Showing papers on "Ultrasonic testing published in 2019"

Internal damage evaluation of composite structures using phased array ultrasonic technique: Impact damage assessment in CFRP and 3D printed reinforced composites

Abstract: The application of non-destructive techniques for the evaluation of internal damage in composite materials is a challenge due to their complexity. The aim of this study is to analyse the ability of ultrasonic technique to identify and evaluate manufacturing defects and internal damage in composite laminates. Artificial object inclusions of different shapes, sizes and materials were embedded in carbon fibre reinforced epoxy (CFRP) laminates. Comprehensive investigation of non-destructive evaluation using phased array ultrasonic testing to trace and characterize the embedded defects is presented. Phased array ultrasonic technique was able to precisely locate most of the artificial inclusion in the composite laminates. Nerveless, the shape and size of the inclusions were not accurately determined due to the high signal attenuation and distortion characteristics of the carbon fibre epoxy composite. Furthermore, a major concern affecting the efficient use of composite laminates is their vulnerability to low velocity impact damage. The dynamic behaviour of composite laminates is very complex as there are many concurrent phenomena during composite laminate failure under impact loading. Thus, the practicality of the previous results is demonstrated by the evaluation of impacted carbon fibre reinforced epoxy laminates using ultrasonic testing. The influence of laminate thickness and impact energy on impact damage is also investigated. The performance of phased array ultrasonic testing for the inspection of composite laminates with barely visible impact damage is evaluated. In addition, fibre-reinforced thermoplastic composites are becoming more significant in industrial applications due to their excellent mechanical performance and potential recycling. In this study, impact damage in 3D printed continuous glass fibre reinforced thermoplastic laminates is also analysed. Phased array ultrasonic testing is performed and the amount of damage is quantified in terms of delaminated area.

Nondestructive Ultrasonic Inspection of Composite Materials: A Comparative Advantage of Phased Array Ultrasonic

Abstract: Carbon- and glass fiber-reinforced polymer (CFRP and GFRP) composite materials have been used in many industries such as aerospace and automobile because of their outstanding strength-to-weight ratio and corrosion resistance. The quality of these materials is important for safe operation. Nondestructive testing (NDT) techniques are an effective way to inspect these composites. While ultrasonic NDT has previously been used for inspection of composites, conventional ultrasonic NDT, using single element transducers, has limitations such as high attenuation and low signal-to-noise ratio (SNR). Using phased array ultrasonic testing (PAUT) techniques, signals can be generated at desired distances and angles. These capabilities provide promising results for composites where the anisotropic structure makes signal evaluation challenging. Defect detection in composites based on bulk and guided waves are studied. The capability of the PAUT and its sensitivity to flaws were evaluated by comparing the signal characteristics to the conventional method. The results show that flaw sizes as small as 0.8 mm with penetration depth up to 25 mm can be detected using PAUT, and the result signals have better characteristics than the conventional ultrasonic technique. In addition, it has been shown that guided wave generated by PAUT also has outstanding capability of flaw detection in composite materials.

Laser ultrasonic inspection of additive manufactured components

Abstract: Additively manufactured components are gaining popularity in aerospace, automotive and medical engineering applications. Additive manufacturing (AM) offers tremendous cost advantages over traditional manufacturing methods. However, inter- and intra-layer defects are observed in AM components. Moreover, the lack of appropriate testing methods for assessing the integrity of AM components deters its use, despite the several functional advantages it has to offer. Non-destructive testing (NDT) forms the most common and convenient way of inspecting parts. In this paper, a laser ultrasonic technique for the inspection of AM components is proposed. The results demonstrate laser ultrasonic testing (LUT) as a promising method for the non-contact inspection of additive manufactured components. Furthermore, the results were validated using X-ray computed tomography (CT) and ultrasonic immersion testing (UIT). The sample used in this study was manufactured through selective laser melting (SLM) AM process with built-in holes representing defects.

