Search Results

Laser lap welding quality is a non-linear response based on a host of categorical and numeric material and process variables. This paper describes a Grammatical Evolution approach to the structure identification of the laser lap welding process and compares its performance with linear regression and a neuro-fuzzy inference system.

Load data representing severe customer usage is required during the chassis development process. One area of current research is the use of road profiles for predicting chassis loads. The most direct method of predicting these loads is to run dynamic simulations of the vehicle using numerous road profiles as the excitation. This onerous task may be avoided, and a greatly reduced number of simulations would be required, if roads having similar characteristics can be grouped. Currently, road profiles are characterized by their spectral content. It has been noted by several researches, however, that road profiles are generally nonstationary signals that contain significant transient events and are not well described in the spectral domain. The objective of this work, then, is to develop a method by which the characteristics of the road can be captured by describing these constitutive transient events.

With the increasing emphasis on Systems Engineering, there is a need to ensure that Electrical/Electronic (E/E) Systems Endurance Testing of vehicles, in a real world environment, prior to Production Launch, is performed in a manner and at a technological level that is commensurate with the high level of electronics and computers in contemporary vehicles. Additionally, validating the design and performance of individual standalone electronic systems and modules “on the bench” does not guarantee that all the permutations and combinations of real-world hardware, software, and driving conditions are taken into account. Traditional Proving Ground (PG) vehicle testing focuses mainly on powertrain durability testing, with only a simple checklist being used by the PG drivers as a reminder to cycle some of the electrical components such as the power window switches, turn signals, etc.

A variety of elastomer mounts are being used for vehicles as isolators/dampers between body and frame, on the engine cradle, etc. These vehicle flexible mounts, made of mainly rubber materials and housed in a metallic tube, are indispensable components affecting the quality of the vehicle ride, noise and vibration. In the auto industry, the usual practice when designing vehicle flexible mounts is to minimally reflect impact considerations in the mount design features. However, in most high-speed vehicle crash events where the mounts fail, the crash responses, including occupant injury severity, are known to be very different from the responses of non-failure cases. Even in low-speed vehicle impact cases, excessive deformation of the flexible mounts could cause significant variance in the compliance of the vehicle acceleration level to the air-bag firing and timing threshold requirements.

The popularity of AWD passenger vehicles presents a challenge to provide car-like drive-train NVH within a relatively small package space. This paper describes a drive-train NVH case study in which analysis and test were used, in conjunction, to solve an NVH problem. Also, it details a systematic process of using the analytical model to identify and resolve similar problems. The particular problem for this case study is a noise and vibration issue occurring at 75 MPH primarily in the middle seat of an all-wheel drive vehicle. Tests indicated that it may be due to propeller shaft imbalance. Analysis results showed good correlation with the tests for that loading condition. Several solutions were identified, which were confirmed by both test and analysis. The most cost-effective of these solutions was implemented.

One of the major NVH concerns for automobile manufacturers is the response of a vehicle to the impact of the tire as it encounters a road discontinuity or bump. This paper describes methods for analyzing the impact response of a vehicle to such events. The test vehicle is driven on a dynamometer, on which a bump simulating cleat is mounted. The time histories of the cleat impact response of the vehicle can be classified as a transient and a repeated signal, which should be processed in a special way. This paper describes the related signal processing issues, which include converting the time data into a continous spectrum, determination of the correct scaling factor for the analyzed spectrum, and smoothing out harmonics and fluctuations in the signal. This procedure yields a smooth frequency spectrum with a correctly scaled amplitude, in which the frequency contents can be easily identified.

This paper contains a review of methods for deriving risk curves from biomechanical data obtained from impact experiments on human surrogates. It covers many of the problems and pitfalls of obtaining realistic human risk curves from impact experiments. The strength and weakness of both parametric and non-parametric methods are evaluated. The limitations of standard analysis of censored impact test data are presented. Methods are given for determining risk curves from both doubly censored data and data obtained from impacts to body regions in which there are more than one mechanism of injury. A detailed set of examples is presented in which different experimental data are analyzed using the Consistent Threshold method and the logistic approach. Finally risk curves for published data are presented for the femur, head, thorax, and neck.

