, is more meaningful for evaluating the statistical properties of the surface, and can be used to evaluate the quality of high-precision surfaces;

[0088] Obtain the maximum grayscale value Imax and the minimum grayscale value Imin in the leaf surface image;

[0089] The maximum grayscale value Imax in the leaf surface image and the minimum grayscale value Imin in the leaf surface image are calculated by difference, and the maximum vertical distance between the highest point and the lowest point of the leaf surface image is obtained, which is recorded as the maximum height R z ;

[0090] It should be noted that the meaning of is very intuitive and easy for non-professionals to understand. It is particularly useful for detecting large-scale defects on the surface, such as scratches, pits, etc. It can be used to detect whether there are large defects or damage on the surface, and it can also evaluate the wear resistance of the surface;

[0091] By formula The arithmetic mean of the absolute values ​​of all pixels in the blade surface image is calculated and recorded as the arithmetic mean roughness Ra;

[0092] It should be noted that the calculation of the arithmetic mean roughness Ra is relatively simple and convenient for quick evaluation. It can provide general information about the surface microstructure, be used to evaluate the overall smoothness of the surface, determine the appropriateness of the surface treatment process, and evaluate the fit and wear resistance of parts.

[0093] By formula CCD = w1 × R q +w2×R z +w3×R a The blade surface roughness score CCD is calculated;

[0094] Where w1 is R q The weight coefficient of R z The weight coefficient of R a The weight coefficient of

[0095] It should be noted that the blade surface roughness score is a comprehensive score obtained by integrating multiple roughness indicators, which can more comprehensively evaluate the surface roughness. According to different application scenarios, the importance of a certain indicator can be emphasized by adjusting the weight. Through standardization, the magnitude differences between different indicators can be eliminated, making the weighted sum more meaningful. In the quality control process, the comprehensive score can help determine whether the blade surface meets the expected quality standards. The change trend of the comprehensive score can guide the optimization of the production process, improve production efficiency and product quality, and the comprehensive score can be used as an indicator to detect whether there are defects on the surface, help to find problems in time and take corrective measures;

[0096] comparing the blade surface roughness score to a blade surface roughness score threshold;

[0097] If the blade surface roughness score is less than or equal to the blade surface roughness score threshold, it means that there are no obvious spots and holes on the blade surface, and a blade surface qualified signal is generated;

[0098] If the blade surface roughness score is greater than the blade surface roughness score threshold, it means that there are obvious spots and holes on the blade surface, and a blade surface unqualified signal is generated.

[0099] Example 3

[0100] Based on the above blade surface unqualified signal, since the abnormal pressure in the turbine blade production process will cause the blade surface to be unqualified, it is analyzed and judged whether the blade surface unqualified is related to the abnormal pressure in the turbine blade production process. If there is a relationship of pressure influence, it is convenient to adjust the pressure in the subsequent turbine blade production process, reduce the influence of pressure on blade quality in the turbine blade production process, effectively maintain the stability of blade quality in the production process, and then improve the qualified rate of blades in the turbine blade production process. Specifically:

[0101] In a specific embodiment:

[0102] During the blade forming process, the monitoring period is divided into a number of time subunits of equal length;

[0103] Obtain the pressure value applied by the equipment when each time sub-unit is in the middle of the blade forming process, and record it as the unit pressure value;

[0104] All unit pressure values ​​are processed to obtain a unit pressure value group;

[0105] The unit pressure values ​​of the unit pressure value group are processed according to the mean value calculation formula to obtain the average value of the unit pressure values;

[0106] Based on the average value of the unit pressure value, the unit pressure value group is processed according to the standard deviation calculation formula to obtain the standard deviation of the unit pressure value;

[0107] The unit pressure value and the average value of the unit pressure value are calculated to obtain the deviation value of the unit pressure value;

[0108] The deviation value of the unit pressure value is calculated by ratio with the standard deviation of the unit pressure value to obtain the standard ratio of the unit pressure value;

[0109] comparing the standard ratio of the unit pressure values ​​to a standard ratio threshold of the unit pressure values;

[0110] If the standard ratio of the unit pressure value is greater than the standard ratio threshold of the unit pressure value, it means that the pressure value applied by the device is abnormal, and a pressure abnormality signal is generated;

[0111] If the standard ratio of the unit pressure value is less than or equal to the standard ratio threshold of the unit pressure value, it means that the pressure value applied by the device is normal, and a normal pressure signal is generated;

