KR102280126B1 - Apparatus and method for ultrasonic signal evaluation in the turbine blade using artificial intelligence - Google Patents

Abstract

Disclosed is an ultrasonic signal evaluation device of a turbine blade using artificial intelligence. An apparatus for evaluating an ultrasonic signal of a turbine blade using artificial intelligence according to the present invention includes: a robot ultrasonic scanner for measuring an ultrasonic signal while moving the turbine blade; a signal acquisition unit for determining the validity of the signal acquired from the robotic ultrasound scanner; a signal evaluation unit for determining whether the turbine blade is defective by comparing the valid ultrasonic signal determined by the signal acquisition unit with an evaluation reference signal; a signal storage unit for storing information on the acquired signal, the evaluation reference signal, and the defect; and an evaluation criterion learning unit for learning the evaluation reference signal and the signal for the defect.

Description

Ultrasonic signal evaluation device and evaluation method of the turbine blade root using artificial intelligence {APPARATUS AND METHOD FOR ULTRASONIC SIGNAL EVALUATION IN THE TURBINE BLADE USING ARTIFICIAL INTELLIGENCE}

The present invention relates to an ultrasonic signal evaluation apparatus and evaluation method for diagnosing the health of the root of a turbine blade using artificial intelligence. More specifically, when inspecting the ultrasonic signal to check the health of the root of the turbine blade, the signal is precisely acquired through an automated device, the valid signal is determined among the acquired signals, and the sound signal is read after reading the soundness of the signal. It relates to a device and method that can continuously increase the accuracy and reliability of a diagnosis by storing it, learning it using artificial intelligence, and using it as a standard for re-evaluation.

A turbine blade is a blade-shaped part that generates rotational force by steam, and is a core part constituting a steam turbine and a gas turbine of a power plant. A turbine blade is a device that generates rotational force when steam or gas generated by a boiler or other equipment passes through a turbine at high temperature and high pressure. A turbine rotor is a device that rotates at high speed. Thousands of tons of centrifugal force is concentrated at the root of the turbine blade, which supports the turbine blade. In particular, since the last blade of a low-pressure turbine is the longest and heaviest, a very large centrifugal force of thousands of tons is concentrated on the blade root during operation. If the turbine blade root part is damaged, it is necessary to check the integrity of the turbine blade root part because it leads to an accident in which the turbine-generator system is all destroyed.

Turbine equipment manufacturers recommend disassembling the turbine blades from the turbine rotor to check the soundness, but since a lot of time and money are incurred, in reality, inspection methods are being developed to check the soundness without disassembling the blades from the rotor.

Since the turbine blade root part is fastened to the turbine rotor and cannot be visually confirmed from the outside, in order to check the health of the root part, the health can be checked by disassembling the turbine blade or sending ultrasonic waves to the root part to check the reflected wave. When inspecting the ultrasonic signal for the turbine blade root, the inspector acquires a signal in a narrow space between the turbine blade and the blade, which takes a lot of time and has a problem in that the accuracy of the acquired signal is also lowered. In addition, since micro cracks as small as 0.5 mm must be detected at the root of the turbine blade, the ultrasonic inspection method is very difficult and requires a lot of experience and high skill. However, ultrasound examination may cause measurement errors and differences in the reading of results depending on the examiner, thereby lowering the reliability of the examination, and may cause distrust and conflict with the examination results.

The robot ultrasonic scanner for facilitating ultrasonic scanning of turbine blades is the movement of the ultrasonic sensor in <Korea Patent Application No. 10-2018-0092787> "Sensor transfer device for ultrasonic inspection of the root of the turbine blade" of the applicant's previous application The method has been realized as a prior art, and the present invention relates to a sensor movement and signal acquisition method for reducing errors compared to the prior art through the robotic ultrasound scanner.

A turbine rotor consists of several stages, and all turbine blades of the same stage have a uniform shape. Therefore, it automatically detects signals through the robot ultrasonic scanner in the turbine blade route, selects only the precisely measured valid signals by determining the validity of the acquired signals, and evaluates the soundness of the signals to improve human factors. Differences in test results caused by In addition, when it is determined as a sound signal, the signal is stored in the form of big data, learned, and the evaluation reference signal is updated and used as a new evaluation criterion, thereby continuously increasing the accuracy and reliability of the examination.

