Student Projects
Contact Information
University: Tokyo Institute of Technology (Japan)
Team Member(s): Koichi Mizukami (2015)
Faculty Adviser(s): Yoshihiro Mizutani, Akira Todoroki, Yoshiro Suzuki
Email Address: kmizukami206@gmail.com
Submission Language: English
Project Information
Title: Damage Detection in CFRP Aircrafts Using Eddy Current Testing and Statistical Diagnosis
Description:
The purpose of this project is to develop a new method of in-service inspection on CFRP (Carbon Fiber Reinforced Plastic) components in aircrafts by using eddy current testing. Multichannel eddy current sensor damage detection system was composed by LabVIEW, NI PXI hardware and multichannel oscilloscope to identify the location of defective zone in the CFRP structure.
Products:
NI products - LabVIEW
- NI PXI-5421 16-Bit 100MS/s Arbitrary Waveform Generator
- NI PXI-5122 14-Bit 100MS/s Digitizer
Additional hardware - Picoscope4424 4-Channel USB Oscilloscope
- NF HSA4101 High Speed Bipolar Amplifier
The Challenge
Background
Carbon Fiber Reinforced Plastic (CFRP) is of a great interest in aerospace industries because of its excellent mechanical properties such as high specific strength and stiffness and potential of weight-saving. However, since CFRP has laminated structure, CFRP structures easily get damaged by external impact load and delamination is caused. Example image of delamination is shown in Figure 1. It is reported that delamination in CFRP greatly decreases the strength of the structure. Although detection of delamination is of great importance, it is technically challenging to detect delamination in CFRP because this damage is not visible from outside of the structure. Testing method currently adopted for detailed inspection on CFRP aircraft is mainly ultrasonic testing. However, detailed ultrasonic testing is conducted only once every four years because disassembly of aircraft is required for inspection and it takes huge time cost. Thus, rapid detection of delamination for in-service CFRP aircraft is in great demand.
.png)
.png)
Fig. 1 Example image of delamination.
Proposal
1. Eddy current testing
We proposed a new flexible eddy current testing sheet attached on the inspected CFRP structure to monitor its structural integrity. The schematic illustration of the proposed testing sheet and the concept for detecting delamination are shown in Figure 2 and Figure 3 respectively.
Proposed testing sheet consists of driver coil and pickup coils as shown in Figure 2. By applying high-frequency AC voltage (1MHz~) to the driver coil, eddy current is induced in the proximity to the surface of the CFRP under test because carbon fiber is electrically conductive. Pickup coils act as magnetic sensors and they output voltage depending on the intensity of magnetic field caused by eddy current. Since intensity of magnetic field changes according to the conductivity of the material under test, conductivity monitoring can be conducted by eddy current testing.
In our proposal, temperature difference is caused between delamination zone and intact zone, and the temperature difference is detected by eddy current testing. Our proposal consists of two stages. First stage is high-frequency induction heating and second stage is eddy current testing. In the first stage, high AC voltage is applied to the driver coil and surface of the CFRP is heated by induced eddy current following the Joule’s law. From the viewpoint of heat transfer, delamination acts as an additional thermal resistance through the thickness of the CFRP under test due to the trapped air, which indicates that delamination produces hot spot on the top external surface of the CFRP as shown in Figure 3. In the second stage, hot spot is detected by eddy current testing. Since electrical conductivity of carbon fiber increases with temperature, hot spot becomes a high electrical conductivity spot. At high electrical conductivity spot, intensity of eddy current is stronger. Since magnetic field caused by eddy current also becomes stronger at high electrical conductivity spot, output signal of the pickup coil on the delamination is different from intact zones. Like this, eddy current thermo-sensing is possible by utilizing temperature dependency of electrical conductivity. Since induction heating (first stage) and eddy current testing (second stage) can be conducted simultaneously, delamination can be detected by comparing the output signal of pickup coils during CFRP is heated by eddy current. It is expected that proposed in-service inspection system can dramatically decrease the cost on detecting delamination and contribute to the improvement of the safety of CFRP aircrafts.
Fig. 2 Schematic illustration of the proposed testing sheet; planar driver coil and pickup coils are attached on the CFRP under test.
Fig.3 Concept for detecting delamination; delamiation produces hot spot during induction heating, and the hot spot is detected by eddy current thermo-sensing.
2. Statistical diagnosis
To clearly judge whether CFRP structure under test is intact or defective, we introduced statistical diagnosis called SI-F method (System Identification-F method). SI-F method is composed by three stages. First stage is training process. In the training process, relationship of pickup coil outputs is investigated when the CFRP structure is not damaged. Second stage is diagnosis process. In diagnosis process, relationship of pickup coil outputs is investigated in the same way as training process when the CFRP structure is in service. Third stage is judgment process. Relationship of pickup coil outputs obtained in diagnosis process is compared with that of training process. If the CFRP structure under test is not damaged, relationship of pickup coils in training process and diagnosis process are almost equivalent. On the other hand, if the CFRP structure is defective, relationship of pickup coils in training process and diagnosis process are not equivalent. Comparison of relationship of pickup coils is conducted by F test. F test is one of the methods of statistical test used to examine the equivalency of sensor responses. If the CFRP structure is intact, it is known that statistic F follows F distribution and average F has a value close to about 0.9~1.2. However, if the CFRP structure is defective, statistic F does not follow F distribution and average F has larger value than 1.2. Thus, by calculating average F from repeated calculation of F, we can judge whether the CFRP structure under test is damaged or not.