[0112] It should be noted that using the above steps to process pressure value data for anomaly detection has multiple advantages and functions: first, this method is simple, intuitive and easy to implement, and can be quickly understood and implemented; second, it supports real-time monitoring and is suitable for instant analysis of data in online systems; in addition, by adjusting parameters, it can flexibly adapt to different data characteristics and scenario requirements; finally, the indicators used (such as mean and standard deviation) are easy to understand, making it easy for people with non-technical backgrounds to interpret the results, and the calculation cost is low, which is suitable for situations with limited resources. In terms of function, it can detect possible problems in the system in advance, so that timely measures can be taken to avoid accidents; in the production process, by monitoring the changes in pressure values, product quality can be ensured. The pressure value is always within an acceptable range; based on the results of anomaly detection, a reasonable maintenance plan can be formulated to reduce unnecessary downtime and maintenance costs; by analyzing abnormal data, potential problems in system design can be discovered and optimized and improved; in key equipment or facilities, abnormal pressure values ​​may indicate an impending dangerous situation, and timely identification of these anomalies helps to ensure the safety of personnel and property; finally, monitoring of pressure values ​​is part of compliance with regulatory requirements, and anomaly detection can ensure that operations comply with relevant standards. Through these advantages and functions, it can be seen that using the above method for pressure value anomaly detection can not only improve the reliability and safety of the system, but also help organizations better manage their operations and services.

[0113] Example 4

[0114] Based on the above blade surface unqualified signals and pressure abnormality signals, the influence relationship between blade surface and equipment pressure is analyzed, specifically:

[0115] In a specific embodiment:

[0116] Get the root mean square roughness R of each blade produced at different pressure levels qs ;

[0117] Get the root mean square roughness R of each blade qs The pressure value P applied by the corresponding equipment;

[0118] The root mean square roughness R of each blade qs One-to-one correspondence is made with the pressure value P applied by the equipment to obtain the data pair (R qs , P);

[0119] All data pairs are integrated to obtain the data set (R qs,i , P i );

[0120] According to the mean calculation formula, the root mean square roughness of all blades is processed to obtain the mean root mean square roughness Rj of the blades. qs ;

[0121] The pressure values ​​applied by all devices are processed according to the mean calculation formula to obtain the mean value Pj of the pressure values ​​applied by the devices;

[0122] By formula The correlation xs between the blade surface roughness and the pressure applied by the equipment is calculated;

[0123] Where i is the number of data pairs, i = 1, 2, ..., n;

[0124] Compare the absolute value of the correlation value xs with the correlation threshold;

[0125] If |xs| is greater than or equal to the correlation threshold, it indicates that there is a strong linear relationship between the blade surface roughness and the pressure applied by the equipment, and a correlation signal is generated;

[0126] If |xs| is less than the correlation threshold, there is no obvious linear relationship between the blade surface roughness and the pressure applied by the equipment, and a non-correlation signal is generated;

[0127] It should be noted that by implementing this correlation and causal analysis, the production quality of turbine blades can be significantly improved. Specifically, by analyzing the relationship between pressure changes and surface roughness, process parameters can be optimized to ensure that the product meets the required surface quality standards, thereby reducing the defective rate and improving product quality. In addition, this analysis also helps process optimization, determine the best pressure setting to obtain the ideal surface roughness, save material and energy costs, and improve production efficiency and output. Through analysis, preventive maintenance can also be carried out. When a specific pressure pattern is found to be associated with a higher surface roughness, this may indicate that certain equipment or tools need maintenance or replacement, thereby formulating a more effective preventive maintenance plan and reducing unexpected downtime. In addition, when an abnormal surface roughness is detected, the root cause of the problem can be diagnosed by analyzing the pressure data, helping to quickly locate the problem and reduce troubleshooting time. By establishing a reliable causal model, the trend of surface roughness changes under given pressure conditions can be predicted, thereby enhancing predictive capabilities and better planning production and inventory management. Finally, continuous monitoring and analysis of data can help identify opportunities for improvement, continuously optimize the manufacturing process, support the research and development of new materials and new technologies, accelerate the innovation process, and ultimately promote the competitiveness and development of the enterprise.