Therefore, in the present invention, an apparatus and method for precisely acquiring an ultrasonic signal for the root of a turbine blade and determining the soundness, and by storing the sound signal and learning using artificial intelligence to continuously update the evaluation criteria, the accuracy and An object of the present invention is to provide an apparatus and method for increasing reliability.

Patent Document 1. Korean Patent Laid-Open No. 10-2014-0017789 (Method and Apparatus for Laser Ultrasonic Imaging of Rotating Structures)

An object of the present invention is to provide an apparatus and method for accurately acquiring an ultrasonic signal and determining the health of a turbine blade root part using artificial intelligence.

The present invention provides an apparatus and method that can continuously increase the accuracy and reliability of diagnosis by storing sound signals and continuously updating the evaluation criteria using artificial intelligence, so that the test results that can be generated by human factors in the prior art are provided. The purpose is to reduce errors and judgment errors, and to secure the reliability of evaluation.

According to an embodiment of the present invention, an apparatus for evaluating an ultrasonic signal of a turbine blade using artificial intelligence includes: a robot ultrasonic scanner for measuring an ultrasonic signal while moving the turbine blade; a signal acquisition unit for determining the validity of the signal acquired from the robotic ultrasound scanner; a signal evaluation unit for determining whether the turbine blade is defective by comparing the valid ultrasonic signal determined by the signal acquisition unit with an evaluation reference signal; a signal storage unit for storing information on the acquired signal, the evaluation reference signal, and the defect; and an evaluation criterion learning unit for learning the evaluation criterion signal and the signal for the defect using artificial intelligence.

In addition, the robot ultrasound scanner may move the root of the turbine blade.

In addition, the robot ultrasonic scanner may obtain a primary signal while moving from the start point to the end point of the turbine blade, and obtain a secondary signal while moving from the end point to the start point.

Also, the signal acquisition unit may determine the validity of the acquired signal by comparing the reflected position and reflected energy intensity (dB) of ultrasonic waves from the primary signal and the secondary signal.

In addition, the robot ultrasound scanner, ultrasound sensors arranged in an array form; a chain for moving the ultrasonic sensor; and a limit sensor attached to the start point and the end point at which the ultrasonic sensor moves in the turbine blade.

In addition, as for the deviation of the primary signal and the secondary signal, a signal in which the difference in the reflection position is within 1% and the difference in the intensity of the reflection energy is within 3% may be determined as a valid signal.

In addition, the signal evaluation unit determines a defect signal including the evaluation reference signal and the ultrasonic reflection signal, calculates the size of the defect by the following equation with respect to the evaluation reference signal and the defect signal, and the size of the defect A signal evaluation process of determining whether to pass or not based on a predetermined size for the .

In addition, the evaluation criterion learning unit stores the sound signal for which the pass determination has been made in the signal storage unit, performs an evaluation criterion learning algorithm to update the evaluation criteria, and reflects the updated evaluation criteria back to the signal evaluation process can do.

According to an embodiment of the present invention, there is provided an ultrasonic signal evaluation method of a turbine blade using artificial intelligence, comprising: an ultrasonic signal acquisition step of measuring an ultrasonic signal by moving a robot ultrasonic scanner along the turbine blade; a valid signal reading step of receiving the ultrasound signal acquired from the robot ultrasound scanner and determining the validity of the ultrasound signal; a signal evaluation step of determining whether the turbine blade is defective by comparing the valid ultrasonic signal determined in the signal acquisition step with an evaluation reference signal; and an evaluation criterion learning step of learning the evaluation criterion signal and the signal for the defect by performing a predetermined artificial intelligence learning algorithm.

Through the ultrasonic signal evaluation apparatus and evaluation method using the artificial intelligence of the present invention, the lack of precision of the inspection result that occurs in the process of manually mounting and measuring the ultrasonic sensor by human factors in the prior art, and the ambiguity of the reading standard of the inspecting party The reliability of evaluation can be secured by reducing judgment errors.

In addition, the evaluation criterion learning unit of the ultrasonic signal evaluation apparatus of the turbine blade using the artificial intelligence of the present invention stores and learns the passed sound signal in the form of big data, updates the evaluation criterion signal, and uses it as a new evaluation criterion to continuously test It can improve accuracy and reliability.

