Large format additive manufacturing (LFAM) proved to have a great potential to become an adjacent technology to traditional manufacturing methods. One of the sectors LFAM is targeting is rapid tool/mold development for composites. This includes large mold structures used for high-temperature molding techniques (in-oven or autoclave). Although, these large printed structures (reaching hundreds of pounds) develop thermal-residual stress during cool-down and can eventually crack, turning the structure into waste. Acoustic emission (AE), a passive non-intrusive global nondestructive evaluation (NDE) technique, was used to monitor crack growth and can provide the right tools that can be used for feedback loop for corrective action. This research performs thermal testing on a large AM mold with preexisting cracks, in an attempt to monitor crack growth using AE. AE was able to detect, identify and locate the crack source by means of acoustic features, waveform characteristics, spectrum analysis, and difference in arrival times.
In the last two decades, nonlinear ultrasonic testing is getting more attention due to their sensitivity to microcracks among a variety of NDT techniques used in infrastructure. Vibro-Acoustic Modulation (VAM) technique is one of the practical methods, that does not need the expensive hardware components required for the conventional nonlinear methods. This method is capable of identifying damage growth using the correlation of the level of nonlinearity to the severity or density of the damage. To be able to determine the sensitivity of VAM technique in comparison with other conventional nondestructive testing methods, Acoustic Emission (AE) technique as a global method and Ultrasonic Testing (UT) and Eddy current Testing (ET) techniques as local methods are investigated in an identical testing condition for similar specimens. The comparison has been conducted by testing a typical steel material used in the steel bridges under cyclic tension load. All these methods have some features in common and some differences. A comprehensive comparison study of these techniques sheds light on their practicality for various applications. Unlike the AE technique that listens to the structure for the received signal of the released elastic energy from the defects, VAM introduces the signals to the specimen and monitors the signal that was modulated by the vibration to get information about the crack. VAM and AE have some similarities such as no need for positioning sensors on the cracks and capability of detecting the crack in the early stages. On the other hand, local techniques such as UT and ET are more accurate than the VAM technique in terms of localization but less sensitive in terms of how soon they detect cracks.
Ultrasonic Phased Array imaging is a key method for fast and reliable nondestructive testing of structures, especially when only one side of the part is accessible. Full matrix capturing (FMC) in combination with the total focusing method (TFM) provides a strong tool for ultrasonic imaging of structures with complex flaw patterns. However, still, operator needs to go through the generated images and manually check for the possible defects. One important task is to separate true and false indications, as some of them are noises or artifacts. Inspecting large structures with TFM Phased Array Imaging generates a huge amount of data which takes a significant time to go through them manually. In this work, we evaluate the possibility of using the neural network as an artificial intelligent toolbox to identify the defects. Using finite element method and an in-house developed TFM code, the phased array images are produced as the input to the neural network. The output of the neural network, target, is defined as the probability of defect existence. After generating TFM final images with different flaw patterns, the network was trained and evaluated based on the stochastic genetic algorithm learning method. This made the training feasible with limited provided data. Results indicate the great potential of machine learning for automatic or assisted defect recognition. The main challenge to pursuing a comprehensive and reliable machine learning toolbox, is to train the system with a satisfactory number of examples in different situations to ensure the final product is able to cover all possible conditions. It is concluded the proposed neural network model is capable of image pattern recognition with limited provided training data.