[0128] Example 5

[0129] Based on the correlation signal, the pressure applied by the equipment is regulated through a closed-loop control strategy, specifically:

[0130] Based on the abnormal pressure signal, obtain the real-time pressure value applied by the equipment;

[0131] The target pressure value and the real-time pressure value are calculated to obtain the error value CZ between the target pressure value and the real-time pressure value;

[0132] Based on the error value CZ between the target pressure value and the real-time pressure value, the formula The pressure value control amount YLZ from the real-time pressure value to the target pressure value is calculated;

[0133] Among them, k1, k2, k3 are the coefficients of the formula, It is an integral term calculated based on the accumulation of the error value CZ over time. CZ(τ) is the error value between the actual pressure and the target pressure at time τ. dτ is a small increment of the time variable τ. t is the current time. is the differential term calculated based on the rate of change of the error value CZ, is the rate of change of the error value CZ with time t, that is, the rate of change of the error;

[0134] The process of obtaining coefficients k1, k2, and k3 is as follows:

[0135] By formula The coefficient k1 is calculated, where Kb is the coefficient value when the adjusted real-time pressure starts to fluctuate continuously, i.e., the critical gain, and Tc is the time required for the adjusted implementation pressure to reach the final stable value for the first time from the initial value, i.e., the rise time;

[0136] It should be noted that the critical gain Kb refers to the coefficient k1 value when the adjusted real-time pressure produces continuous fluctuations under the control ratio. When the coefficient k1 exceeds this value, the real-time pressure will become unstable and show continuous fluctuation behavior;

[0137] The process of obtaining the critical gain Kb is:

[0138] In a specific embodiment:

[0139] Preset an initial coefficient k 01 , initial coefficient k 02 and the initial coefficient k 03 ;

[0140] Based on the initial coefficient k 01 , initial coefficient k 02 and the initial coefficient k 03 , obtain the initial pressure control amount;

[0141] Based on the initial pressure control amount, the pressure applied by the equipment is adjusted, and the adjusted real-time pressure is monitored in real time;

[0142] Keeping the initial coefficient k 02 and the initial coefficient k03 Under the condition of no change, gradually increase the initial coefficient k 01 When the real-time pressure starts to fluctuate continuously, stop and record the initial coefficient k when the real-time pressure starts to fluctuate continuously. 01 Value, denoted as critical gain Kb;

[0143] The process of obtaining the rise time Tc is:

[0144] Monitor the pressure change when the pressure applied to the equipment is adjusted based on the initial pressure control amount, obtain the pressure change over time, and draw a pressure response curve;

[0145] Based on the pressure response curve, the time required for the pressure to change from the initial value to the final stable value is obtained, that is, the rise time Tc;

[0146] By formula Calculate the coefficient k 02 ;

[0147] By formula k 03 =Kb*Tc to calculate the coefficient k 03 ;

[0148] Based on the pressure value control amount from the real-time pressure value to the target pressure value, the pressure applied by the device to the blade is regulated;

[0149] After regulation, the blade surface continues to be monitored during the production process;

[0150] If it is still abnormal, repeat the above steps;

[0151] It should be noted that, through closed-loop control, the modal analysis system for turbine blade production can realize real-time monitoring of blade morphological characteristics during the production process and automatically adjust the applied pressure to ensure the quality of the blades. The advantages of closed-loop control include precise control, improved stability, strong adaptability, improved response speed, reduced impact of human factors and optimized performance. Specifically, the closed-loop control system can monitor the surface roughness, internal bubbles and blade deformation of the blades in real time, and promptly identify and issue an alarm when abnormal blade morphology is detected. If abnormal pressure is one of the causes of abnormal blade morphology, the closed-loop control system can automatically adjust the pressure to achieve the predetermined target value. The data collected by the closed-loop control system can be used for a more in-depth cause analysis to find out the root cause of the problem, and automatically adjust the applied pressure without human intervention to ensure that the quality of the blades meets the standards. The closed-loop control strategy helps to improve production efficiency, reduce scrap rate, improve product quality, and reduce scrap rate caused by quality problems, thereby significantly improving the overall quality and production efficiency of turbine blades.

[0152] Working principle of the present invention: Aiming at the shortcomings of the existing turbine blade surface quality detection technology, the present invention proposes a system based on modal analysis. The system can accurately identify and locate tiny defects on the surface of turbine blades by introducing advanced vibration signal processing technology and machine learning algorithms. Compared with traditional detection methods, the present invention has significant advantages. On the one hand, it greatly improves the detection efficiency and can complete a comprehensive inspection of a large number of blades in a short time; on the other hand, due to the use of high-precision sensors and complex algorithm models, the detection accuracy is greatly improved, thereby effectively avoiding the safety hazards caused by missed detection or false detection. In addition, the present invention also pays special attention to the intelligence of the system. By integrating intelligent technology, it realizes automated data analysis and decision support functions, reduces the workload of operators, and can continuously optimize the detection strategy based on historical data, further enhancing the stability and reliability of the system.