1 is a conceptual diagram of an automatic ultrasonic evaluation apparatus for diagnosing the health of a turbine blade root portion according to the present invention.
2 is a block diagram of a robot ultrasonic scanner connected to an ultrasonic signal evaluation apparatus of a turbine blade using artificial intelligence according to the present invention.
3 is an operation flowchart of an ultrasound signal evaluation process of the ultrasound signal evaluation apparatus according to the present invention.
4 is a view for explaining an operation process of the ultrasonic signal evaluation apparatus according to the present invention.
5 is an operation flowchart for each process of the ultrasound evaluation apparatus according to the present invention.
6 is an example of the shape of a turbine blade root portion to which the present invention is applied.
7 is an example of the ultrasonic wave signal of the healthy part and the defective part evaluated in the present invention.
8 is an example of ultrasound image signals of a healthy part and a defective part evaluated in the present invention.

Since the present invention can have various changes and can have various embodiments, specific embodiments are illustrated in the drawings and will be described with reference to the detailed description. However, this is not intended to limit the present invention to specific embodiments, and it should be understood to include all modifications, equivalents and substitutes included in the spirit and scope of the present invention. In describing each figure, like reference numerals have been used for like elements. In describing the present invention, if it is determined that a detailed description of a related known technology may obscure the gist of the present invention, the detailed description thereof will be omitted.

The terms used herein are used only to describe specific embodiments, and are not intended to limit the present invention. The singular expression may include the plural expression unless the context clearly dictates otherwise.

In the present specification, terms such as "comprise", "have" or "include" are intended to designate that the features, numbers, steps, operations, components, parts, or combinations thereof described in the specification exist, and one It may be understood that this does not preclude the possibility of the presence or addition of or more other features or numbers, steps, operations, components, parts, or combinations thereof.

In addition, the components shown in the embodiment of the present invention are shown independently to represent different characteristic functions, and it does not mean that each component is composed of separate hardware or a single software component. That is, each component is listed as each component for convenience of description, and at least two components of each component are combined to form one component, or one component can be divided into a plurality of components to perform a function. Integrated embodiments and separate embodiments of each of these components are also included in the scope of the present invention without departing from the essence of the present invention.

In addition, the following embodiments are provided to more clearly explain to those of ordinary skill in the art, and the shapes and sizes of elements in the drawings may be exaggerated for more clear description.

Hereinafter, with reference to the accompanying drawings, a preferred embodiment according to the present invention will be described.

1 is a conceptual diagram of an ultrasonic signal evaluation apparatus of a turbine blade using artificial intelligence according to the present invention, and FIG. 2 is a block diagram of a robot ultrasonic scanner connected to the ultrasonic evaluation apparatus according to the present invention.

1 and 2, the ultrasonic signal evaluation apparatus 001 of a turbine blade using artificial intelligence in an embodiment of the present invention is an ultrasonic signal evaluation apparatus for diagnosing the health of the root of a turbine blade, and a uniform and precise signal It is provided for automated ultrasonic signal acquisition by mounting the robot ultrasonic scanner 100 to the root of the turbine blade as shown in FIG. 1 , instead of passively acquiring the signal with a measuring instrument as in the prior art for acquisition.

The ultrasonic signal automatic evaluation apparatus 001 determines the accuracy and validity of the acquired ultrasonic signal and determines the soundness of only the valid signal, thereby reducing the error of the inspection.

In addition, it is possible to increase the accuracy and reliability of evaluation by converting the passed sound signal into big data, storing and learning, and continuously updating the evaluation criteria.

Referring to FIG. 2 , the ultrasonic signal evaluation apparatus 001 includes a robot ultrasonic scanner 100 for automatically and precisely measuring an ultrasonic signal of the root of a turbine blade, and the precision of a signal obtained from the robot ultrasonic scanner 100 and A signal acquisition unit 200 for determining the validity, a signal evaluation unit 300 for determining the presence or absence of an ultrasonic reflected signal with respect to the valid signal, and reading whether or not it is a defective signal, and storing the acquired signal into big data and a signal storage unit 400 to store and learn a signal read as a sound signal using artificial intelligence, and an evaluation criterion learning unit 500 to continuously update and refine the evaluation criterion.

