No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
Wavelet neural networks for EEG modeling and classification
Abstract
Ph.D. George Vachtsevanos
... For example, in [1-3] wavelet networks represent an implementation of wavelet decomposition. This recently-developed technique, is an efficient tool utilized in many applications applied in the signal processing domain; such as function approximation and 3D data modeling [4,5]. The wavelet network structure has also been applied in nonlinear system identification [6][7][8]. ...
... Once the initialization set W of wavelet basis functions has been constructed, the next step is to select the best M-size subset (M < L) of wavelet basis functions in W to estimate f (x). However, in general, a search through all the M-size subsets is a computationally expensive procedure suffering from a combinatorial optimization problem [4]. ...
... One nonlinear optimization technique consists in using genetic algorithms which have been utilized successfully by Echauz et al. in [4], for radial wavelet networks, and Chen in [22] for RBF networks. In fact, genetic algorithms can provide optimal or near-optimal network topologies, at the expense of extensive computational requirements [22]. ...
We propose, in this paper, a novel technique for large Laplacian boundary deformations using estimated rotations. The introduced method is used in the domain of Region of Interest (ROI) to align features of mesh based on Multi Mother Wavelet Neural Network (MMWNN) structure found in several mother wavelet families. The wavelet network allows the alignment of the characteristic points of the original mesh towards the target mesh. The key component of our correspondence scheme is a deformation energy that penalizes geometric distortion, encourages structure preservation and simultaneously allows mesh topology changes. To ensure the design of wavelet neural network architecture, an optimization algorithm should be applied to estimate and optimize the network parameters. Therefore, we compare our approach of 3d mesh deformation using MMWNN architecture based on genetic algorithm and our approach relying on Levenberg-Marquardt Method. We also discuss the existing comparison metrics for static and deformed triangle meshes employing the two mentioned approaches. Besides, we enumerate their strengths, weaknesses and relative performance.
... WNNs have recently emerged as a powerful new type of ANN [4], [5]. They resemble radial basis function (RBF) networks because of the localized support of their wavelet basis functions [9]. In contrast to classical sigmoidal-based ANNs, wavelet networks provide efficient network construction techniques, faster training times, and multiresolution analysis capabilities. ...
... The reason for this is that wavelet networks are a linear combination of localized basis functions that offer several advantages including orthogonality, efficient numerical procedures and multiresolution analysis capabilities over an RBF network. Wavelet networks have been applied to a wide variety of applications including: nonlinear functional approximation and nonparametric estimation [4], system identification and control tasks [13], and modeling and classification [9]. ...
... Once the initialization set of wavelet basis functions has been constructed, the next step is to select the "best"-size subset of wavelet basis functions in for estimating . However, in general, a search through all the-size subsets is a computationally expensive combinatorial optimization problem, i.e., it is NP complete [9]. One nonlinear optimization technique would be to use genetic algorithms (GAs), which have been utilized successfully by Echauz [9] for radial wavelet networks and Chen [15], [16] for RBF networks. ...
A novel objective function is presented that incorporates both local and global errors as well as model parsimony in the construction of wavelet neural networks. Two methods are presented to assist in the minimization of this objective function, especially the local error term. First, during network initialization, a locally adaptive grid is utilized to include candidate wavelet basis functions whose local support addresses the local error of the local feature set. This set can be either user-defined or determined using information derived from the wavelet transform modulus maxima representation. Next, during the network construction, a new selection procedure based on a subspace projection operator is presented to help focus the selection of wavelet basis functions to reduce the local error. Simulation results demonstrate the effectiveness of these methodologies in minimizing local and global error while maintaining model parsimony and incurring a minimal increase on computational complexity.
... Lancet Neurology 2002; 1: [22][23][24][25][26][27][28][29][30] Epilepsy affects more than 50 million individuals worldwide-about 1% of the world's population. Two-thirds of affected individuals have seizures that are controlled by antiepileptic drugs. ...
... For most EEG studies, embedding dimensions of seven to ten have been used. 30 In a variation of these methods, called dynamical similarity, the EEG is not directly embedded from all of its digitally sampled points but only the timepoints when the trace crosses the zero voltage axis, and the slope of the voltage curve is upward, are used to reconstruct a trajectory. The trajectory of the EEG signal is calculated (by the method of delays and an embedding dimension of ten) in a 5 min reference window far from seizure onset; it is then compared with computed trajectories in 30 s test data epochs in the EEG as seizures approach. ...
For almost 40 years, neuroscientists thought that epileptic seizures began abruptly, just a few seconds before clinical attacks. There is now mounting evidence that seizures develop minutes to hours before clinical onset. This change in thinking is based on quantitative studies of long digital intracranial electroencephalographic (EEG) recordings from patients being evaluated for epilepsy surgery. Evidence that seizures can be predicted is spread over diverse sources in medical, engineering, and patent publications. Techniques used to forecast seizures include frequency-based methods, statistical analysis of EEG signals, non-linear dynamics (chaos), and intelligent engineered systems. Advances in seizure prediction promise to give rise to implantable devices able to warn of impending seizures and to trigger therapy to prevent clinical epileptic attacks. Treatments such as electrical stimulation or focal drug infusion could be given on demand and might eliminate side-effects in some patients taking antiepileptic drugs long term. Whether closed-loop seizure-prediction and treatment devices will have the profound clinical effect of their cardiological predecessors will depend on our ability to perfect these techniques. Their clinical efficacy must be validated in large-scale, prospective, controlled trials.