This project investigated the effect of carburizing carbon-potential and thermal history on the bending fatigue performance of carburized SAE 4320 gear steel. Modified-Brugger cantilever bending fatigue specimens were carburized at carbon potentials of 0.60, 0.85, 1.05, and 1.25 wt. pct. carbon, and were either quenched and tempered or quenched, tempered, reheated, quenched, and tempered. The reheat treatment was designed to lower the solute carbon content in the case through the formation of transition carbides and refine the prior austenite grain size. Specimens were fatigue tested in a tension/tension cycle with a minimum to maximum stress ratio of 0.1. The bending fatigue results were correlated with case and core microstructures, hardness profiles, residual stress profiles, retained austenite profiles, and component distortion.

For the purposes of analyzing and understanding the general effects of a set of different vehicle attributes on overall crash outcome a fleet model is used. It represents the impact response, in a one-dimensional sense, of two vehicle frontal crashes, across the frontal crash velocity spectrum. The parameters studied are vehicle mass, stiffness, intrusion, pulse shape and seatbelt usage. The vehicle impact response parameters are obtained from the NCAP tests. The fatality risk characterization, as a function of the seatbelt use and vehicle velocity, is obtained from the NASS database. The fatality risk is further mapped into average acceleration to allow for evaluation of the different vehicle impact response parameters. The results indicate that the effects of all the parameters are interconnected and none of them is independent. For example, the effect of vehicle mass on fatality risk depends on seatbelt use, vehicle stiffness, available crush, intrusion and pulse shape.

A 50 Hz Solid-State Relay (SSR) was used to provide pulse-width-modulated power to engine cooling fans for continuous speed control, to reduce airflow noise and improve efficiency. However, this caused the cooling fans to vibrate at the switching frequency and harmonics, thus degrading vehicle NVH performance. This paper describes the problem associated with SSR- powered cooling fans, including root-cause analysis, and identification of areas sensitive to vibration affected by the switching power supply. Based on our analysis, we found several solutions to the problem. Our production solution and some generic recommendations for shroud design are presented in the paper.

The damping loss factor of a structural panel plays a significant role in its vibro-acoustic performance. The objective of this paper is to present a new procedure for evaluating the damping loss factors of these panels. Traditionally, the damping loss factors are determined by using the decay rate of the decay curves which are experimentally obtained from the structure. However, this is time consuming and the accuracy is limited by fluctuations in the decay curve. In this paper, the envelope signal of each decay curve is determined through its Hilbert transform, and the remaining small fluctuations in the envelope signal are further smoothed out by the exponential average method. Finally, the damping loss factor is estimated based on the smoothed envelope signal of each decay curve. A computer program has been developed to implement this procedure. It is shown that this procedure improves both accuracy and efficiency of the decay rate method for estimating damping loss factor.

Brake torque variation is a source of objectionable NVH body/chassis response. Such input commonly results from brake disk thickness variation. The NVH dynamic characteristics of a vehicle can be assessed and quantified through experimental modal testing for determination of mode resonance frequency, damping property, and shape. Standard full vehicle modal testing typically utilizes a random input excitation into the vehicle frame or underbody structure. An alternative methodology was sought to quantify and predict body/chassis sensitivity to brake torque variation. This paper presents a review of experimental modal test methodologies investigated for the reproduction of vehicle response to brake torque variation in a static laboratory environment. Brake caliper adapter random and sine sweep excitation input as well as body sine sweep excitation in tandem with an intentionally locked brake will be detailed.

The problem of predicting the quality of weld is critical to manufacturing. A great deal of data is collected under multiple conditions to predict the quality. The data generated at Daimler Chrysler has been used to develop a model based on grammatical evolution. Grammatical Evolution Technique is based on Genetic Algorithms and generates rules from the data which fit the data. This paper describes the development of a software tool that enables the user to choose input variables such as the metal types of top and bottom layers and their thickness, intensity and speed of laser beam, to generate a three dimensional map showing weld quality. A 3D weld quality surface can be generated in response to any of the two input variables picked from the set of defining input parameters. This tool will enable the user to pick the right set of input conditions to get an optimal weld quality. The tool is developed in Matlab with Graphical User Interface for the ease of operation.