Ultrasonic phased array inspection of a Wire + Arc Additive Manufactured (WAAM) sample with intentionally embedded defects

Abstract: In this study, Wire + Arc Additive Manufacture (WAAM) was employed to manufacture a steel specimen with intentionally embedded defects which were subsequently used for calibration of an ultrasonic phased array system and defect sizing. An ABB robot was combined with the Cold Metal Transfer (CMT) Gas Metal Arc (GMA) process to deposit 20 layers of mild steel. Tungsten-carbide balls (o1-3 mm) were intentionally embedded inside the additive structure after the 4th, 8th, 12th and 18th layers to serve as ultrasonic reflectors, simulating defects within the WAAM sample. An ultrasonic phased array system, consisting of a 5 MHz 64 Element phased array transducer, was used to inspect the WAAM sample non-destructively. The majority of the reflectors were detected successfully using Total Focusing Method (TFM), proving that the tungsten carbide balls were successfully embedded during the WAAM process and also that these are good ultrasonic reflectors. Owing to a lack of standards and codes for the ultrasonic inspection of WAAM samples (Lopez et al., 2018), a calibration method and step-by-step inspection strategy were introduced and then used to estimate the size and shape of an unknown lack of fusion (LoF) indication. This was then validated by destructive analysis, showing a good correlation with the phased array results.

A novel pulsed eddy current method for high-speed pipeline inline inspection

Abstract: High-speed inspection of inner diameter (ID) and outer diameter (OD) surface defects on thick-wall steel pipes is an important aspect to advance the pipeline inline inspection (ILI) in the oil and gas industry. The state-of-the-art ILI methods including magnetic flux leakage (MFL), ultrasonic testing (UT), electromagnetic acoustic transducer (EMAT), eddy current (EC) and pulsed eddy current (PEC) are hardly applicable to practical high-speed pipeline ILI due to the reasons of long inspection time required for sensor response and low detection sensitivity caused by severe motion. New sensing techniques that offer faster inspection speed, deeper signal penetration depth, better detection sensitivity and linearity, as well as capability of ID/OD discrimination are imperatively needed. This paper proposes a novel PEC sensing method to detect and discriminate ID/OD defects by utilizing the conductivity-dependent and permeability-dependent distribution patterns of induced eddy current at the ID surface of steel pipes. For ID defects, the pattern is caused by the discontinuous conductivity distribution, while for OD defects, the pattern is caused by the non-uniform incremental permeability distribution. A pulse-excited current with short width (2.5 μs), low duty cycle (1%) and fast-falling edge (100 ns) is injected into an excitation coil, so that the secondary magnetic field at the fast-falling edge will produce a transient oscillation in a pair of differential pick-up coils. Then, the time-domain transient oscillatory pick-up signal is extended, filtered, amplified, and extracted to be one feature by an envelope detector and an average sample method, which is processed in real-time by the developed probe for facilitating the back-end data recording. Meanwhile, a novel high-speed pipeline inspection gauge (PIG) with a sensor array is developed for field testing to validate the effectiveness of the proposed PEC method that achieved high inspection speed, deep detection depth, superior sensitivity, good linearity, low power consumption, easy implementation, ID/OD discrimination and crack detection capability.

Reliability assessment of pulsed thermography and ultrasonic testing for impact damage of CFRP panels

Abstract: In order to quantitatively compare the reliability of pulsed thermography and ultrasonic testing techniques, a set of thirty-five Carbon Fiber Reinforced Plastic (CFRP) composite panels with impact damages are inspected by pulsed thermography and ultrasonic C-scan. The comparative experimental results and Probability of Detection (PoD) analysis results are presented. The quantitative comparison shows that pulsed thermography testing has smaller defect size at 90% PoD with 95% confidence level, i.e. a90/95 values than ultrasonic testing for the parameters and setup used in the inspections of these thirty-five CFRP composite panels.

Phased Array Ultrasonic Inspection of Metal Additive Manufacturing Parts

Abstract: The adoption of wire and arc additive manufacturing (WAAM) in the market has been retained by the need to find a suitable method to ensure the quality of the parts produced. WAAM processes build up parts through the deposition of weld beads, consequently components with rough finish surfaces are characteristics of the method. Non-destructive testing (NDT) by ultrasonic (UT) method, namely the phased array technique (PAUT), is usually used to detect these defects in welding. However, the roughness of the parts represents a challenge for the UT application, since these variations influence the interaction between the emitted UT beam and the component. This study is thus focused on assessing the capability of detecting WAAM defects. The effectiveness of phased array ultrasonic testing to detect defects in aluminium WAAM components with several degrees of surface finish was evaluated. Simulations were first performed with CIVA software to characterize the beam emitted and select the probes and inspection parameters. Afterwards, physical inspections were performed on three reference specimens. Experimental outcomes prove that PAUT techniques are suitable for WAAM defects detection, including sizing, morphology and location. In addition, the experimental results were consistent with the simulated ones. The probes were able to overcome the limitations caused by the surface roughness of the samples, for a maximum of 89.6 μm average waviness profile. Also, defects ranging from 2 to 5 mm were characterized, in size and depth. These preliminary results represent an essential step for the development of an NDT system for inspecting WAAM parts.