... It is often not feasible to capture the continuous time temperature operational history over the use life of the product . Current health management systems provide nearly zero visibility into health of electronics and packaging for prediction of impending failures [ McCann 2005 ; Marko 1996 ; Schauz 1996 ; Shiroishi 1997 ] . The built - in - self tests ( BIST ) are generally used to give electronic assemblies the ability to test and diagnose themselves with minimal interaction from external test equipment . ...
Gold wire bonding has been widely used as first-level interconnect in semiconductor packaging. The increase in the gold price has motivated the industry search for alternative to the gold wire used in wire bonding and the transition to copper wire bonding technology. Potential advantages of transition to Cu-Al wire bond system includes low cost of copper wire, lower thermal resistivity, lower electrical resistivity, higher deformation strength, damage during ultrasonic squeeze, and stability compared to gold wire. However, the transition to the copper wire brings along some trade-offs including poor corrosion resistance, narrow process window, higher hardness, and potential for cratering. Formation of excessive Cu-Al intermetallics may increase electrical resistance and reduce the mechanical bonding strength. Current state-of-art for studying the Cu-Al system focuses on accumulation of statistically significant number of failures under accelerated testing. In this paper, a new approach has been developed to identify the occurrence of impending apparently-random defect fall-outs and pre-mature failures observed in the Cu-Al wirebond system. The use of intermetallic thickness, composition and corrosion as a leading indicator of failure for assessment of remaining useful life for Cu-al wirebond interconnects has been studied under exposure to high temperature and temperature-humidity. Damage in wire bonds has been studied using x-ray Micro-CT. Microstructure evolution was studied under isothermal aging conditions of 150°C, 175°C, and 200°C till failure. Activation energy was calculated using growth rate of intermetallic at different temperatures. Effect of temperature and humidity on Cu-Al wirebond system was studied using Parr Bomb technique at different elevated temperature and humidity conditions (110°C/100%RH, 120°C/100%RH, 130°C/100%RH) and failure mechanism was developed. The present methodology uses evolution of the IMC thickness, composition in conjunction with the Levenberg-Marquardt algorithm to identify accrued damage in wire bond subjected to thermal aging. The proposed method can be used for quick assessment of Cu-Al parts to ensure manufactured part consistency through sampling.
... It is often not feasible to capture the continuous time temperature operational history over the use life of the product. Current health management systems provide nearly zero visibility into health of electronics and packaging for prediction of impending failures [McCann 2005; Marko 1996; Schauz 1996; Shiroishi 1997]. The built-in-self tests (BIST) are generally used to give electronic assemblies the ability to test and diagnose themselves with minimal interaction from external test equipment. ...
Gold wire bonding has been widely used as first-level interconnect in semiconductor packaging. The increase in the gold price has motivated the industry search for alternative to the gold wire used in wire bonding and the transition to copper wire bonding technology. Potential advantages of transition to Cu-Al wire bond system includes low cost of copper wire, lower thermal resistivity, lower electrical resistivity, higher deformation strength, damage during ultrasonic squeeze, and stability compared to gold wire. However, the transition to the copper wire brings along some trade-offs including poor corrosion resistance, narrow process window, higher hardness, and potential for cratering. Formation of excessive Cu-Al intermetallics may increase electrical resistance and reduce the mechanical bonding strength. Current state-of-art for studying the Cu-Al system focuses on accumulation of statistically significant number of failures under accelerated testing. In this paper, a new approach has been developed to identify the occurrence of impending apparently-random defect fall-outs and pre-mature failures observed in the Cu-Al wirebond system. The use of intermetallic thickness, composition and corrosion as a leading indicator of failure for assessment of remaining useful life for Cu-al wirebond interconnects has been studied under exposure to high temperature and temperature-humidity. Damage in wire bonds has been studied using x-ray Micro-CT. Microstructure evolution was studied under isothermal aging conditions of 150°C, 175°C, and 200°C till failure. Activation energy was calculated using growth rate of intermetallic at different temperatures. Effect of temperature and humidity on Cu-Al wirebond system was studied using Parr Bomb technique at different elevated temperature and humidity conditions (110°C/100%RH, 120°C/100%RH, 130°C/100%RH) and failure mechanism was developed. The present methodology uses evolution of the IMC thickness, composition in conjunction with the Levenberg-Marquardt algorithm to identify accrued damage in wire bond subjected to thermal aging. The proposed method can be used for quick assessment of Cu-Al parts to ensure manufactured part consistency through sampling.
... Advanced health management techniques for EPS and avionic systems are required to meet the safety, reliability, maintainability , and supportability requirements of aeronautics and space systems. Current health management techniques in EPS and avionic systems provide very-limited or no-visibility into health of power electronics, and packaging to predict impending failures.34353637. Maintenance has evolved over the years from corrective maintenance to performing time-based preventive maintenance. Future improvements in reduction of system downtime require emphasis on early detection of degradation mechanisms. ...
... Current health management techniques in EPS and Avionic Systems provide very-limited or no-visibility into health of power electronics, and packaging to predict impending failures. [McCann 2005, Marko 1996, Schauz 1996, Shiroishi 1997]. Electronics systems may be subjected to prolonged periods of thermal exposure over wide temperature extremes and long periods of thermal aging at often high ambient temperatures. ...
