Search Results

Battery terminal voltage modelling is crucial for various applications, including electric vehicles, renewable energy systems, and portable electronics. Terminal voltage models are used to determine how a battery will respond under load and can be used to calculate run-time, power capability, and heat generation and as a component of state estimation approaches, such as for state of charge. Previous studies have shown better voltage modelling accuracy for long short-term memory (LSTM) recurrent neural networks than other traditional methods (e.g., equivalent circuit and electrochemical models). This study presents two new approaches – sequence training and data shuffling – to improve LSTM battery voltage models further, making them an even better candidate for the high-accuracy modelling of lithium-ion batteries. Because the LSTM memory captures information from past time steps, it must typically be trained using one series of continuous data.

Battery state-of-charge (SOC) is critical information for the vehicle energy management system and must be accurately estimated to ensure reliable and affordable electrified vehicles (xEV). However, due to the nonlinear temperature, health, and SOC dependent behaviour of Li-ion batteries, SOC estimation is still a significant automotive engineering challenge. Traditional approaches to this problem, such as electrochemical models, usually require precise parameters and knowledge from the battery composition as well as its physical response. In contrast, neural networks are a data-driven approach that requires minimal knowledge of the battery or its nonlinear behaviour. The objective of this work is to present the design process of an SOC estimator using a deep feedforward neural network (FNN) approach. The method includes a description of data acquisition, data preparation, development of an FNN, FNN tuning, and robust validation of the FNN to sensor noise.

An investigation was performed to study the effects of convection, diffusion and thermophoresis on diesel exhaust soot deposition inside a plate-type rectangular cooling section for recirculation (EGR) applications since deposited soot can be detrimental to the heat transfer efficiency of EGR cooling devices. A non-destructive neutron radiography technique was used to measure the soot deposition thickness distribution on the plate surface inside the cooling chamber. The chamber cooled with an inlet water coolant of 20 and 40°C, was installed in a modified exhaust system of a 2.4kW diesel engine and subjected to a mass flow rate of 20kg/hr of diesel exhaust ranging 0 to 3 hours with the exhaust gas temperature at 260°C upstream of the cooling chamber. In this work, the effect of cooling temperature and operation time on thermophoretic deposition was investigated.

An investigation was performed to study the soot deposition and its effect on heat transfer in a cooled cylindrical section. The soot layer thickness was measured using a non-destructive neutron radiography technique. Experiments were performed for a diesel exhaust mass flow rate of 20kg/hr or Reynolds number of approximately 9,000, initial inlet coolant temperatures of approximately 22 and 40°C, and exposure times from 1 to 3 hours. The results show that the nominal soot layer thickness was approximately uniform in the flow direction, hence, the thicker soot layer observed near the entrance by Ismail et al. [8] and de la Cruz et al. [9] was due to entrance effects. The deposited soot layer shows evidence of long wavy thickness variations that appears to be due to a soot re-entrainment and re-deposition moving bed type mechanism. The soot thickness increased and the long wavy variations persisted for larger soot thicknesses when the coolant temperature or wall temperature was lower.

Hybrid electric vehicle (HEV) powertrains are characterized by a complex design environment as a result of both the large number of possible layouts and the need for dedicated energy management strategies. When selecting the most suitable hybrid powertrain architecture at an early design stage of HEVs, engineers usually focus solely on fuel economy (directly linked to tailpipe emissions) and vehicle drivability performance. However, high voltage batteries are a crucial component of HEVs as well in terms of performance and cost. This paper introduces a multitarget assessment framework for HEV powertrain architectures which considers both fuel economy and battery lifetime. A multi-objective formulation of dynamic programming is initially presented as an off-line optimal HEV energy management strategy capable of predicting both fuel economy performance and battery lifetime of HEV powertrain layout options.

Magnesium AZ80 is a medium strength alloy with good corrosion resistance and very good forging capability which offers an affordable commercial alternative to the Mg ZK60 alloy used for wheels in racing cars. Extending the market of Mg AZ80 alloy to automotive wheels requires a better understanding of macro- and micro-properties of this structural material, especially its forging behaviour. In this study the deformation behaviour of Mg AZ80 alloy is characterized by uniaxial compression tests from ambient to 420°C at a variety of strain rates using a Gleeble 1500 simulator. A constitutive relationship coupling materials work hardening and strain rate and temperature dependences is calibrated based on test results. This flow behaviour is input into a finite element model to simulate the forging operation of an automotive wheel with ABAQUS codes.

Accurate battery state of charge (SOC) estimation is essential for safe and reliable performance of electric vehicles (EVs). Lithium-ion batteries, commonly used for EV applications, have strong time-varying and non-linear behaviour, making SOC estimation challenging. In this paper, a processor in the loop (PIL) platform is used to assess the execution time and memory use of different SOC estimation algorithms. Four different SOC estimation algorithms are presented and benchmarked, including an extended Kalman filter (EKF), EKF with recursive least squares filter (EKF-RLS) feedforward neural network (FNN), and a recurrent neural network with long short-term memory (LSTM). The algorithms are deployed to two different NXP S32Kx microprocessors and executed in real-time to assess the algorithms' computational load. The algorithms are benchmarked in terms of accuracy, execution time, flash memory, and random access memory (RAM) use.