Prediction of Concrete Strength with P-, S-, R-Wave Velocities by Support Vector Machine (SVM) and Artificial Neural Network (ANN)

Abstract: Mechanical waves, such as ultrasonic waves, have shown promise for use in non-destructive methods used in the evaluation of concrete properties, such as strength and elasticity. However, accurate estimation of the concrete compressive strength is difficult if only the pressure waves (P-waves) are considered, which is common in non-destructive methods. P-waves cannot reflect various factors such as the types of aggregates and cement, the fine aggregate modulus, and the interfacial transition zone, influencing the concrete strength. In this study, shear waves (S-waves) and Rayleigh waves (R-waves) were additionally used to obtain a more accurate prediction of the concrete strength. The velocities of three types of mechanical waves were measured by recent ultrasonic testing methods. Two machine learning models—a support vector machine (SVM) and an artificial neural network (ANN)—were developed within the MATLAB programming environment. Both models were successfully used to model the relationship between the mechanical wave velocities and the concrete compressive strength. The machine learning model that included the P-, S-, and R-wave velocities was more accurate than the model that included only the P-wave velocity.

Quantitative detection of lamination defect in thin-walled metallic pipe by using circumferential Lamb waves based on wavenumber analysis method

Abstract: Lamination defect is one of the common defects in the manufacturing process of seamless pipes. In this paper, the quantitative detection of a lamination defect in thin-walled metallic pipe using circumferential Lamb waves is studied. To interpret the received time-domain signals and extract useful information about the lamination defect, wavenumber analysis is performed on these signals. A three-dimensional finite-element model is established using Matlab and ABAQUS commercial software. Owing to the processing technique, an aluminum ring structure with a three-quarters circumference is considered to represent the metallic pipe. The lamination defect constructed in the model is a “zero-volume” crack, which stretches from θ = 180° to 270° and locates in the mid-plane of the wall. A five-cycle 0.41-MHz sinusoidal tone-burst signal modulated by a Hanning window is carefully chosen to generate the appropriate excitation wave, in CL0 mode. According to the received signals, the conclusion that the incident CL0 mode interacts with the lamination defect for numbers of times can be obtained. The space-amplitude curve of incident waves is also depicted to reveal the amplitude distribution of incident waves. A fully non-contact experimental platform that adopts an electromagnetic acoustic transducer as a transmitter and a laser ultrasonic inspection system as a receiver is set up to verify the finite-element model. Three different wavenumber analysis methods are performed on the wavefield signals to explain the detectability of the lamination defect separately through both numerical and experimental studies. It can be concluded that, from the variation of wavenumber, the continuity of structure can be deduced. Not only can the location be calculated with an error of less than 10%, but the profile of the lamination defect can also be imaged. It is also found that very good consistency exists between numerical and experimental results.

Laser ultrasonics inspection for defect evaluation on train wheel

Abstract: Passengers’ safety and in-service life of wheelset axles play an important role in railway vehicles. For this reason, periodic inspections are necessary. Among non-destructive techniques, ultrasonic ones are widely applied in this field. The main disadvantage of conventional ultrasonic techniques is that the overall inspection of wheels requires the train to be put out-of-service and disassembly each part, which is time-consuming and expensive. In this paper, a non-conventional non-contact laser ultrasonic inspection for train wheels is proposed. The proposed method uses a laser interferometer to receive the ultrasonic wave without contact. The receiving system allows choosing the distance between the surface to be inspected and the interferometer, overcoming any encumbrance issue. The experimental investigation is carried out on standard-reproduced defects in order to evaluate the reliability and the accuracy of the technique and to verify its applicability for railway components, as wheels, which have a complex geometry. The experimental setup consists of a pulsed laser for the ultrasonic wave generation. The receiving unit combines a continuous-wave laser and an interferometer in order to acquire the surface out-of-plane displacements. Surface and internal standard defects are detected by collecting all the A-scans in a B-scan map. The results are promising for the application of the laser technique to detect both surface and internal defects on in-service components.