Electronic systems are often stored for long periods prior to deployment in the intended environment. Aging has been previously shown to effect the reliability and constitutive behavior of second-level leadfree interconnects. Deployed systems may be subjected to cyclic thermo-mechanical loads subsequent to deployment. Prognostication of accrued damage and assessment of residual life is extremely critical for ultra-high reliability systems in which the cost of failure is too high. The presented methodology uses leading indicators of failure based on microstructural evolution of damage to identify impending failure in electronic systems subjected to sequential stresses of thermal aging and thermal cycling. The methodology has been demonstrated on area-array ball-grid array test assemblies with Sn3 Ag0.5Cu interconnects subjected to thermal aging at 125°C and thermal cycling from -55 to 125°C for various lengths of time and cycles. Damage equivalency methodologies have been developed to map damage accrued in thermal aging to the reduction in thermo-mechanical cyclic life based on damage proxies. Assemblies have been prognosticated to assess the error with interrogation of system state and assessment of residual life. Prognostic metrics including α-λ metric, sample standard deviation, mean square error, mean absolute percentage error, average bias, relative accuracy, and cumulative relative accuracy have been used to compare the performance of the damage proxies.
... Current health management techniques in EPS and Avionic Systems provide very-limited or no-visibility into health of power electronics, and packaging to predict impending failures. [McCann 2005, Marko 1996, Schauz 1996, Shiroishi 1997]. Electronics systems may be subjected to prolonged periods of thermal exposure over wide temperature extremes and long periods of thermal aging at often high ambient temperatures. ...
Field deployed electronics may accrue damage due to environmental exposure and usage after finite period of service but may not often have any macro-indicators of failure such as cracks or delamination. A method to interrogate the damage state of field deployed electronics in the pre-failure space may allow insight into the damage initiation, progression, and remaining useful life of the deployed system. Aging has been previously shown to effect the reliability and constitutive behavior of second-level leadfree interconnects. Prognostication of accrued damage and assessment of residual life can provide valuable insight into impending failure. In this paper, field deployed parts have been extracted and prognosticated for accrued damage and remaining useful life in an anticipated future deployment environment. A subset of the field deployed parts have been tested to failure in the anticipated field deployed environment to validate the assessment of remaining useful life. In addition, some parts have been subjected to additional know thermo-mechanical stresses and the incremental damage accrued validated with respect to the amount of additional damage imposed on the assemblies. The presented methodology uses leading indicators of failure based on micro-structural evolution of damage to identify accrued damage in electronic systems subjected to sequential stresses of thermal aging and thermal cycling. Damage equivalency methodologies have been developed to map damage accrued in thermal aging to the reduction in thermo-mechanical cyclic life based on damage proxies. The expected error with interrogation of system state and assessment of residual life has been quantified. Prognostic metrics including a-X metric, sample standard deviation, mean square error, mean absolute percentage error, average bias, relative accuracy, and cumulative relative accuracy have been used to compare the performance of the damage proxies.
... Current health management t�c�niques in EPS and Avionic Systems provide very limited or no-visibility into health of power electronics, and packaging to predict impending failures. [McCann 2005, Marko 1996, Schauz 1996, Shiroishi 1997]. Ele � tronics systems may be subjected to prolonged perIods of thermal exposure over wide temperature extremes and long periods of thermal aging at often high ambient temperatures. ...
Field deployed electronics may accrue damage due to environmental exposure and usage after finite period of service but may not often have any macro-indicators of failure such as cracks or delamination. A method to interrogate the damage state of field deployed electronics in the pre-failure space may allow insight into the damage initiation, progression, and remaining useful life of the deployed system. Aging has been previously shown to effect the reliability and constitutive behavior of second-level leadfree interconnects. Prognostication of accrued damage and assessment of residual life can provide valuable insight into impending failure. In this paper, field deployed parts have been extracted and prognosticated for accrued damage and remaining useful life in an anticipated future deployment environment. A subset of field deployed parts has been tested to failure in the anticipated field deployed environment to validate the assessment of remaining useful life. In addition, some parts have been subjected to additional know thermo-mechanical stresses and the incremental damage accrued validated with respect to the amount of additional damage imposed on the assemblies. The presented methodology uses leading indicators of failure based on micro-structural evolution of damage to identify accrued damage in electronic systems subjected to sequential stresses of thermal aging and thermal cycling. Damage equivalency methodologies have been developed to map damage accrued in thermal aging to the reduction in thermo-mechanical cyclic life based on damage proxies. The expected error with interrogation of system state and assessment of residual life has been quantified. Prognostic metrics including α-λ metric, sample standard deviation, mean square error, mean absolute percentage error, average bias, relative accuracy, and cumulative relative accuracy have been used to compare the performance of the damage proxies.
... Current health management techniques in EPS and Avionic Systems provide very-limited or no-visibility into health of power electronics, and packaging to predict impending failures. [McCann 2005, Marko 1996, Schauz 1996, Shiroishi 1997]. ...
Aerospace-electronic systems usually face a very harsh environment, requiring them to survive the high strain rates, e.g. during launch and re-entry and thermal environments including extreme low and high temperatures. Traditional health monitoring methodologies have relied on reactive methods of failure detection often providing little or no insight into the remaining useful life of the system. In this paper, a mathematical approach for interrogation of system state under cyclic thermo-mechanical stresses has been developed for 6-different leadfree solder alloy systems. Data has been collected for leading indicators of failure for alloy systems including, Sn3Ag0.5Cu, Sn3Ag0.7Cu, Sn1Ag0.5Cu, Sn0.3Ag0.5Cu0.1Bi, Sn0.2Ag0.5Cu0.1Bi0.1Ni, 96.5Sn3.5Ag second-level interconnects under the application of cyclic thermo-mechanical loads. Methodology presented resides in the pre-failure space of the system in which no macro-indicators such as cracks or delamination exist. Systems subjected to thermo-mechanical damage have been interrogated for system state and the computed damage state correlated with known imposed damage. The approach involves the use of condition monitoring devices which can be interrogated for damage proxies at finite time-intervals. Interrogation techniques are based on derivation of damage proxies, and system prior damage based non-linear least-squares methods including the Levenberg-Marquardt Algorithm. The systempsilas residual life is computed based on residual-life computation algorithms.