Health monitoring of CFRP structures by using statistical diagnosis is benefitial and innovative for two reasons. Firstly, statistical diagnosis can automatically judge whether the CFRP structure under test is defective or not. This means that subjective judgement of experienced inspectors of eddy current test will be less required. Secondly, statistical diagnosis does not require destructive tests of CFRP specimens for obtaining database of output signal at defective zone. That is because statistical diagnosis focuses on change of relationship of pickup coil output and require only data of initial state. For those reasons, it is expected that statistical diagnosis can greatly contribute to reducing cost on inspection.
The Solution
Experiment
To prove the feasibility of the proposal, experiment to detect delamination in the CFRP plate was conducted. Eddy current damage detection system is shown in Figure 4, and the monitoring part fabricated in the front panel of LabVIEW is shown in Figure 5. Driver coil and three pickup coils are attached on the CFRP plate with a delamination. 2.4Vpp, 2.0MHz AC voltage supplied from NI PXI-5421 arbitrary waveform generator is amplified by NF HAS 4101 high speed bipolar amplifier, then 126Vpp, 2.0MHz AC voltage is applied to the driver coil. Eddy current is induced in CFRP by the driver coil and its surface is heated. During the induction heating, output voltage of pickup coils are measured by Picoscope 4424 4channel USB Oscilloscope at measurement points A, B, C in Figure 4. A and C are intact zones, and B is on delamination zone. For eddy current thermo-sensing, output voltage and transimpedance of each pickup coil have to be monitored. Output voltage of pickup coil is a parameter which represents the intensity of magnetic field. Transimpedance of pickup coil is defined as output voltage divided by electric current applied to driver coil, and it is reported that transimpedance should be measured to examine the influence of the temperature change of the testing sheet itself. Data from Picoscope Oscilloscope can be read by LabVIEW and monitoring of pickup coil output voltage and calculation of transimpedance are conducted by LabVIEW program as shown in Figure 5.
Fig. 4 Schematic illustration of the eddy current damage detection system; high voltage is applied to the driver coil and the output signal of pickup coils are monitored by the multichannel oscilloscope whose data is read in LabVIEW program.
Fig. 5 Monitoring part of eddy current damage detection system; measured output signal are displayed and data acquisition settings can be conducted in this part.
Result
Thermal image of the CFRP under test and the change of output voltage of pickup coils 90s after induction heating was started are shown in Figure 6. The thermal image in Figure 6(a) obtained from thermography shows that delamination at measurement point B produces a hot spot on the top external surface of the CFRP plate. As shown in Figure 6(b), change of output voltage of the pickup coil at B is higher than A and C. That is because surface temperature of delamination zone rises faster than intact zone and increase of electrical conductivity at B is higher than A and C. Those results indicate that induction heating can cause the difference of temperature between intact zone and delamination zone, and the delamination zone is detected by eddy current thermo-sensing.
Figure 7 shows the result of statistical diagnosis for both intact specimen and defective specimen. Average of statistic F is calculated and average F should be 0.923~1.2 for the specimen to be judged intact. Moreover, LEDs are assigned to each measurement point in the front panel and LEDs of measurement point with no defect is green. In case of defective zone, LED is red and we can know the location of the defective zone. As Figure 7(a) shows, in case of the intact specimen, average F is 1.08 and the specimen is judged intact. On the other hand, in case of defective specimen, average F is 1.44 and the specimen is judged defective as shown in Figure 7(b). LED of defective zone B is red and we can know measurement point B is defective zone.
Therefore, it is proven that location of defect in CFRP can be identified by using multichannel eddy current testing and statistical diagnosis.
(a) (b)
Fig. 6 Result of the experiment. (a)Thermal image of the CFRP under test obtained from thermography. (b)Comparison of change of output voltage of pickup coil 90s after induction heating is started.
.png)
(a)
.png)
(b)
Fig. 7 Result of the statistical diagnosis. (a)Diagnosis for the intact specimen. (b)Diagnosis for the defective specimen with delamination at B.
Benefit of using LabVIEW and NI tools
LabVIEW and NI tools mainly have three benefits in our project.
First benefit is high frequency waveform generator obtained by NI PXI-5421. For induction heating, AC voltage at 100 ~ 400kHz is enough in case of metal. However, AC voltage at 1.0MHz ~ 10MHz is pre-requisite for the induction heating of CFRP. Since NI PXI-5421can supply AC frequency up to 43.0MHz, induction heating of CFRP became possible in our project. Moreover, although AC frequency must be changed depending on the thickness of CFRP to guarantee appropriate eddy current distribution, test condition can easily be changed in NI PXI-5421.
Second benefit is monitoring system obtained by LabVIEW. In the monitoring part of Figure 6, output signal of each pickup coil is monitored and the chart can enable us to know how the output signal changes during eddy current thermo-sensing. More importantly, system fabrication by LabVIEW is beneficial in terms of cost. To fabricate this eddy current damage detection system by using conventional products, we need multichannel oscilloscope with recorder, display and signal processer device which can modify test condition. On the other hand, cost to fabricate this system by LabVIEW is less than 40% of conventional product (calculated from the price of conventional eddy current electrical conductivity meter and oscilloscope). Furthermore, because of easy system integration in LabVIEW, monitoring part of our project is completed only in two weeks and it provides us more chances to study physical aspects and improve our project.
Third benefit is easy fabrication of statistical diagnosis part of Figure 7. Although SI-F method is based on complex calculations of matrix, Beta function and probability density analysis, the system for diagnosis can easily be fabricated by using functions precomposed in LabVIEW. Surprisingly, statistical diagnosis system was fabricated within a week.
Conclusion
System for damage detection in CFRP structures using eddy current testing and statistical diagnosis was proposed. The damage detection system was fabricated on LabVIEW and NI hardware. Through the experiment on CFRP specimen, feasibility of the proposal was proven. More importantly, because of easy system fabrication on LabVIEW, our project team could investigate the feasibility of our study in a short term and have enough time to focus more on physical aspect.