KEYWORDS: Sensors, Signal detection, Neural networks, Acoustic emission, Algorithm development, Acoustics, Machine learning, Data modeling, Data analysis, Signal generators
Boiler tubes in power plants develop defects including creep and thermal fatigue damage that can lead to fluid leakage over the operation period. Such leakage is the main cause of outages and power generation losses in thermal power plants. Therefore, early detection of leaks in boiler tubes is necessary to avoid more than 60% of boiler outages. To monitor and detect tube leaks in real-time, Acoustic Emission (AE) technique is widely used in power plants. A boiler tube leak could be detected using Average Signal Level (ASL) of the acquired AE signal using a network of sensors attached to the body of the boiler. Changes in ASL are proportional to the tube leakage; however, background signals generated by operating soot blowers bury the features which represent the tube leaks in the boiler and makes it nearly impossible to detect them automatically with established threshold methods. Soot blowers are used to remove the soot that is deposited on the tubes to maintain the efficacy and continuous operation of boilers. In this study, a bidirectional long short-term memory (LSTM) recurrent neural network (RNN) is developed to automatically detect tube leaks in power plant boilers. This detection method aims at identifying abnormal acoustic signals which differ from the reference/normal data that the system was trained with. The neural networks are trained on a sample boiler and the evaluation was done on the same boiler on the intervals with leak presence. Once the developed machine learning algorithm was tested with AE signals acquired from boiler tubes, the results show that this novel approach can detect anomalies in the signal levels as an indication of tube defects with an acceptable accuracy.
Quality assurance and structural integrity evaluation are the crucial parts of the successful design and service of additively manufactured (AM) components. Discontinuities and flaws in AM parts can affect the mechanical properties of a component during manufacturing and service. It is very important to identify the discontinuities in AM parts in terms of location, size, and geometrical properties using nondestructive testing (NDT) techniques. Existing research in both mechanical testing and nondestructive evaluation involves developing methods for characterizing and inspecting AM components as the use of such materials continues to rise. Although there exist relatively mature ultrasonic inspection techniques for defect detection, AM polymer components face the challenge of considerable internal inhomogeneities caused by the design and printing strategies. It has been shown that the ultrasonic signals are very sensitive to the material inhomogeneities, consequently the reflection/diffractions from the defects will be significantly influenced and defect detection will be very challenging. This work aims to present the potentials and challenges in ultrasonic detection of defects in polymer AM parts. Air-coupled ultrasonic tests to be demonstrated and followed by results and discussions. The role of porosity on detectability in the ultrasonic NDT tests is described and a possible way for attenuation assessment is demonstrated. Finally, the effect of AM part inhomogeneities on detection probability of seeded defects with different sizes and locations in AM parts is presented.
Acoustic tomography method facilitates mapping internal defects in real-time and in-situ without destructive testing. The method requires certain number of transmitter and receiver paths to reconstruct the slowness map of scanned area depending upon the target resolution. Once the hardware component is determined, the major software output to feed into the algorithm is the time of flight. There are sophisticated signal processing methods reported in literature to determine the time of flight (TOF) with better accuracy as compared to conventional threshold-based method. The most common approaches are wavelet-based or energy-based methods, which require transforming time history signal into different domains. Domain transformation is typically applied in laboratory-scale experiments. In this paper, a new arrival time pick-up approach based on defining outliers in the derivative of transient signal in time domain is evaluated in terms of accuracy, computational effort and power as compared to threshold-based and wavelet/energy-based methods reported in literature. The waveforms from experiments is used to study the influence of materials and signal-to-noise ratio on the accuracy of detecting the fastest wave mode. In addition, waveforms are also artificially generated with fixed wave velocity using numerical models to further evaluate the performance of the different methods (outlier-based, threshold-based and energy-based). The influence of tomography quality by using these to this method performs better in accuracy and efficiency.
Comb-drive transducers are made of interdigitized fingers formed by the stationary part known as stator and the moving part known as rotor, and based on the transduction principle of capacitance change. They can be designed as area-change or gap-change mechanism to convert the mechanical signal at in-plane direction into electrical output. The comb-drive transducers can be utilized to differentiate the wave motion in orthogonal directions when they are utilized with the outof- plane transducers. However, their sensitivity is weak to detect the wave motion released by newly formed damage surfaces. In this study, Micro-Electro-Mechanical System (MEMS) comb-drive Acoustic Emission (AE) transducer designs with two different mechanisms are designed, characterized and compared for sensing high frequency wave propagation. The MEMS AE transducers are manufactured using MetalMUMPs (Metal Multi-User MEMS Processes), which use electroplating technique for highly elevated microstructure geometries. Each type of the transducers is numerically modeled using COMSOL Multiphysics program in order to determine the sensitivity based on the applied load. The transducers are experimentally characterized and compared to the numerical models. The experiments include laser excitation to control the direction of the wave generation, and actual crack growth monitoring of aluminum 7075 specimens loaded under fatigue. Behavior and responses of the transducers are compared based on the parameters such as waveform signature, peak frequency, damping, sensitivity, and signal to noise ratio. The comparisons between the measured parameters are scaled according to the respective capacitance of each sensor in order to determine the most sensitive design geometry.