... Advanced HM techniques for electrical power systems and avionic systems are required to meet the safety, reliability, maintainability, and supportability requirements of aeronautics and space systems. Current HM techniques in EPS and avionic systems provide very limited or no visibility into the health of power electronics and packaging to predict impending failures [51], [52], [62], An incentive for the development of prognostics and HM methodologies has been provided by the need for reduction in operation and maintenance process costs. Future improvements in the reduction of system downtime require emphasis on the early detection of degradation mechanisms. ...
Requirements for system availability for ultrahigh reliability electronic systems such as airborne and space electronic systems are driving the need for advanced health monitoring techniques for the early detection of the onset of damage. Aerospace electronic systems usually face a very harsh environment, requiring them to survive the high strain rates, e.g., during launch and reentry, and thermal environments, including extremely low and high temperatures. Traditional health monitoring methodologies have relied on reactive methods of failure detection often providing little or no insight into the remaining useful life of the system. In this paper, a mathematical approach for the interrogation of the system state under cyclic thermomechanical stresses has been developed for six different lead-free solder alloy systems. Data have been collected for leading indicators of failure for alloy systems, including Sn3Ag0.5Cu, Sn0.3Ag0.7Cu, Sn1Ag0.5Cu, Sn0.3Ag0.5Cu0.1Bi, Sn0.2Ag0.5Cu0.1Bi0.1Ni, and 96.5 Sn3.5Ag second-level interconnects under the application of cyclic thermomechanical loads. The methodology presented resides in the prefailure space of the system in which no macroindicators such as cracks or delamination exist. Systems subjected to thermomechanical damage have been interrogated for the system state and the computed damage state correlated with the known imposed damage. The approach involves the use of condition monitoring devices which can be interrogated for damage proxies at finite time intervals. The interrogation techniques are based on the derivation of damage proxies and system prior-damage-based nonlinear least square methods, including the Levenberg-Marquardt algorithm. The system's residual life is computed based on residual-life computation algorithms.
... Current health management techniques in EPS and avionic systems provide very-limited or no-visibility into health of power electronics, packaging to predict impending failures. [McCann 2005, Marko1996, Schauz 1996, Shiroishi 1997]. Maintenance has evolved over the years from corrective maintenance to performing time-based preventive maintenance. ...
Structural damage to BGA interconnects incurred during vibration testing has been monitored in the pre-failure space using resistance spectroscopy based state space vectors, rate of change of the state variable, and acceleration of the state variable. The technique is intended for condition monitoring in high reliability applications where the knowledge of impending failure is critical and the risks in terms of loss-of-functionality are too high to bear. Future state of the system has been estimated based on a second order Kalman Filter model and a Bayesian Framework. The measured state variable has been related to the underlying interconnect damage in the form of inelastic strain energy density. Performance of the prognostication health management algorithm during the vibration test has been quantified using performance evaluation metrics. The methodology has been demonstrated on leadfree area-array electronic assemblies subjected to vibration. Model predictions have been correlated with experimental data. The presented approach is applicable to functional systems where corner interconnects in area-array packages may be often redundant. Prognostic metrics including α-λ metric, sample standard deviation, mean square error, mean absolute percentage error, average bias, relative accuracy, and cumulative relative accuracy have been used to assess the performance of the damage proxies. The presented approach enables the estimation of residual life based on level of risk averseness.
Electronic systems may be subjected to prolonged and intermittent periods of storage prior to deployment or usage. Prior studies have shown that the lead-free solder interconnects show measurable degradation in the mechanical properties even after the brief exposures to high temperature. In this paper, a method has been developed for determining the equivalent storage time to produce identical damage at a different temperature. Electronics subjected to accelerated tests often have a well-defined thermal profile for a specified period of time. Quantification of the thermal profile in field-deployed electronics may be often difficult because of the variance in the environment conditions and usage profile. There is a need for tools and techniques to quantify the damage in deployed systems in the absence of macroindicators of damage without the knowledge of prior stress history. The approach for mapping damage in the lead-free second-level interconnects between different thermal conditions is new. High-reliability applications, such as avionics and missile systems, may be often exposed to long periods of storage prior to deployment. The effect of storage at different temperature conditions can be mapped using the presented approach. A framework has been developed to investigate the system state and estimate the remaining useful life of the solder ball subjected to a variety of isothermal aging conditions, including 60 °C, 75 °C, and 125 °C for periods of time between 1 and 4 weeks. Data on damage precursors, including the rate of change in the normalized phase growth and the normalized IMC thickness, has been collected and analyzed to derive physics-based damage mapping relationships for aging. Mathematical relationships have been derived for the damage mapping to various thermal storage environments to facilitate determining an appropriate time-temperature combination to reach a particular level of damage state. Activation energy for the leading indicators of failure is also computed. Specific damage proxies examined include the phase-growth indicator and the intermetallic thickness. The viability of the approach has been demonstrated for the lead-free test assemblies subjected to multiple thermal aging at 60 °C, 75 °C, and 125 °C. Damage mapping relationships are derived from the data based on the two separate leading indicators.