Real-time monitoring of welding process using air-coupled ultrasonics and acoustic emission

Abstract: The automated weld quality assurance can improve efficiency and productivity. This paper presents the development of real-time weld quality assurance approach for gas tungsten arc welding (GTAW) using acoustic emission (AE) and air-coupled ultrasonic testing (UT). The major weld defect of interest in this paper is burn through, that is, melting through the base metal during welding that creates a hole/gap. The in situ monitoring system evaluates the changes in weld size leading to burn through by changing the weld heat input. Different categories of burn through are defined that include melting of the back of the plate without any molten metal exiting to formation of a hole in the plate. It is demonstrated that complete air-coupled UT cannot be used simultaneously with welding due to the influence of the magnetic field that develops in the weld torch during welding, which weakens the ultrasonic signal. Consequently, a rolling UT transmitter is combined with air-coupled UT receiver to increase the signal/noise value. Wave dispersion is detected due to the different levels of burn through. While UT method provides quantitative information about the weld state, any localized surface discontinuity causes sudden surges in the AE energy indicating non-uniform welding qualitatively. It is concluded that passive and active nondestructive evaluation methods should be combined to monitor weld quality real time for qualitative and quantitative assessment.

Design and optimisation of a phased array transducer for ultrasonic inspection of large forged steel ingots

Abstract: Large steel forgings are used in numerous industries. This paper investigates the possibility of adapting ultrasonic phased array testing to the inspection of forged steel blocks with dimensions of up to 1000 mm in every direction. The paper looks at two objectives in order to provide the best inspection performance: (1) the ultrasonic phased array probe optimisation and (2) the ultrasonic wave transmission sequence in a total focusing method imaging scenario. The CIVA software suite was used to optimise the phased array probe element count and width in a full matrix capture configuration. Based on the simulation results, a 32-element transducer was then built and tested in a 777 mm forged steel block using full matrix capture, plane wave and Hadamard transmission sequences. It was observed that plane wave and Hadamard sequences transmit significantly more energy inside the test sample because the elements are emitting simultaneously, therefore leading to an improved signal-to-noise ratio. However, the horizontal resolution was a strong limitation for every transmission sequence, especially for plane waves, because of the limited range of angles available in a block of large dimensions. The Hadamard transmission sequence was shown to represent the best compromise in terms of defect reflected amplitude, signal-to-noise ratio and resolution. The experimental results were compared with simulations and were found to be in very good agreement.

Effect of ultrasonic shot peening on the surface defects of thin struts built by electron beam melting: Consequences on fatigue resistance

Abstract: Ti-6Al-4V thin struts, built by selective electron beam melting have been submitted to hot isostatic pressing and further processed with ultrasonic shot peening to investigate the effect of this surface treatment on surface defects. The consequence of those surface defects on the fatigue resistance of thin struts has been evaluated before and after treatment. X-ray microtomography has been used to track the defect population with repeated scans before and after ultrasonic shot peening as well as after fatigue tests conducted to failure. Our results show that the fatigue resistance of surface-treated struts (measured at 105 cycles) is doubled compared to the fatigue resistance of struts with an as-built surface. This enhancement in fatigue resistance is attributed to a refinement of the sub-surface microstructure, to the introduction of local compressive residual stresses and to a significant surface smoothing effect induced by ultrasonic shot peening. Ultrasonic shot peening fails to “heal” the most tortuous surface defects observed in the as-built condition, leaving residual defects in the sub-surface region. The presence of those residual surface defects limits the fatigue resistance when compared to struts with a machined surface. © 2019 Elsevier B.V.

Non-destructive Damage Detection in Fibre Metal Laminates

Abstract: The work explores the possibilities of application of ultrasonic testing in the assessment of fibre metal laminates. Basic problems concerning the use of ultrasonic methods in the research of laminates are explained, and methods for solving these problems are suggested. Tests were conducted using two phased array methods: ultrasonic pulse-echo and through transmission. The efficiency of both selected ultrasonic methods are compared with respect to detecting and dimensioning defects in laminate structures. Based on the analyses and proposed solutions, it has been proven that the developed through transmission phased array (TTPA) method ensures a much more effective, in terms of quality and quantity, assessment of the condition of hybrid laminates than conventional ultrasonic methods, irrespective of the degree of surface deformation and the type of laminate structure.