3 is an operation flowchart of an ultrasound signal evaluation process of the ultrasound evaluation apparatus according to the present invention, and FIG. 4 is a view for explaining the operation process of the ultrasound evaluation apparatus according to the present invention.

3 and 4 , the ultrasonic signal evaluation apparatus 001 of the turbine blade using artificial intelligence receives the ultrasonic signal detected from the turbine blade root from the robot ultrasonic scanner 100 as shown in FIG. 3 . Step (S021), a valid signal reading step (S022) for determining the precision and validity of the ultrasonic signal, and comparing the acquired valid ultrasonic signal with an evaluation reference signal to read the presence or absence of defects in the signal to determine soundness pass/fail A signal evaluation step (S023), a step of storing the signal read as a sound signal into big data in the signal storage unit (S025), and performing an arbitrary learning algorithm to learn and update the evaluation criteria to continuously refine the evaluation criteria And it is possible to perform a process including the evaluation criteria learning step (S024) to accurately improve. In addition, by utilizing the learned and updated evaluation reference signal as the evaluation reference signal of the signal evaluation step ( S023 ) again, it is possible to continuously increase the precision and accuracy in diagnosing the health of the turbine blade root part.

In order to secure the reliability of the ultrasonic inspection on the turbine blade root, uniform and precise signal acquisition must first be made. To this end, the signal must be acquired by an automated ultrasound signal acquisition method rather than a manual method by an examiner's hand.

In general, the ultrasonic signal acquisition is completed by the inspector manually moving the sensor from the beginning to the end of the blade measurement position. In the present invention, in the signal acquisition of the robot ultrasound scanner 100, the primary signal acquisition is completed at the point where the sensor moves from the starting point and ends. In addition, the secondary signal acquisition is completed by returning to the starting point at the part where the sensor is completed. Using this method, it is possible to determine whether a uniform and precise signal is obtained according to the degree to which the two signals acquired in the first and second phases coincide.

Referring to FIG. 4 , the robot ultrasound scanner 100 connected to the ultrasound signal evaluation device 001 moves the ultrasound sensor 107 mounted on the turbine blade 101 to the chain 103 connected to the driving device 102 . to obtain an ultrasonic signal from the blade root. In addition, by attaching a limit sensor to the start point 104 and the end point 105 at which the ultrasonic sensor 107 moves in the turbine blade 101, it can be recognized that the movement of the sensor and signal acquisition are completed. Also, the position of the ultrasonic sensor 107 on the turbine blade 10 may be recorded by the encoder 106 when the ultrasonic sensor 107 is moved by the drive device 102 .

In addition, the ultrasonic sensor 107 moves from the start point 104 to the end point 105 to obtain a primary signal, and returns to the start point 104 from the end point 105 to obtain a secondary signal, By comparing the reflected position and reflected energy intensity (dB) of ultrasonic waves in the acquired primary and secondary signals, it can be read whether precise signal acquisition within a certain standard is achieved despite the difference in blade conditions.

The turbine blade 101 has a single stage of a uniform shape, but a signal acquisition difference may occur for each turbine blade 101 due to the roughness of the surface of the turbine blade 10, minute shape difference, etc. depending on use. . Therefore, when two signals are acquired from one turbine blade 101 and the signals are compared, it can be confirmed whether precise signal acquisition is made even for the condition difference of the turbine blade 10 .

The criteria for whether the ultrasonic signals acquired in the 1st and 2nd stages are precisely made can be set such that the difference in the reflected position of the ultrasonic wave is within ±1% and the difference in the reflected energy intensity (dB) is within ±3%. If this criterion is not satisfied, it is regarded as a measurement error and the signal can be reacquired.

Referring to FIG. 2 , the signal acquisition unit 200 of the ultrasound signal evaluation apparatus 001 is connected to the robotic ultrasound scanner 100 to identify automatically detected signals and select uniform and precise signals, thereby inputting ultrasound signals maintain its effectiveness and adequacy.

Referring to FIG. 5 , in the signal acquisition unit 200 , when the ultrasonic sensor 107 that starts the photosensor at the start point 104 of the robot ultrasonic scanner 100 recognizes the photosensor at the end point 105 , the primary signal Acquisition is made, and secondary signal acquisition can be made as the ultrasonic sensor 107 starts again at the end point 105 and returns to the start point 104 .