Gold wire bonding has been widely used as the first-level interconnect in semiconductor packaging. The increase in the gold price has motivated the industry search for an alternative to the gold wire used in wire bonding and the transition to a copper wire bonding technology. Potential advantages of transition to a Cu-Al wire bond system include low cost of copper wire, lower thermal resistivity, lower electrical resistivity, higher deformation strength, damage during ultrasonic squeeze, and stability compared with gold wire. However, the transition to the copper wire brings along some tradeoffs, including poor corrosion resistance, narrow process window, higher hardness, and potential for cratering. Formation of excessive Cu-Al intermetallics may increase the electrical resistance and reduce the mechanical bonding strength. Current state of the art for studying the Cu-Al system focuses on the accumulation of statistically significant number of failures under accelerated testing. In this paper, a new approach has been developed to identify the occurrence of impending apparently random defect fall-outs and premature failures observed in the Cu-Al wire bond system. The use of intermetallic thickness, composition, and corrosion as a leading indicator of failure for the assessment of the remaining useful life for Cu-Al wire bond interconnects has been studied under exposure to high temperature. Damage in the wire bonds has been studied using an X-ray micro-Computed Tomography (CT). Microstructure evolution was studied under the isothermal aging conditions of 150 °C, 175 °C, and 200 °C until failure. Activation energy was calculated using the growth rate of intermetallic at different temperatures. An effect of temperature and humidity on a Cu-Al wire bond system was studied using the Parr bomb technique at different elevated temperature and humidity conditions (110 °C/100%RH, 120 °C/100%RH, and 130 °C/100%RH), and a failure mechnism was developed. The present methodology uses the evolution of the intermetallic compound thickness and composition in conjunction with the Levenberg-Marquardt algorithm to identify accrued damage in wire bond subjected to thermal aging. The proposed method can be used for a quick assessment of Cu-Al parts to ensure manufactured part consistency through sampling.
Prognostics has received considerable attention recently as an emerging sub-discipline within SHM. Prognosis is here strictly defined as "predicting the time at which a component will no longer perform its intended function." Loss of function is oftentimes the time at which a component fails. The predicted time to that point then becomes the remaining useful life (RUL). For prognostics to be effective, it must be performed well before deviations from normal performance propagate to a critical effect. This enables a failure preclusion or prevention function to repair or replace the offending components, or, if the components cannot be repaired, to retire the system (or vehicle) before the critical failure occurs. Therefore, prognosis has the promise to provide critical information to system operators that will enable safer operation and more cost-efficient use. To that end, the US Department of Defense (DoD), NASA, and industry have been investigating this technology for use in their vehicle health management solutions. Dedicated prognostic algorithms (in conjunction with failure detection and fault isolation algorithms) must be developed that are capable of operating in an autonomous and real-time vehicle health management system software architecture that is possibly distributed in nature. This envisioned prognostic and health management system will be realized in a vehicle-level reasoner that must have visibility and insight into the results of local diagnostic and prognostic technologies implemented at the line replaceable unit (LRU) and subsystem levels. Accomplishing this effectively requires an integrated suite of prognostic technologies that compute failure effect propagation through diverse subsystems and that can capture interactions that occur in these subsystems. In this chapter a generic set of selected prognostic algorithm approaches is presented and an overview of the required vehicle-level reasoning architecture needed to integrate the prognostic information across systems is provided.
Wire bonding is predominant mode of interconnect in electronics packaging. Traditionally material used for wire bonding is gold. But industry is slowly replacing gold wire bond by copper-aluminum wire bond because of the lower cost and better mechanical properties than gold, such as high strength, high thermal conductivity etc. Numerous studies have been done to analyze failure mechanism of Cu-Al wire bonds. Cu-Al interface is a predominant location for failure of the wirebond interconnects. In this paper, the use of intermetallic thickness as leading indicator-of-failure for prognostication of remaining useful life for Cu-Al wire bond interconnects has been studied. For analysis, 32 pin chip scale packages were used. Packages were aged isothermally at 200°C and 250°C for 10 days. Packages were withdrawn periodically after 24 hours and its IMC thickness was measured using SEM. The parts have been prognosticated for accrued damage and remaining useful life in current or anticipated future deployment environment. The presented methodology uses evolution of the IMC thickness in conjunction with the Levenberg-Marquardt Algorithm to identify accrued damage in wire bond subjected to thermal aging. The proposed method can be used for equivalency of damage accrued in Cu-Al parts subjected to multiple thermal aging environments.
Electronics in automotive underhood applications may be subjected to temperatures in the neighborhood of 150°C to 175°C. Several of the electronics functions such as lane departure warning systems, collision avoidance systems are critical to vehicle operation. Prior studies have shown that low silver leadfree SnAgCu alloys exhibit pronounced deterioration in mechanical properties even after short exposure to high temperatures. Current life prediction models for second level interconnects do not provide a method for quick-turn assessment of the effect of mean temperature on cyclic life. In this paper, a method has been developed for assessment of the effect of mean cyclic temperature on the thermal fatigue reliability based on physics based leading damage indicators including phase-growth rate and the intermetallic thickness. Since the quantification of the thermal profile in the field applications may be often very difficult, the proposed method does not require the acquisition of the thermal profile history. Three environments of -50°C to +50°C, 0°C to 100°C, 50°C to 150°C with identical thermal excursion and different mean temperatures have been studied. Test assemblies with three different packages including CABGA 144, PBGA 324, and PBGA 676 have been used for the study. Damage-proxy based damage-equivalency relationships have been derived for the three thermal cycles. Weibull distributions have been developed for the three test assemblies to evaluate the effect of the mean cyclic temperature on the thermal fatigue life. Data indicates that the thermal fatigue lie drops with the increase in mean temperature of the thermal cycle even if the thermal excursion magnitude is kept constant. Damage equivalency model predictions of the effect of mean temperature of the thermal cycle have been validated versus weibull life distributions. The damage proxy based damage equivalency methodology shows good correlation with experimental data.
