Search Results

As electrification of powertrains is progressing, diversification of hybrid powertrains increases. This generally imposes the challenge for a supervisory controller of how to optimally control the torque of the electric machine(s). Architectures, which have at least one belt driven electric machine, are an essential part of the portfolio. This paper describes a strategy on how to include the losses of the belt device in the determination of optimal electric machine torque command. It first depicts a physics-based method for controlling optimal electric machine torque command for systems without a belt connected electric machine. This method considers the constraints of the electric machine(s) as well as the power limitations from the electric devices, which supply power to the motors.

Engineering design-decisions often involve many attributes which can differ in the levels of their importance to the decision maker (DM), while also exhibiting complex statistical relationships. Learning a decision-making policy which accurately represents the DM’s actions has long been the goal of decision analysts. To circumvent elicitation and modeling issues, this process is often oversimplified in how many factors are considered and how complicated the relationships considered between them are. Without these simplifications, the classical lottery-based preference elicitation is overly expensive, and the responses degrade rapidly in quality as the number of attributes increase. In this paper, we investigate the ability of deep preference machine learning to model high-dimensional decision-making policies utilizing rankings elicited from decision makers.

Through real-time online optimization, the full potential of the performance and energy efficiency of multi-gear, multi-mode, series–parallel hybrid powertrains can be realized. The framework allows for the powertrain to be in its most efficient configuration amidst the constantly changing hardware constraints and performance objectives. Typically, the different gears and hybrid/electric modes are defined as discrete states, and for a given vehicle speed and driver power demand, a formulation of optimization costs, usually in terms of power, are assigned to each discrete states and the state which has the lowest cost is naturally selected as the desired of optimum state. However, the optimization results would be sensitive to numerical exactitude and would typically lead to a very noisy raw optimum state. The generic approach to stabilization includes adding hysteresis costs to state-transitions and time-debouncing.

Making a sturdy battery box or enclosure is one of the many challenging issues that the expansion of electrification entails. Many characteristics of an effective battery housing contribute to the safety of passengers and shield the battery from the harsh environment created by vibrations and shocks due to varying road profiles in the vehicle. This results in stress and deformations of different degrees. There is a need to understand and develop a correlation between structural performance and lightweight design of battery enclosure as this can increase the range of the drive and the life cycle of a battery pack. This paper investigates the following points: I) A conceptualized CAD model of battery enclosure is developed to understand the design parameters such as utilization of different material for strength and structural changes for performance against vibration and strength.

This paper presents an optimal control co-design framework of a parallel electric-hydraulic hybrid powertrain specifically tailored for heavy-duty vehicles. A pure electric powertrain, comprising a rechargeable lithium-ion battery, a highly efficient electric motor, and a single or double-speed gearbox, has garnered significant attention in the automotive sector due to the increasing demand for clean and efficient mobility. However, the state-of-the-art has demonstrated limited capabilities and has struggled to meet the design requirements of heavy-duty vehicles with high power demands, such as a class 8 semi-trailer truck. This is especially evident in terms of a driving range on one battery charge, battery charging time, and load-carrying capacity. These challenges primarily stem from the low power density of lithium-ion batteries and the low energy conversion efficiency of electric motors at low speeds.

In this paper, a control strategy for state of charge (SOC) allocation using navigation data for Hybrid Electric Vehicle (HEV) propulsion systems is proposed. This algorithm dynamically defines and adjusts a SOC target as a function of distance travelled on-line, thereby enabling proactive management of the energy store in the battery. The proposed approach incorporates variances in road resistance and adheres to geolocation constraints, including ultra-low emission zones (uLEZ). The anticipated advantages are particularly pronounced during scenarios involving extensive medium-to-long journeys characterized by abrupt topological changes or the necessity for exclusive electric vehicle (EV) mode operation. This novel solution stands to significantly enhance both drivability and fuel economy outcomes.

This paper focuses on an inherent problems of active damping control prevalent in contemporary hybrid torque controls. Oftentimes, a supervisory torque controller utilizes simplified system models with minimal system states representation within the optimization problem, often not accounting for nonlinearities and stiffness. This is motivated by enabling the generation of the optimum torque commands with minimum computational burden. When inherent lash and stiffness of the driveline are not considered, the resulting command can lead to vibrations and oscillations in the powertrain, reducing performance and comfort. The paper proposes a Linear Quadratic Integral (LQI)-based compensator to be integrated downstream the torque supervisory algorithm, which role is to shape transient electric machine torques, compensating for the stiffness and backlash present in the vehicle while delivering the driver-requested wheel torque.