이내, 반사 에너지 세기(dB)가

이내인 신호 만을 정밀하게 측정된 유효 신호로 판별할 수 있다(S204). 다음으로, 신호 저장부(400)에 저장함으로써(S205) 측정 상의 오류가 발생하지 않은 유효한 신호 만을 판별해 내는 유효 신호 취득 프로세스(S200)를 수행할 수 있다. 이 기준을 만족하지 못할 경우 경고 신호가 발생하고 로봇 초음파 스캐너(100)에 신호 재취득을 지시하게 된다.Acquisition of the primary signal (S201) measured while moving from the start point 104 to the end point 105, which is generated according to the movement of the ultrasonic sensor 107, and 2 measured while moving from the end point 105 to the start point 104 again By calculating the deviation between the difference signal acquisition (S202) (S203), the difference in the ultrasonic reflection position is

Within, the reflected energy intensity (dB) is

Only a signal within the range can be determined as a precisely measured valid signal (S204). Next, by storing in the signal storage unit 400 (S205), it is possible to perform a valid signal acquisition process (S200) of discriminating only valid signals in which no measurement error occurs. If this criterion is not satisfied, a warning signal is generated and the robot ultrasound scanner 100 is instructed to acquire a signal again.

The signal evaluation unit 300 of the ultrasound signal evaluation apparatus 001 may read and evaluate the presence or absence of a defect with respect to the valid signal confirmed by the signal acquisition unit 200 with appropriate precision. The signal evaluator 300 compares (301) the evaluation criterion input signals 504 stored from the evaluation criterion learning unit 500 to read whether the signal is a healthy signal or a defective signal.

The evaluation primarily evaluates whether an ultrasonic reflected signal is generated in the monitoring target range (S301), and if there is no ultrasonic reflection in the monitoring target range, it is read as a healthy part, and if there is an ultrasonic reflected signal in the monitoring range, it is read as a defect. It can be evaluated whether or not (S302).

In the case of a defect signal having an ultrasonic reflected signal, the sensitivity of the reflected received sound pressure can be calculated as follows based on the dB evaluation standard (S303).

A signal having a size of 0.5 mm or more is rejected, and a defect smaller than 0.5 mm can be evaluated as an area to be observed in the future (S304). No defects were found, but the signal with ultrasonic reflection can be evaluated by re-verification as to whether it is sound or not based on big data for image reading in the future.

The signal storage unit 400 of the ultrasound signal evaluation apparatus 001 may store the signals confirmed that the acquired signals are uniform and precise signals into big data. Depending on the type, stage, and blade number of the turbine, a unique number of the signal can be assigned and stored, and the stored ultrasonic signal can be referenced from the evaluation criterion learning unit 500 for automatic evaluation through artificial intelligence learning of the signal. there is.

The evaluation criterion learning unit 500 of the ultrasound signal evaluation apparatus 001 may provide an evaluation reference signal based on two criteria: an evaluation of an ultrasound waveform signal as shown in FIG. 7 and an evaluation of an ultrasound image signal as shown in FIG. 8 . there is. The evaluation criterion learning unit 500 can use artificial intelligence learning, and stores a sound signal that has been passed from the signal evaluation process to the signal storage unit 400 (S501), and performs the evaluation criterion learning algorithm (S502) , an evaluation criterion may be updated (S503), and an evaluation criterion learning process may be performed such that the updated evaluation criterion signal (S504) may be reflected again in the step (S302) of the signal evaluation process. Here, the artificial intelligence learning is a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), a Restricted Boltzmann Machine (RBM), and a deep neural network. One or more of a Deep Brief Network (DBN) may be used.

The waveform signal input standard can be initialized by inputting a numerical value indicating a monitoring section at a location where cracks are expected to occur in the blade root. In addition, the ultrasonic image signal of the sound part and the ultrasonic image signal of the crack part, which are criteria for evaluating defects by ultrasonic image recognition, may be input and converted into big data.

In the above, the present invention has been described with specific matters such as specific components and limited embodiments and drawings, but these are only provided to help a more general understanding of the present invention, and the present invention is not limited to the above embodiments, Those of ordinary skill in the art to which the present invention pertains may devise various modifications and variations from these descriptions.