Electronic systems are often stored for long periods prior to deployment in the intended environment. Aging has been previously shown to effect the reliability and constitu-tive behavior of second-level leadfree interconnects. Deployed systems may be subjected to cyclic thermo-mechanical loads subsequent to deployment. Prognostication of accrued damage and assessment of residual life is extremely critical for ultrahigh reliability systems in which the cost of failure is too high. The presented methodology uses leading indicators of failure based on microstructural evolution of damage to identify impending failure in electronic systems subjected to sequential stresses of thermal aging and thermal cycling. The methodology has been demonstrated on area-array ball-grid array test assemblies with Sn3Ag0.5Cu interconnects subjected to thermal aging at 125 °C and thermal cycling from −55 to 125 °C for various lengths of time and cycles. Damage equivalency methodologies have been developed to map damage accrued in thermal aging to the reduction in thermo-mechanical cyclic life based on damage proxies. Assemblies have been prognosticated to assess the error with interrogation of system state and assessment of residual life. Prognostic metrics including α − λ metric, sample standard deviation, mean square error, mean absolute percentage error, average bias, relative accuracy (RA), and cumulative RA have been used to compare the performance of the damage proxies.
Electronic interconnects may encounter damage under exposure to vibration and mechanical shock. The damage may manifest itself as an open circuit after a finite period of operation. The traditional methods for damage detection such as microscopy and x-ray are destructive in nature and provide limited information offline. In this paper, resistance spectroscopy and phase-sensitive detection based state space vectors have been used to monitor the health of interconnects and assess remaining useful life. Two techniques including Kalman Filter and Extended Kalman Filter have been used to estimate the future state of the system. The measured state variable has been related to the underlying interconnect damage in the form of inelastic strain energy density. Performance of both the prognostication health management algorithms during vibration and mechanical shock has been quantified using performance evaluation metrics. The methodology has been demonstrated on lead-free area-array and advanced interconnect electronic assemblies. Model predictions have been correlated with experimental data. The presented approach is applicable to functional systems where corner interconnects in area-array packages may be often redundant. Prognostic metrics including α-λ metric, sample standard deviation, mean square error, mean absolute percentage error, average bias, relative accuracy, and cumulative relative accuracy have been used to assess the performance of the damage proxies. The presented approach enables the estimation of residual life based on level of risk averseness.
Field deployed electronics may accrue damage due to environmental exposure and usage after finite period of service but may not often have any macro-indicators of failure such as cracks or delamination. A method to interrogate the damage state of field deployed electronics in the pre-failure space may allow insight into the damage initiation, progression, and remaining useful life of the deployed system. Aging has been previously shown to effect the reliability and constitutive behavior of second-level leadfree interconnects. Prognostication of accrued damage and assessment of residual life can provide valuable insight into impending failure. In this paper, field deployed parts have been extracted and prognosticated for accrued damage and remaining useful life in an anticipated future deployment environment. A subset of the field deployed parts have been tested to failure in the anticipated field deployed environment to validate the assessment of remaining useful life. In addition, some parts have been subjected to additional know thermo-mechanical stresses and the incremental damage accrued validated with respect to the amount of additional damage imposed on the assemblies. The presented methodology uses leading indicators of failure based on micro-structural evolution of damage to identify accrued damage in electronic systems subjected to sequential stresses of thermal aging and thermal cycling. Damage equivalency methodologies have been developed to map damage accrued in thermal aging to the reduction in thermo-mechanical cyclic life based on damage proxies. The expected error with interrogation of system state and assessment of residual life has been quantified. Prognostic metrics including α-λ metric, sample standard deviation, mean square error, mean absolute percentage error, average bias, relative accuracy, and cumulative relative accuracy have been used to compare the performance of the damage proxies.
Electronic systems may be subjected to prolonged and intermittent periods of storage prior to deployment or usage. Prior studies have shown that leadfree solder interconnects show measurable degradation in the mechanical properties even after brief exposures to high temperature. In this paper, a method has been developed for the determining equivalent storage time to produce identical damage at a different temperature. Electronics subjected to accelerated tests often have a well-defined thermal profile for a specified period of time. Quantification of the thermal profile in field deployed electronics may be often difficult because of variance in the environment conditions and usage profile. There is need for tools and techniques to quantify damage in deployed systems in absence of macro-indicators of damage without knowledge of prior stress history. Approach for mapping damage in leadfree second-level interconnects under between thermal conditions is new. High reliability applications such as avionics and missile systems may be often exposed to long periods of storage prior to deployment. Effect of storage at different temperature conditions can be mapped using the presented approach. A framework has been developed to investigate the system state and estimate the remaining useful life of solder ball subjected to a variety of isothermal aging conditions including 60°C, 75°C and 125°C for periods of time between 1-week and 4-week. Data on damage precursors has been collected and analyzed to derive physics based damage mapping relationships for aging. Mathematical relationships have been derived for the damage mapping to various thermal storage environments to facilitate determining appropriate time-temperature combination to reach a particular level of damage state. Activation energy for the leading indicators of failure is also computed. Specific damage proxies examined include the phase-growth indicator and the intermetallic thickness. The v- ability of the approach has been demonstrated for leadfree test assemblies subjected to multiple thermal aging at 60° C, 75°C and 125°C. Damage mapping relationships are derived from data based on the two separate leading indicators.
Electronic systems are often stored for long periods prior to deployment in the intended environment. Aging has been previously shown to effect the reliability and constitutive behavior of second-level leadfree interconnects. Deployed systems may be subjected to cyclic thermo-mechanical loads subsequent to deployment. Prognostication of accrued damage and assessment of residual life is extremely critical for ultrahigh reliability systems in which the cost of failure is too high. The presented methodology uses leading indicators of failure based on microstructural evolution of damage to identify impending failure in electronic systems subjected to sequential stresses of thermal aging and thermal cycling. The methodology has been demonstrated on area-array ball-grid array test assemblies with Sn3Ag0.5Cu interconnects subjected to thermal aging at 125 °C and thermal cycling from -55 to 125 °C for various lengths of time and cycles. Damage equivalency methodologies have been developed to map damage accrued in thermal aging to the reduction in thermo-mechanical cyclic life based on damage proxies. Assemblies have been prognosticated to assess the error with interrogation of system state and assessment of residual life. Prognostic metrics including α - λ metric, sample standard deviation, mean square error, mean absolute percentage error, average bias, relative accuracy (RA), and cumulative RA have been used to compare the performance of the damage proxies.