Reconstruction of Barely Visible Impact Damage in Composite Structures Based on Non-Destructive Evaluation Results.

Abstract: The occurrence of barely visible impact damage (BVID) in aircraft composite components and structures being in operation is a serious problem, which threatens structural safety of an aircraft, and should be timely detected and, if necessary, repaired according to the obligatory regulations of currently applied maintenance methodologies. Due to difficulties with a proper detection of such a type of damage even with non-destructive testing (NDT) methods as well as manual evaluation of the testing results, supporting algorithms for post-processing of these results seem to be of a high interest for aircraft maintenance community. In the following study, the authors proposed new approaches for BVID reconstruction based on results of ultrasonic and X-ray computed tomographic testing using authored advanced image processing algorithms. The studies were performed on real composite structures taking into consideration failure mechanisms occurring during impact damaging. The developed algorithms allow extracting relevant diagnostic information both from ultrasonic B-and C-Scans as well as from tomographic 3D arrays used for the validation of ultrasonic reconstructed damage locations, which allows for a significant improvement of the detectability of BVID in tested structures. The developed approach can be especially useful for NDT operators evaluating the results of structural NDT inspections.

Evaluation of Different Non-destructive Testing Methods to Detect Imperfections in Unidirectional Carbon Fiber Composite Ropes

Abstract: Online monitoring of carbon fiber reinforced plastic (CFRP) ropes requires non-destructive testing (NDT) methods capable of detecting multiple damage types at high inspection speeds. Three NDT methods are evaluated on artificial and realistic imperfections in order to assess their suitability for online monitoring of CFRP ropes. To support testing, the microstructure and electrical conductivity of a carbon fiber rope is characterized. The compared methods are thermography via thermoelastic stress analysis, ultrasonic testing with commercial phased array transducers, and eddy current testing, supported by tailor-made probes. While thermoelastic stress analysis and ultrasonics proved to be accurate methods for detecting damage size and the shape of defects, they were found to be unsuitable for high-speed inspection of a CFRP rope. Instead, contactless inspection using eddy currents is a promising solution for real-time online monitoring of CFRP ropes at high inspection speeds.

A flexible numerical approach for non-destructive ultrasonic testing based on a time-domain spectral-element method: Ultrasonic modeling of Lamb waves in immersed defective structures and of bulk waves in damaged anisotropic materials

Abstract: This paper deals with the introduction of a spectral-element method, in the time domain, to address problems of ultrasonic wave propagation in the field of the nondestructive ultrasonic testing and evaluation. Two kinds of problems are addressed. The first focuses on guided waves that are often used to control immersed layered defective structures. The second focuses on the use of bulk waves to inspect layered anisotropic media that may contain a defect. This full-wave technique allows for the use of significantly coarser meshes compared to other standard finite-element methods that can be used for non-flat defective or damaged models, as using one element per shortest wavelength is sufficient. It is thus well suited to the simulation of wave propagation phenomena in complex structures at high frequencies. With the goal of first validating our approach for simple cases, we begin by presenting results of simulations for a homogeneous plate. We compare the dispersion curves obtained with the analytical ones. The results are in excellent agreement, and the spectral-element method is fast enough to allow for simulations having a high level of accuracy. We then reproduce and analyze the transmission losses of a quasi plane wave across the immersed plate, and discuss the influence of the finite size of the ultrasonic beam in real physical experiments. We subsequently study the transmission of an ultrasonic wave through an immersed tri-layer composed of two aluminum plates glued together and with a defect, a model that is already not accessible to quasi-analytical calculation techniques. We again determine the dispersion curves, and then study transmission through the structure having a delamination defect compared to transmission through a healthy structure. We then compute the dispersion curves of the same tri-layer structure but with a varying glue thickness. These dispersion curves are compared to those of the healthy structure. We then perform a three-dimensional simulation of a pulse-echo experiment in an immersed medium composed of layers of transversely-isotropic austenitic steel, the axis of symmetry of each layer being tilted differently. This medium represents a very simple model of a weld, in which we include two kinds of defects: a gas bubble resulting from the welding process, and then a branching and crossing Y-shaped crack resulting e.g. from aging. We illustrate their effect on ultrasonic waves by computing the scattered field.