During input tracking, closed-loop performance is strongly influenced by the dynamic of the system under control. Internal and external delays, such as actuation and measurement delays, have a detrimental effect on the bandwidth and stability. Additionally, production controllers are discrete in nature and the sampling time selection is another critical factor to be considered. In this paper we analyze the impact of both transported delay and controller sampling time on tracking performance using an electric machine speed-control problem as an example. A simple linear PI controller is used for this exercise. Furthermore, we show how the PI parameters can be adjusted to maintain a certain level of performance as the delays and sampling times are modified. This is achieved through an optimization algorithm that minimizes a specifically designed cost function.

A finite element/contact mechanics (FE/CM) method is used to determine the tooth contact forces, static transmission error, and tooth pair stiffnesses for spur gear pairs that have pitting damage. The pitting damage prevents portions of the tooth surface from carrying load, which results in meaningfully different contact pressure distribution on the gear teeth and deformations at the mesh. Pits of elliptical shape are investigated. Parametric analyses are used to investigate the effect of pit width (along the tooth face) and height (along the tooth profile) on the gear tooth mesh interface. Pitting damage increases static transmission error and decreases tooth pair stiffness. Tooth contact forces differ only in the portions of the mesh cycle when multiple pairs of teeth are in contact and share the transmitted load. Pitting damage does not change the loads when only a single pair of teeth are in contact.

One of the building blocks of the Stellantis hybrid powertrain embedded control software computes the maximum and minimum values of objective functions, such as output torque, as a function of engine torque, hybrid motor torque and other variables. To test such embedded software, an offline reference function was created. The reference function calculates the ideal minimum and maximum values to be compared with the output of the embedded software. This article presents the offline reference function with an emphasis on mathematical novelties. The reference function computes the minimum and maximum points of a linear objective function as a function of six independent variables, subject to 42 linear and two nonlinear constraints. Concave domains, curved surfaces, disjoint domains and multiple local extremum points challenge the algorithm. As a theorem, the conditions and methods for running trigonometric calculations in 6D Euclidean space are presented.

Many chassis and powertrain components in the transportation and automotive industry experience multi-axial cyclic service loading. A thorough load-history leading to durability damage should be considered in the early vehicle production steps. The key feature of rubber fatigue analysis discussed in this study is how to define local critical location strain time history based on nominal and complex load time histories. Material coupon characterization used here is the crack growth approach, based on fracture mechanics parameters. This methodology was utilized and presented for a truck engine mount. Temperature effects are not considered since proving ground (PG) loads are generated under isothermal high temperature and low frequency conditions without high amounts of self-heating.

The Los Angeles City Traffic (LACT) brake test is well known acclaimed procedure used by many vehicle manufacturers to assess the brake pad wear behavior and to investigate the Noise, Vibration and Harness (NVH) performance of the brake system. The LACT driving route consists of a set of real-world driving conditions, which has been considered representative of the US passenger vehicle market. The scope of this study is to mimic the LACT test using finite element analysis (FEA) to calculate the wear displacement based on Rhee’s theory. The Leading-edge and trailing edge of the brake pad’s wear tendency is also predicted from the simulation. The finite element model for wear simulation consists of brake system viz., Rotor, Knuckle, Pads, Anchor bracket, Piston, and Caliper.

Fasteners, commonly used in automotive industry, play an important role in the safety and reliability of the vehicle structural system. In practical application, bolted joints would never undergo fully reversed loading; there always will be positive mean stress on bolt. The mean stress has little influence on the fatigue life if the maximum stress is lower than a threshold which is near the yield stress of the bolt. However, when the sum of the mean stress and the stress amplitude exceeds the threshold, the endurance limit stress amplitude decreases fast as the mean stress increases. The purpose of this paper is to research the fatigue endurance limit of a fastener and establish the threshold for safe design in automotive application. In order to obtain the fatigue endurance limit at different mean stress levels, various mechanical tests were performed on M12x1.75 and M16x1.5 Class 10.9 fasteners using MTS test systems.

This paper proposes a new nonlinear model predictive control (NMPC) for autonomous vehicle path tracking problem. The vehicle is equipped with active rear steering, allowing independent control of front and rear steering. Traditional NMPC, which runs at fixed sampling rate, has been shown to provide satisfactory control performance in this problem. However, the high throughput of NMPC limits its implementation in production vehicle. To address this issue, we propose a novel event-triggered NMPC formulation, where the NMPC is triggered to run only when the actual states deviate from prediction beyond certain threshold. In other words, the event-triggered NMPC will formulate and solve a constrained optimal control problem only if it is enabled by a trigger event. When NMPC is not triggered, the optimal control sequence computed from last NMPC instance is shifted to determine the control action.

