Jesper Martinsson

Complete post-separation of overlapping ultrasonic signals by combining hard and soft modeling

In some ultrasonic measurement situations, an adequate signal separation is difficult to achieve. A typical situation is material characterization of thin media using pulse-echo or through-transmission techniques, when the time-of-flight in the media is shorter than the emitted signals time support. Separated signals are necessary to obtain accurate estimates of material properties and transit times. In this paper a new method is proposed that enables complete post-separation of measured coinciding signals. The method is based on a combination of hard physical and soft empirical models, which allows for a description of both known and unknown properties making a complete separation possible. The validity and limitations of the model and the separation results are thoroughly addressed. The proposed technique is verified using real measurements on thin dispersive samples and validated using residual analysis. The experimental results show a complete separation with uncorrelated and normally distributed residuals. The method enables characterization and/or flow analysis in difficult overlapping situations.

Model-Based Estimation of Thin Multi-Layered Media Using Ultrasonic Measurements

In ultrasonic measurement situations, when dealing with media of multi-layered structures consisting of 1 or more thin layers, analysis of the measured ultrasonic waveform can be difficult because of overlapping and reverberant echoes. Information from the individual layers is then difficult to extract because the individual echoes cannot be detected. In this study, we use a parametric layer model to analyze the multi-layered material in a system identification approach. The parameters of the model are connected to physical properties of the investigated material, e.g., the reflection coefficients, the time-of-flight, and the attenuation. The main advantage using this model is that the complexity of the model is connected to the number of layers rather than the number of observable echoes in the received ultrasonic waveform. A system of linear equations is presented, giving the opportunity to find the model for both pulse-echo and through-transmission measurements. A thorough effort is made on the parameter estimation and optimization algorithm. The model is validated with practical measurements on a 3-layered structure using both pulse-echo and through-transmission techniques. The 3-layered material consists of a thin embedded middle layer with the time-of-flight in that layer shorter than the emitted signals time support, giving rise to overlapping echoes. Finally the relation between the model parameters and physical properties of the material is established.

Model-based phase velocity and attenuation estimation in wideband ultrasonic measurement systems

A parametric method to estimate frequency-dependent phase velocity and attenuation is presented in this paper. The parametric method is compared with standard nonparametric Fourier analysis techniques using numerical simulations as well as real pulse-echo experiments. Approximate standard deviations are derived for both methods and validated with numerical simulations. Compared to standard Fourier analysis, the parametric model gives considerably lower variance when estimating attenuation and phase velocity. In contrast to nonparametric techniques, the proposed estimator avoids the phase unwrapping problem because analytical expressions for the continuous phase velocity and attenuation can be derived

Estimating the underlying signal waveform, noise covariance and synchronization jitter from unsynchronized measurements

In this paper a new synchronization technique is presented for applications using repeated measurements or experiments with periodically excited signals. The objective with repeated or periodic measurements is often to retrieve an estimate of the (noise reduced) signal and its uncertainties. However, these measurements need to be synchronized to obtain accurate estimates. Existing synchronization techniques are limited to specific signal and noise conditions, such as white Gaussian noise or narrowband signals, to achieve good performance. The proposed method, not limited by these conditions, extracts statistical information regarding the underlying signal and the noise contained in the measurements, to obtain good synchronization (asymptotically optimal). The Cram´ er–Rao lower bound (CRLB) is derived for the synchronization problem, including bounds for the underlying signal waveform and the covariance of the measurement noise, both considered unknown. The method, which is shown to be the maximum-likelihood estimator (MLE) in both white and colored Gaussian noise, is compared with the CRLB along with standard sub-sample estimation and aligning techniques using Monte Carlo simulations. The results show significant mean square error (MSE) improvements compared to standard synchronization techniques. Synchronization results using the proposed technique are presented for repeated ultrasonic measurements, to validate the method in a real measurement situation, and to experimentally support theoretical results.