Multi-Frequency Piezoelectric Micromachined Ultrasonic Transducers

Abstract: In this paper multi-frequency piezoelectric MEMS ultrasonic transducers (pMUTs) are designed, characterized and tested for nondestructive evaluation (NDE) of solids. The transducers operate in flexural mode, and are tuned to three different frequencies namely 1 MHz, 1.5 MHz and 2 MHz. The microstructural layers consist of aluminum nitride (AlN) as an active sensing layer sandwiched between metal and doped silicon electrodes. pMUTs are designed with octagonal and circular membranes. The vibration of silicon membrane assists piezoelectric element to convert energies. The transducers are modeled numerically to obtain their dynamic characteristics. Piezoelectric Multi-User MEMS Processes (PiezoMUMPs) are utilized to manufacture pMUTs. The electromechanical characterization shows that the circular design has higher figure of merit as compared to the octagonal design due to more uniform stress distribution transferred between silicon and AlN layers. It is demonstrated that the piezeoelectric layer should be deposited up to the inflection point of diaphragm deformation. This avoids the signal cancellation due to opposite polarization. The performance of pMUTs as receiver is evaluated to detect the creep damage by implementing nonlinear ultrasonic testing (NLUT). NLUT is based on detecting higher harmonics in solids due to heterogeneity in materials. Significant amplification in the second harmonics is obtained due to highly narrowband and low damping characteristics of pMUTs that improves the resolution of NLUT to detect subwavelength damage. Higher harmonics can be detected using small footprint pMUT device, which is not possible with conventional piezoelectric transducers. This allows better spatial resolution of nonlinear measurement.

Traceable Porosity Measurements in Industrial Components Using X-Ray Computed Tomography

Abstract: Manufacturing technologies deliver products that can suffer from various defects, one of which is internal porosity. Pores are present in most of the parts produced by, e.g., casting, additive manufacturing, and injection molding and can significantly affect the performance of the final products. Due to technological and economic limits, typically porosity cannot be completely removed by optimizing process parameters. It is therefore essential to have a measurement technique that can detect and evaluate these defects accurately. Apart from conventional nondestructive techniques, such as ultrasonic testing or Archimedes’ method that suffer from various limitations, X-ray computed tomography has emerged as a promising solution capable of measuring size, spatial distribution, and shape of pores. In this paper, a method to achieve traceable computed tomography measurements of internal porosity using a reference object with calibrated internal artificial defects is described and demonstrated on an industrial case study. Furthermore, the possibility to improve measurement results by optimizing parameters used for the evaluation of acquired data is discussed. The optimization method is based on an iterative procedure that reduces to ±5 × 10−5 mm3 the error of the measured values of total void content in the reference object.

Delamination Detection in Polymeric Ablative Materials Using Pulse-Compression Thermography and Air-Coupled Ultrasound.

Abstract: Ablative materials are used extensively in the aerospace industry for protection against high thermal stresses and temperatures, an example being glass/silicone composites. The extreme conditions faced and the cost-risk related to the production/operating stage of such high-tech materials indicate the importance of detecting any anomaly or defect arising from the manufacturing process. In this paper, two different non-destructive testing techniques, namely active thermography and ultrasonic testing, have been used to detect a delamination in a glass/silicone composite. It is shown that a frequency modulated chirp signal and pulse-compression can successfully be used in active thermography for detecting such a delamination. Moreover, the same type of input signal and post-processing can be used to generate an image using air-coupled ultrasound, and an interesting comparison between the two can be made to further characterise the defect.

Flaw Detection from Ultrasonic Images using YOLO and SSD

Abstract: Non-destructive ultrasonic testing (UT) of materials is used for monitoring critical parts in power plants, aeronautics, oil and gas industry, and space industry. Due to a vast amount of time needed for a human expert to perform inspection it is practical for a computer to take over that task. Some attempts have been made to produce algorithms for automatic UT scan inspection mainly using older, non-flexible analysis methods. In this paper, two deep learning based methods for flaw detection are presented, YOLO and SSD convolutional neural networks. The methods' performance was tested on a dataset that was acquired by scanning metal blocks containing different types of defects. YOLO achieved average precision (AP) of 89.7% while SSD achieved AP of 84.5 %.