Reducing emissions and the carbon footprint of our society have become imperatives requiring the automotive industry to adapt and develop technologies to strive for a cleaner sustainable transport system and for sustainable economic prosperity. Electrified hybrid electric vehicles (HEV), plug-in hybrid electric vehicles (PHEV) and range extender powertrains provide potential solutions for reducing emissions, but they present challenges in terms of thermal management. A key requirement for meeting these challenges is accurately to predict the thermal loading and temperatures of an internal combustion engine (ICE) quickly under multiple full-load and part-load conditions. Computational Fluid Dynamics (CFD) and thermal survey database methods are used to derive thermal loading of the engine structure and are well understood but typically only used at full-load conditions.

Cargo box is one of the indispensable structures of a pickup truck which makes it capable of transporting heavy cargo weights. This heavy cargo weight plays an important role in durability performance of the box structure when subjected to road load inputs. Finite element representation for huge cargo weight is always challenging, especially in a linear model under dynamic proving ground road load durability analysis using a superposition approach. Any gap in virtual modeling technique can lead to absurd cargo box modes and hence durability results. With the existing computer aided engineering (CAE) approach, durability results could not correlate much with physical testing results. It was crucial to have the right and robust CAE modeling technique to represent the heavy cargo weight to provide the right torsional and cargo modes of the box structure and in turn good durability results.

A finite element/contact mechanics formulation is used to analyze the dynamic tooth forces that arise from damage-induced vibrations in spur gear pairs. Tooth root crack damage of varying sizes are analyzed for a wide range of speeds that include resonant gear speeds. The added localized compliance from tooth root crack damage leads to a re-distribution of the forces on the individual gear teeth in mesh. At speeds away from resonance, smaller dynamic forces occur on the damaged tooth and larger dynamic forces occur on the tooth that engages immediately after it. These dynamic tooth contact forces cause additional transient dynamic response in the gear pair. For certain speeds and sufficiently large tooth root cracks, the damage-induced dynamic response causes large enough vibration that tooth contact loss nonlinearity occurs. For some speeds near resonance, the damage-induced vibrations cause teeth that normally lose contact to remain in contact due to vibration.

Stringent requirements for high fuel economy and energy efficiency mandate using increasingly complex vehicle thermal systems in most types of electrified vehicles (xEVs). Enabling the maximum benefits of such complex thermal systems under the full envelope of their operating modes demands designing complex thermal control systems. This is becoming one of the most challenging problems for electrified vehicles. Typically, the thermal systems of such vehicles have several modes of operation, constituting nonlinear multiple-input/multiple-output (MIMO) dynamic systems that cannot be efficiently controlled using classical or rule based strategies. This paper covers the different steps towards the design of a model-based control (MBC) strategy that can improve the overall performance of xEV thermal control systems. To achieve the above objective, the latter MBC strategy is applied to control cooling of the cabin and high voltage battery.

In the automotive industry, a Malfunction Indicator Light (MIL) is commonly employed to signify a failure or error in a vehicle system. To identify the root cause that has triggered a particular fault, a technician or engineer will typically run diagnostic tests and analyses. This type of analysis can take a significant amount of time and resources at the cost of customer satisfaction and perceived quality. Predicting an impending error allows for preventative measures or actions which might mitigate the effects of the error. Modern vehicles generate data in the form of sensor readings accessible through the vehicle’s Controller Area Network (CAN). Such data is generally too extensive to aid in analysis and decision making unless machine learning-based methods are used. This paper proposes a method utilizing a recurrent neural network (RNN) to predict an impending fault before it occurs through the use of CAN data.

Accurate prediction of in-vehicle powertrain bounce mode is necessary to ensure optimum responses are achieved at driver’s touch points during 4post shake or rough road shake events. But, during the early stages of vehicle development, building a detailed vehicle finite element (FE) model is not possible and often powertrain bounce modes are computed assuming the powertrain to be a stand-alone unit. Studies conducted on FE models of a large SUV with body on frame architecture showed that the stand-alone approach overestimates the powertrain bounce mode. Consequently, there is a need for a simplified version of vehicle model which can be built early on to compute powertrain modes. Previously, representing all the major components as rigid entities, simplified unibody vehicle models have been built to compute powertrain modes. But such an approach would be inaccurate here, for a vehicle with body on frame architecture due to the flexible nature of the frame (even at low frequencies).