3B-1 Flaw Detection in Layered Media Based on Parametric Modeling of Overlapping Ultrasonic Echoes

In materials consisting of several thin layers, multiple reflections within the structure give rise to received ultrasonic signals composed of overlapping echoes. In this paper we present a parametric model that can be used to decompose such signals into the individual reflections. We derive a maximum likelihood estimator for the the model parameters, which are then used in a generalized likelihood ratio test (GLRT) to detect flaws in multi-layered structures. We show with simulations how the presence of a thin bonding layer in a three-layer structure can be detected. The probability of detection is shown to be ap96%, for a signal-to-noise ratio (SNR) of 15 dB and a probability of false alarm of 5 %

Parametric modeling of wave propagation in gas mixtures - a system identification approach

In ultrasonic pulse-echo systems, the observable properties are restricted to frequency dependent attenuation and phase velocity, which in turn are related to the material properties of the investi ...

Calibration of simulation models for ultrasonic transducers based on implicit calibration

There are numerous software packages available for modeling of the sound pressure fields emitted by ultrasound transducers and transducer arrays. Accurate modeling of a real-world transducer requires knowledge of several parameters that are generally not known. This paper presents an estimation principle that can be used to calibrate such models, based on measurements of the sound field. The model separates the problem into estimation of the transducers electro-mechanical impulse response and its spatial impulse response. The latter being what the software packages generally model. We demonstrate the principle with measurements of the sound field from a non-focused 5 MHz transducer. The results show that the modeled sound field agrees well with measurements.

3E-4 Estimating the Underlying Signal Waveform and Synchronization Jitter from Repeated Measurements

In this paper we present a synchronization technique, for applications using repeated or periodically excited measurements. The problem with existing techniques is their limitations to specific signal and noise conditions, such as white Gaussian noise or narrowband signals. The proposed method extracts statistical information about the underlying signal and noise in the measurements to obtain good synchronization (asymptotically optimal). The Cramer-Rao lower bound (CRLB) is derived for the synchronization problem, including bounds for the underlying signal waveform and the covariance of the noise. The method, which is the maximum-likelihood estimator for both white and colored Gaussian noise, is compared with standard sub-sample estimation and aligning techniques using Monte Carlo simulations. The results show significant improvements compared to standard synchronization techniques

Ultrasonic imaging of thin layers within multi-layered structures

In the area of process control, non-destructive testing (NDT) using ultrasound is valuable due to its noninvasive properties. In process control, imaging of surface profiles is used to locate defects or problematic areas in order to quickly steer the manufacturing process on track again. This paper presents a method for imaging of parallel thin layers within multi-layered structures. Due to the application in process control a parametric model is used, and all subsequent analysis is performed on the model parameters rather than on the signal waveforms, resulting in a necessary data reduction. The parameters in the model are directly connected to physical properties, such as the reflection coefficients, time-of-flights, and attenuation coefficients. Experimental results shows that the estimated model parameters can be used in imaging of thin layer properties within the material structure. Images of embedded layers with a thickness about the wavelength is shown. Result also show that flaws can be detected in such structures. The results are verified by comparing the images to visual inspections of photographs.

Bayesian hierarchical model-based prognostics for lithium-ion batteries

To optimise operation and maintenance, knowledge of the ability to perform the required functions is vital. The ability is governed by the usage of the system (operational issues) and availability aspects like reliability of different components. This paper proposes a Bayesian hierarchical model (BHM)-based prognostics approach applied to Li-ion batteries, where the goal is to analyse and predict the discharge behaviour of such batteries with variable load profiles and variable amounts of available discharge data. The BHM approach enables inferences for both individual batteries and groups of batteries. Estimates of the hierarchical model parameters and the individual battery parameters are presented, and dependencies on load cycles are inferred. A BHM approach where the operational and reliability aspects end of life (EoD) and end of life (EoL) is studied where its shown that predictions of EoD can be made accurately with a variable amount of battery data. Without access to measurements, e.g. predicting a new battery, the predictions are based only on the prior distributions describing the similarity within the group of batteries and their dependency on the load cycle. A discharge cycle dependency can also be identified in the result giving the opportunity to predict the battery reliability.