[0153] In the specific application process, the technical solution of the present invention fully considers the complexity and variability in the actual industrial environment. By conducting in-depth research on the modal characteristics of different parts of the blades, the present invention can adjust the detection parameters according to the characteristics of different types of turbine blades in a customized manner to ensure that the optimal state can be maintained throughout the entire detection process. At the same time, in order to better meet the needs of on-site operations, the present invention also adds portable design elements, making the entire system more flexible and easy to use. It can be fixedly installed at the end of the production line as a final inspection device, and can also be moved to any position of the production line at any time for online inspection, which greatly improves the versatility and practicality of the equipment. In addition, the present invention also takes into account the cost control issue. While ensuring high performance, low-cost but reliable components are used as much as possible, which reduces the overall purchase and maintenance costs, and helps promote the popularization and application of this technology in a wider range.

[0154] From a long-term perspective, the present invention is of great significance for improving the technical level and product quality of the entire turbine manufacturing industry. With the continuous improvement of the performance requirements for turbine blades in the fields of aviation, energy, etc., how to ensure the surface quality of blades has become a major bottleneck restricting the development of the industry. The present invention uses innovative modal analysis technology to not only solve this problem, but also provide a new development direction for future turbine blade manufacturing. Especially with the development of new generation information technologies such as big data and cloud computing, the intelligent detection solution proposed in the present invention is expected to be deeply integrated with these advanced technologies to form a more powerful comprehensive solution, which will not only help enterprises achieve digital transformation, but also effectively promote collaborative innovation of the entire upstream and downstream of the industrial chain, and accelerate the industry's pace towards high-end and intelligent directions. In summary, the present invention not only fills the current market gap, but also lays a solid foundation for future technological progress.

[0155] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A modal analysis system for turbine blade production, characterized in that: It includes an image acquisition and processing module, a pressure monitoring and analysis module, a correlation analysis module and a closed-loop control module; The image acquisition and processing module is used to acquire blade surface images; Processing the leaf surface image to extract the morphological features of the leaf surface; Among them, the morphological characteristics include the root mean square roughness, maximum height and arithmetic mean roughness of the leaf surface; The morphological features of the blade surface are processed to obtain the blade surface roughness score, and the qualified and unqualified signals of the blade surface are generated; The pressure monitoring and analysis module obtains the pressure value applied by the equipment based on the unqualified signal on the blade surface; Process the pressure value to generate normal and abnormal pressure signals; The correlation analysis module analyzes the correlation between the blade surface roughness and the pressure applied by the equipment based on the blade surface unqualified signal and the pressure abnormality signal; The closed-loop control module performs closed-loop control based on the correlation between the blade surface roughness and the pressure applied by the equipment.

2. A modal analysis system for turbine blade production according to claim 1, characterized in that: The process of obtaining the morphological characteristics of the leaf surface is as follows: During the production process of turbine blades, blade surface images of turbine blades are collected; Processing the blade surface image, mapping all pixel points in the blade surface image to a plane coordinate system; Obtain the grayscale value I of each pixel, process the grayscale values ​​of all pixels according to the mean calculation formula, and obtain the average grayscale value Ij of the leaf surface image; By formula The root mean square roughness R of the blade surface is calculated q ; Where M is the width of the leaf surface image, N is the height of the leaf surface image, I(x, y) is the gray value of the pixel (x, y), and Ij is the average gray value of the leaf surface image; Obtain the maximum grayscale value Imax and the minimum grayscale value Imin in the leaf surface image; The maximum grayscale value Imax in the leaf surface image and the minimum grayscale value Imin in the leaf surface image are calculated by difference, and the maximum vertical distance between the highest point and the lowest point of the leaf surface image is obtained, which is recorded as the maximum height R z ; By formula The arithmetic mean of the absolute values ​​of all pixels in the blade surface image is calculated and recorded as the arithmetic mean roughness Ra.

3. A modal analysis system for turbine blade production according to claim 2, characterized in that: The specific process of obtaining the blade surface roughness score is as follows: By formula CCD = w1 × R q +w2×R z +w3×R a The blade surface roughness score CCD is calculated; Where w1 is R q The weight coefficient of R z The weight coefficient of R a The weight coefficient of .