In this paper, new MEMS strain sensors are introduced. The transduction principle of the sensors is the resistance change due to piezoresistive property of polysilicon. Five different sensors are designed on the same device and tuned to resistance values of 350 Ω and 120 Ω. The sensors are aligned in horizontal, vertical and 45° directions in order to extract the principle strains. The geometry of the sensing element is a rectangular bar anchored at two ends and suspended above silicon substrate. The sensors are numerically modeled using COMSOL Multiphysics software. The model consists of all the micromachining layers, including silicon substrate, 0.7 μm thick polysilicon layer (sensing element) sandwiched between two layers of 0.35 μm thick silicon nitride layers and trenching under polysilicon layer, in order to estimate the strain that piezoresistive element is exposed to. The MEMS strain sensors are manufactured using MetalMUMPs process. The sensors are attached to aluminum and steel plates, and their gauge factors are compared with conventional foil gauges under uniaxial and biaxial loading. It is demonstrated that the MEMS strain sensors can detect both static and dynamic strains with the gauge factor reaching significantly high values. High gauge factor occurs because of unique geometry design and trenching, which amplify the strain that the polysilicon layer senses. The MEMS strain sensor can be fused with other sensing elements on the same device such as accelerometer, acoustic emission in order to have redundant measurement from a single point.
In this paper, new MEMS Acoustic Emission (AE) sensors are introduced. The transduction principle of the sensors is capacitance due to gap change. The sensors are numerically modeled using COMSOL Multiphysics software in order to estimate the resonant frequencies and capacitance values, and manufactured using MetalMUMPS process. The process includes thick metal layer (20 μm) made of nickel for freely vibration layer and polysilicon layer as the stationary layer. The metal layer provides a relatively heavy mass so that the spring constant can be designed high for low frequency sensor designs in order to increase the collapse voltage level (proportional to the stiffness), which increases the sensor sensitivity. An insulator layer is deposited between stationary layer and freely vibration layer, which significantly reduces the potential of stiction as a failure mode. As conventional AE sensors made of piezoelectric materials cannot be designed for low frequencies (<300 kHz) with miniature size, the MEMS sensor frequencies are tuned to 50 kHz and 200 kHz. The each sensor contained several parallel-connected cells with an overall size of approximately 250μm × 500 μm. The electromechanical characterizations are performed using high precision impedance analyzer and compared with the numerical results, which indicate a good fit. The initial mechanical characterization tests in atmospheric pressure are conducted using pencil lead break simulations. The proper sensor design reduces the squeeze film damping so that it does not require any vacuum packaging. The MEMS sensor responses are compared with similar frequency piezoelectric AE sensors.
Silicon has piezoresistive property that allows designing strain sensor with higher gauge factor compared to conventional
metal foil gauges. The sensing element can be micro-scale using MEMS, which minimizes the effect of strain gradient
on measurement at stress concentration regions such as crack tips. The challenge of MEMS based strain sensor design is
to decouple the sensing element from substrate for true strain measurement and to compensate the temperature effect on
the piezoresistive coefficients of silicon. In this paper, a family of MEMS strain sensors with different geometric designs
is introduced. Each strain sensor is made of single crystal silicon and manufactured using deposition/ etching/oxidation
steps on a n- doped silicon wafer in (100) plane. The geometries include sensing element connected to the free heads of
U shape substrate, a set of two or more sensing elements in an array in order to capture strain gradients and two
directional sensors. The response function and the gauge factor of the strain sensors are identified using multi-physics
models that combine structural and electrical behaviors of sensors mounted on a strained structure. The relationship
between surface strain and strain at microstructure is identified numerically in order to include the relationship in the
response function calculation.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.