Therefore, the spirit of the present invention should not be limited to the above-described embodiments, and not only the claims described below, but also all modifications equivalently or equivalently to the claims described below belong to the scope of the spirit of the present invention. will do it

001: Ultrasonic signal evaluation device of turbine blades using artificial intelligence
100: robotic ultrasound scanner
101: turbine blade blades
102: ultrasonic sensor driving device
103: chain
104: sensor starting point limit sensor
105: sensor endpoint limit sensor
106: encoder
107: ultrasonic sensor

Claims (9)

As an ultrasonic signal evaluation device of a turbine blade using artificial intelligence,
a robot ultrasonic scanner that moves the turbine blade and measures an ultrasonic signal;
a signal acquisition unit for determining the validity of the signal acquired from the robotic ultrasound scanner;
a signal evaluation unit for determining whether the turbine blade is defective by comparing the valid ultrasonic signal determined by the signal acquisition unit with an evaluation reference signal;
a signal storage unit for storing information on the acquired signal, the evaluation reference signal, and the defect; and
and an evaluation criterion learning unit for learning the evaluation criterion signal and the signal for the defect using artificial intelligence;
The robot ultrasound scanner,
moving from the start point to the end point of the turbine blade to obtain a primary signal, and moving from the end point to the start point to obtain a secondary signal,
The signal acquisition unit,
By comparing the reflected position and reflected energy intensity (dB) of the ultrasonic wave from the primary signal and the secondary signal, the ultrasonic signal evaluation apparatus of a turbine blade using artificial intelligence to determine the validity of the acquired signal.
According to claim 1,
The robot ultrasonic scanner is an ultrasonic signal evaluation apparatus of a turbine blade using artificial intelligence to move the root of the turbine blade.
delete delete According to claim 1,
The robot ultrasound scanner,
Ultrasonic sensors arranged in an array form;
a chain for moving the ultrasonic sensor; and
Ultrasonic signal evaluation apparatus of a turbine blade using artificial intelligence, including a; limit sensor attached to the start point and the end point at which the ultrasonic sensor moves in the turbine blade.
According to claim 1,
The difference between the primary signal and the secondary signal is that the difference in the reflected position is within 1%, and the difference in the reflected energy intensity is within 3% of a signal that is determined as a valid signal as a valid signal. of ultrasonic signal evaluation device.
According to claim 1,
The signal evaluation unit,
determining a defect signal including the evaluation reference signal and the ultrasonic reflection signal;
The size of the defect is calculated by the following equation for the evaluation reference signal and the defect signal,

An ultrasonic signal evaluation apparatus of a turbine blade using artificial intelligence to perform; a signal evaluation process for determining whether or not the defect has passed based on a predetermined size.
8. The method of claim 7,
The evaluation standard learning unit,
Turbine using artificial intelligence to store a sound signal that has received a pass decision in the signal storage unit, perform an evaluation criterion learning algorithm to update the evaluation criterion, and reflect the updated evaluation criterion back to the signal evaluation process Ultrasonic signal evaluation device of the blade.
An ultrasonic signal evaluation method of a turbine blade using artificial intelligence, comprising:
an ultrasonic signal acquisition step of measuring an ultrasonic signal by moving the robot ultrasonic scanner along the turbine blade;
a valid signal reading step of receiving the ultrasound signal acquired from the robot ultrasound scanner and determining the validity of the ultrasound signal;
a signal evaluation step of determining whether the turbine blade is defective by comparing the valid ultrasonic signal determined in the signal acquisition step with an evaluation reference signal; and
An evaluation criterion learning step of learning the evaluation criterion signal and the signal for the defect by performing a predetermined artificial intelligence learning algorithm;
The ultrasonic signal acquisition step includes:
The robot ultrasound scanner moves from the start point to the end point of the turbine blade to obtain a primary signal, and moves from the end point to the start point to obtain a secondary signal,
The valid signal reading step includes:
By comparing the reflected position and reflected energy intensity (dB) of the ultrasonic wave from the primary signal and the secondary signal, the ultrasonic signal evaluation method of the turbine blade using artificial intelligence to determine the validity of the acquired signal.

Images