A technique has been developed for monitoring the structural damage accrued in ball grid array interconnects during operation in vibration environments. The technique uses resistance spectroscopy based state space vectors, rate of change of the state variable, and acceleration of the state variable in conjunction with extended Kalman filter, and is intended for the pre-failure time-history of the component. Condition monitoring using the presented technique can provide knowledge of impending failure in high reliability applications where the risks associated with loss-of-functionality are too high to bear. The methodology has been demonstrated on 96.5%Sn3.0%Ag0.5%Cu (SAC305) lead-free area-array electronic assemblies subjected to vibration. The future state of the system has been estimated based on a second order extended Kalman filter model and a Bayesian Framework. The measured state variable has been related to the underlying interconnect damage using plastic strain. Performance of the prognostication health management algorithm during the vibration test has been quantified using performance evaluation metrics. Model predictions have been correlated with experimental data. The presented approach is applicable to functional systems where corner interconnects in area-array packages may be often redundant. Prognostic metrics including α-λ metric, beta, and relative accuracy have been used to assess the performance of the damage proxies. The presented approach enables the estimation of residual life based on level of risk averseness.
Due to the increasing desire for having more autonomous vehicle platforms and life cycle support mechanisms, there is a great
need for the development of prognostic health management technologies that can detect, isolate and assess remaining useful
life of critical subsystems. To meet these needs for next generation systems, dedicated prognostic algorithms must be developed
that are capable of operating in an autonomous and real-time vehicle health management system that is distributed in nature
and can assess overall vehicle health and its ability to complete a desired mission. This envisioned prognostic and health
management system should allow vehicle-level reasoners to have visibility and insight into the results of local diagnostic
and prognostic technologies implemented down at the LRU and subsystem levels. To accomplish this effectively requires an integrated
suite of prognostic technologies that can be applied to critical systems and can capture fault/failure mode propagation and
interactions that occur in these systems, all the way up through the vehicle level. In the chapter, the authors will present
a generic set of selected prognostic algorithm approaches, as well as provide an overview of the required vehicle-level reasoning
architecture needed to integrate the prognostic information across systems.
Wavelet Neural Networks have been developed for fault diagnosis and prognosis with unique capabilities in addressing identification and classification problems. A fault diagnostic and prognostic system is presented by using Wavelet Neural Networks and Dynamic Wavelet Neural Networks. We used a Matlab Simulink model of a chiller system and applied the Wavelet Neural Network to detect and recover sensor errors. It has been shown that this method gains good performance in sensor fault diagnostics.
Electronic assemblies deployed in harsh environments may be subjected to multiple thermal environments during the use-life of the equipment. Often the equipment may not have any macro-indicators of damage such as cracks or delamination. Quantification of thermal environments during use-life is often not feasible because of the data-capture and storage requirements, and the overhead on core-system functionality. There is need for tools and techniques to quantify damage in deployed systems in absence of macro-indicators of damage without knowledge of prior stress history. The presented PHM framework is targeted towards high reliability applications such as avionic and space systems. In this paper, Sn3.0Ag0.5Cu alloy packages have been subjected to multiple thermal cycling environments including -55 to 125C and 0 to 100C. Assemblies investigated include area-array packages soldered on FR4 printed circuit cards. The methodology involves the use of condition monitoring devices, for gathering data on damage pre-cursors at periodic intervals. Damage-state interrogation technique has been developed based on the Levenberg-Marquardt Algorithm in conjunction with the micro structural damage evolution proxies. The presented technique is applicable to electronic assemblies which have been deployed on one thermal environment, then withdrawn from service and targeted for redeployment in a different thermal environment. Test cases have been presented to demonstrate the viability of the technique for assessment of prior damage, operational readiness and residual life for assemblies exposed to multiple thermo-mechanical environments. Prognosticated prior damage and the residual life show good correlation with experimental data, demonstrating the validity of the presented technique for multiple thermo-mechanical environments.
Leading indicators of failure have been developed based on high-frequency characteristics, and system-transfer function derived from resistance spectroscopy measurements during shock and vibration. The technique is intended for condition monitoring in high reliability applications where the knowledge of impending failure is critical and the risks in terms of loss-of-functionality are too high to bear. Previously, resistance spectroscopy measurements [Constable 1992, Lizzul 1994, Prassana 1995] have been used during thermal cycling tests to monitor damage progression due to thermo-mechanical stresses. The development of resistance spectroscopy based damage pre-cursors for prognostication under shock and vibration is new. In this paper, the high-frequency characteristics, and system transfer function based on resistance spectroscopy measurements have been correlated with the damage progression in electronics during shock and vibration. Packages being examined include ceramic area-array packages. Second level interconnect technologies examined include copper-reinforced solder column, SAC305 solder ball, and 90Pb10Sn high-lead solder ball. Assemblies have been subjected to 1500g, 0.5 ms pulse [JESD-B2111]. Continuity has been monitored in-situ during the shock test for identification of part-failure. Resistance spectroscopy based damage pre-cursors have been correlated with the optically measured transient strain based feature vectors. High speed cameras have been used to capture the transient strain histories during shock-impact. Statistical pattern recognition techniques have been used to identify damage initiation and progression and determine the statistical significance in variance between healthy and damaged assemblies. Models for healthy and damaged packages have been developed based on package characteristics. Data presented shows that high-frequency characteristics and system-transfer characteristics based on resistance spectroscopy measurements can be used for-
condition-monitoring, damage initiation and progression in electronic systems. A positive prognostic distance has been demonstrated for each of the interconnect technologies tested.