Low frequency torsional vibrations can be a significant source of objectionable vehicle vibrations and in-vehicle boom, especially with changes in engine operation required for improved fuel economy. These changes include lower torque converter lock-up speeds and cylinder deactivation. This paper has two objectives: 1) Examine the effect of increased torsional vibrations on vehicle NVH performance and ways to improve this performance early in the program using test and simulation techniques. The important design parameters affecting vehicle NVH performance will be identified, and the trade-offs required to produce an optimized design will be examined. Also, the relationship between torsional vibrations and mount excursions, will be examined. 2) Investigate the ability of simulation techniques to predict and improve torsional vibration NVH performance. Evaluate the accuracy of the analytical models by comparison to test results.

Laminated steel sheets sandwiched with a polymer core are increasingly used for automotive applications due to their vibration and sound damping properties. However, it has become a major challenge in finite element modeling of laminated steel structures and forming processes due to the extremely large differences in mechanical properties and in the gauges of the polymer core and the steel skins. In this study, circular cup deep drawing and V-bending experiments using laminated steels were conducted in order to develop a modeling technique for laminate forming processes. The effectiveness of several finite element modeling techniques was investigated using the commercial FEM code LS-Dyna. Furthermore, two production parts were selected to verify the modeling techniques in real world applications.

Faced with an ever increasing supply of road load data and no reasonable means to keep track of it, Managing Road Load Data and The Process is about one engineer's charge to deliver a solution. This paper summarizes the approach taken by this engineer and his team to produce a system that not only provides a place to store their data, but facilitates the entire data collection, validation and dissemination process. By following the steps outlined in this document, virtually every opportunity for improvement will be identified. That is, the process is thoroughly explored; the wants and needs of the users are identified, and then, as warranted, turned into functional elements of the system. The result is a central repository that is accessible to all and with the capability of significantly reducing cost and timing. The development process presented here is not a difficult one to accomplish, but does require keeping track of a lot of detail. It can therefore be quite time consuming.

Different methodologies to test transfer case imbalance were investigated in this study. One method utilized traditional standard single plane and two plane methods to measure the imbalance of the transfer case when running it on a dynamic balance machine at steady RPM, while a second method utilized accelerometers and a laser vibrometer to measure vertical vibration on the transfer case when running it on a dynamic balance machine in 4 Hi open mode during a run up from 1000 to 4000 RPM with a 40 RPM difference between the input and output shaft speeds. A comparison of all of the measurements for repeatability and accuracy was done with the goal of determining an appropriate and efficient method that generates the most consistent results. By using the traditional method, the test results were not repeatable. This may be due to the internal complexity of transfer cases. With the second method, good correlation between the measurements was obtained.

In the early stages of a vehicle program, it is a common practice to set target ranges for the global body, suspension and powertrain modes. This modal management process allows engineers to avoid potential noise and vibration problems stemming from strong overlap of major global modes. Before the first prototype hardware is built, finite element models of the body, suspension and powertrain are usually exercised to compare predicted versus targeted ranges of the major system modes in the form of a modal management chart. However, uncertainty associated with the design parameters, manufacturing process and other sources can lead to a major departure from the design intent when the first hardware prototype is built. In this study, a first order reliability method is used to predict variance of the eigen values due to parameter uncertainties. This allows the CAE engineers to add a “three sigma” bound on the eigen values reported in the modal management chart.

Automotive electronics and software is getting complex day by day. More and more features and functions are offered and supported by software in place of hardware. Communication is carried out on the CAN bus instead of hard wired circuits. This architectural transition facilitates lots of flexibility, agility and economy in development. However, it introduces risk of unexpected failures due to insufficient testing and million of possible combinations, which can be created by users during the life time of a product. Model based development supports an effective way of handling these complexities during simulation and also provide oracle for its validation. Based on priorities and type of applications, test vectors can be auto generated and can be used for formal verification of the models. These auto-generated test vectors are valuable assets in testing and can be effectively reused for target hardware (ECU) verification.

In embedded software world, development and testing are becoming far more complex with growing functionality and fail safe strategies. As a result of that, model-based software development is getting increasingly popular in capturing the functional requirements and auto generating the code from these validated models to avoid any functional deficiency. However, the complexity in the model may not be correctly interpreted by the code generation tool and may result to an incorrect code behavior. In this paper, a methodology has been proposed and implemented to validate the generated code against the models. Simulation test scripts are recorded in the modeling environment to generate the desired set of test inputs. These input scripts are designed to get complete transition and state exposure to maximize the functional coverage. With these test scripts, expected outputs are recorded for downstream validation in the simulation environment with mature models.