4. A modal analysis system for turbine blade production according to claim 3, characterized in that: The process of generating the pass and fail signals for the blade surface is: comparing the blade surface roughness score to a blade surface roughness score threshold; If the blade surface roughness score is less than or equal to the blade surface roughness score threshold, a blade surface qualified signal is generated; If the blade surface roughness score is greater than the blade surface roughness score threshold, a blade surface unqualified signal is generated.

5. A modal analysis system for turbine blade production according to claim 4, characterized in that: Based on the blade surface unqualified signal, the process of processing the pressure value is as follows: During the blade forming process, the monitoring period is divided into a number of time subunits of equal length; Obtain the pressure value applied by the equipment when each time sub-unit is in the middle of the blade forming process, and record it as the unit pressure value; All unit pressure values ​​are processed to obtain a unit pressure value group; The unit pressure values ​​of the unit pressure value group are processed according to the mean value calculation formula to obtain the average value of the unit pressure values; Based on the average value of the unit pressure value, the unit pressure value group is processed according to the standard deviation calculation formula to obtain the standard deviation of the unit pressure value; The unit pressure value and the average value of the unit pressure value are calculated to obtain the deviation value of the unit pressure value; The deviation value of the unit pressure value is calculated by ratio with the standard deviation of the unit pressure value to obtain the standard ratio of the unit pressure value.

6. A modal analysis system for turbine blade production according to claim 5, characterized in that: Based on the standard ratio of the unit pressure values, comparing the standard ratio of the unit pressure values ​​with a standard ratio threshold of the unit pressure values; If the standard ratio of the unit pressure value is greater than the standard ratio threshold of the unit pressure value, a pressure abnormality signal is generated; If the standard ratio of the unit pressure value is less than or equal to the standard ratio threshold of the unit pressure value, a pressure normal signal is generated.

7. A modal analysis system for turbine blade production according to claim 6, characterized in that: Based on the blade surface unqualified signal and pressure abnormality signal, the specific process of analyzing the correlation between the blade surface roughness and the pressure applied by the equipment is as follows: Get the root mean square roughness R of each blade produced at different pressure levels qs ; Get the root mean square roughness R of each blade qs The pressure value P applied by the corresponding equipment; The root mean square roughness R of each blade qs One-to-one correspondence is made with the pressure value P applied by the equipment to obtain the data pair (R qs , P); All data pairs are integrated to obtain the data set (R qs,i , P i ); According to the mean calculation formula, the root mean square roughness of all blades is processed to obtain the mean root mean square roughness Rj of the blades. qs ; The pressure values ​​applied by all devices are processed according to the mean calculation formula to obtain the mean value Pj of the pressure values ​​applied by the devices; By formula The correlation xs between the blade surface roughness and the pressure applied by the equipment is calculated; Wherein, i is the number of data pairs, i=1,2,...,n.

8. A modal analysis system for turbine blade production according to claim 7, characterized in that: Based on the correlation xs between the surface roughness of the blade and the pressure applied by the device, the absolute value of the correlation value xs is compared with a correlation threshold value; If |xs| is greater than or equal to the correlation threshold, it indicates that there is a strong linear relationship between the blade surface roughness and the pressure applied by the equipment, and a correlation signal is generated; If |xs| is less than the correlation threshold, there is no obvious linear relationship between the blade surface roughness and the pressure applied by the device, and a non-correlation signal is generated.

9. A modal analysis system for turbine blade production according to claim 8, characterized in that: The specific process of regulating the pressure applied by the equipment through a closed-loop control strategy is as follows: Closed-loop control is performed based on the correlation between the blade surface roughness and the pressure applied by the equipment; By formula The pressure value control amount YLZ from the real-time pressure value to the target pressure value is calculated; Among them, CZ is the error value between the target pressure value and the real-time pressure value, k1, k2, k3 are the coefficients of the formula, It is an integral term calculated based on the accumulation of the error value CZ over time. CZ(τ) is the error value between the actual pressure and the target pressure at time τ. dτ is a small increment of the time variable τ. t is the current time. is the differential term calculated based on the rate of change of the error value CZ, is the rate of change of the error value CZ with time t; Based on the pressure value control amount, adjust the pressure applied by the device.

10. A modal analysis system for turbine blade production according to claim 9, characterized in that: After regulating the pressure applied by the equipment, the blade surface continues to be monitored during the production process; If the surface of the produced blade is still abnormal, repeat the above steps.

Images