Leading indicators of failure have been developed based on high-frequency characteristics, and system-transfer function derived from resistance spectroscopy measurements during shock and vibration. The technique is intended for condition monitoring in high reliability applications where the knowledge of impending failure is critical and the risks in terms of loss-of-functionality are too high to bear. Previously, resistance spectroscopy measurements [Constable 1992, Lizzul 1994, Prassana 1995] have been used during thermal cycling tests to monitor damage progression due to thermo-mechanical stresses. The development of resistance spectroscopy based damage pre-cursors for prognostication under shock and vibration is new. In this paper, the high-frequency characteristics, and system transfer function based on resistance spectroscopy measurements have been correlated with the damage progression in electronics during shock and vibration. Packages being examined include ceramic area-array packages. Second level interconnect technologies examined include copper-reinforced solder column, SAC305 solder ball, and 90Pb10Sn high-lead solder ball. Assemblies have been subjected to 1500g, 0.5 ms pulse [JESD-B2111]. Continuity has been monitored in-situ during the shock test for identification of part-failure. Resistance spectroscopy based damage pre-cursors have been correlated with the optically measured transient strain based feature vectors. High speed cameras have been used to capture the transient strain histories during shock-impact. Statistical pattern recognition techniques have been used to identify damage initiation and progression and determine the statistical significance in variance between healthy and damaged assemblies. Models for healthy and damaged packages have been developed based on package characteristics. Data presented shows that high-frequency characteristics and system-transfer characteristics based on resistance spectroscopy measurements can be used for-
-
condition-monitoring, damage initiation and progression in electronic systems. A positive prognostic distance has been demonstrated for each of the interconnect technologies tested.
In recent years, three-phase boost rectifiers, due to their high efficiency, good current quality and low EMI emissions are widely used in industry as Power Factor Correction (PFC) converters. Performance criteria of these converters significantly improve with increasing the switching frequency, and highly depend on the control strategy used. This paper presents a novel approach to control of three phase boost rectifiers. The proposed method is a hybrid of wavelet and neural network (WNN). Simulation results show that this control strategy is very robust, flexible and also the response of the system is very fast. With applying WNN to the three-phase boost rectifier, the controlled system has unity power factor, sinusoidal input currents and regulated output voltage.
Prognostic algorithms for condition based maintenance of critical machine components are presenting major challenges to software designers and control engineers. Predicting time-to-failure accurately and reliably is absolutely essential if such maintenance practices are to find their way into the industrial floor. Moreover, means are required to assess the performance and effectiveness of these algorithms. This paper introduces a prognostic framework based upon concepts from dynamic wavelet neural networks and virtual sensors and demonstrates its feasibility via a bearing failure example. Statistical methods to assess the performance of prognostic routines are suggested that are intended to assist the user in comparing candidate algorithms. The prognostic and assessment methodology proposed here may be combined with diagnostic and maintenance scheduling methods and implemented on a conventional computing platform to serve the needs of industrial and other critical processes
A general-purpose, systematic algorithm is presented, consisting of a genetic programming module and a k-nearest neighbor classifier to automatically create artificial features--computer-crafted features possibly without a known physical meaning--directly from the reconstructed state-space trajectory of intracranial EEG signals that reveal predictive patterns of epileptic seizures. The algorithm was evaluated with IEEG data from seven patients, with prediction defined over a horizon of 1-5 min before unequivocal electrographic onset. A total of 59 baseline epochs (nonseizures) and 55 preictal epochs (preseizures) were used for validation purposes. Among the results, it is shown that 12 seizures out of 55 were missed while four baseline epochs were misclassified, yielding 79% sensitivity and 93% specificity.
Patient-specific epilepsy seizure detectors were designed based on the genetic programming artificial features algorithm, a general-purpose, methodic algorithm comprised by a genetic programming module and a k-nearest neighbor classifier to create synthetic features. Artificial features are an extension to conventional features, characterized by being computer-coded and may not have a known physical meaning. In this paper, artificial features are constructed from the reconstructed state-space trajectories of the intracranial EEG signals intended to reveal patterns indicative of epileptic seizure onset. The algorithm was evaluated in seven patients and validation experiments were carried out using 730.6 hr of EEG recordings. The results with the artificial features compare favorably with previous benchmark work that used a handcrafted feature. Among other results, 88 out of 92 seizures were detected yielding a low false negative rate of 4.35%.
Artificial neural networks (ANN) constitute a powerful class of nonlinear function approximate for model-free estimation. ANN has been widely used in pattern recognition, prediction and classification. In the artificial neural network approach, we compare radial basis function neural networks (RBFNN) and wavelet neural networks for multispectral image classification. The aim of this study is to examine the effectiveness of the neural network model for multispectral image classification. Radial basis function neural network is used for its advantages of rapid training, generality and simplicity over feedforward backpropagation neural network. The k-means clustering is used to choose the initial radial basis centers and widths for the RBFNN. The wavelet is a localized function that is capable of detecting some features in signals. A wavelet basis function is assigned for each neuron and each synaptic weight is determined by learning. An attempt is also made to study the performance of the RBFNN with the centers and widths chosen using the classical k-means clustering.
ResearchGate has not been able to resolve any references for this publication.