The constitutive behavior of aluminum alloy sheet and their resistance spot welds at large strains is critical for light weight vehicle design analysis and life prediction. However, data from uniaxial tensile tests are usually limited to small strains or by material instability. A novel technique was developed using digital image correlation coupled with a modified shear test to directly measure the stress - strain curves of aluminum alloy sheet at large strains. The modified shear sample prevents end rotation of the shear zone as compared to the ASTM B831 test. The results show that the effective stress - effective strain curves from shear tests match those obtained by uniaxial tension, but only by incorporating material anisotropy using the Barlat-Lian yield function. For the first time, the technique was applied to aluminum resistance spot welds to determine both the shear strength and stress-strain curves of spot welds at large strains.

Hybrid, plug-in hybrid, and electric vehicles have enthusiastically embraced rechargeable Li-ion batteries as their primary/supplemental power source of choice. Because the state of charge (SoC) of a battery indicates available remaining energy, the battery management system of these vehicles must estimate the SoC accurately. To estimate the SoC of Li-ion batteries, we derive a normalized state-space model based on Li-ion electrochemistry and apply a Bayesian algorithm. The Bayesian algorithm is obtained by modifying Potter's squareroot filter and named the Potter SoC tracker (PST) in this paper. We test the PST in challenging test cases including high-rate charge/discharge cycles with outlier cell voltage measurements. The simulation results reveal that the PST can estimate the SoC with accuracy above 95% without experiencing divergence.

Due to their high energy density, power density, and durability, lithium-ion (Li-ion) batteries are rapidly becoming the most popular energy storage method for electric vehicles. Difficulty arises in accurately estimating the amount of left capacity in the battery during operation time, commonly known as battery state of charge (SOC). This paper presents a comparative study between six different Equivalent Circuit Li-ion battery models and two different state of charge (SOC) estimation strategies. The Battery models cover the state-of-the-art of Equivalent Circuit models discussed in literature. The Li-ion battery SOC is estimated using non-linear estimation strategies i.e. Extended Kalman filter (EKF) and the Smooth Variable Structure Filter (SVSF). The models and the state of charge estimation strategies are compared against simulation data obtained from AVL CRUISE software.

This paper presents a compact, partially laminated busbar design to connect the DC-link capacitor, high-voltage DC (HVDC) connector, and power module using a single integrated busbar. The proposed busbar design is designed for a high-power and high-voltage Silicon Carbide (SiC) traction inverter. The proposed solution eliminates the need for using separate busbars: one for the connection between the HVDC connector and the DC-link capacitor, and the other one between the connection of the DC-link capacitor and the power module. Incorporating two busbars in a single traction inverter increases the total volume of the inverter and the parasitic components. Thus, the main design goals in this paper are minimizing the parasitic inductances, increasing the power density, and achieving a uniform current distribution across the capacitor cores.

In recent years, environmental regulations in the automotive industry have become increasingly strict, particularly with respect to emissions from diesel engines. Large amounts of these harmful emissions are released during the cfold start of a vehicle, due to the catalytic converter system not yet reaching its light-off temperature. This paper presents an induction heating system which heats the catalytic converter during a cold start, reducing the time for it to reach light-off temperature, and thus reducing cold-start emissions. Detailed dynamometer testing results are used to develop vehicle models of the induction heating system for a diesel Peugeot 308 light duty vehicle. The model is used to quantify the changes in hydrocarbons (HC), carbon monoxide (CO), carbon dioxide (CO2), oxygen (O2), nitrogen oxide (NOx), and fuel consumption on a variety of standard drive cycles.

We have studied the formability of continuous strip cast (CC) AA5754 aluminum alloy for automotive applications. Strip casting technology can considerably reduce material cost compared with conventional direct chill (DC) cast aluminum sheets. However, the CC material tends to exhibit much less post-localization deformation and lower fracture strains compared with DC sheets with similar Fe content, although both alloys show similar strains for the onset of localization. Bendability of the CC alloy is also found to be inferior. The inferior behavior (post-necking and bendability) of the CC alloy can be attributed to the higher incidence of stringer-type particle distributions in the alloy. The formability of the AA5754 alloy has also been studied using two dimensional microstructure-based finite element modeling. The microstructures are represented by grains and experimentally measured particle distributions.

Advanced engine test methods incorporate several different sensing and signal processing techniques for identifying and locating manufacturing or assembly defects of an engine. A successful engine test method therefore, requires advanced signal processing techniques. This paper introduces a novel signal processing technique to successfully detect a faulty internal combustion engine in a quantitative manner. Accelerometers are mounted on the cylinder head and lug surfaces while vibration signals are recorded during engine operation. Using the engine's cam angular position, the vibration signals are transformed from the time domain to the crank-angle domain. At the heart of the transformation lies interpolation. In this paper, linear, cubic spline and sinc interpolation methods are demonstrated for reconstructing vibration signals in the crank-angle domain.

This paper presents a Differential Flatness-Based Control (FBC) approach for the current control of Switched Reluctance Machines (SRMs), a potential candidate for the automotive industry. The main challenges in SRM control methods stem from motor nonlinearity. In electrical drives, FBC has been applied in doubly-fed induction generators, permanent magnet motors, and magnet-assisted synchronous reluctance motors. Among the few papers that have used FBC for SRM, this research distinguishes itself by addressing current control and considering both current and flux-linkage separately as a flat output, an approach not found in previous literature. The performance of the proposed controls is assessed in a three-phase 12/8 SRM against the conventional hysteresis current controller (HCC) and PI controller. Additionally, it is integrated into a torque-sharing function based on a maximum torque per ampere control strategy.

Using urea-based Selected Catalytic Reduction (SCR) systems is an effective way in diesel engine after-treatment systems to meet increasingly stringent emission regulations. The amount of urea injection is critical to achieve high NOx reduction efficiency and low ammonia slip and overdosing or under-dosing of urea injection need to be avoided. One of the difficulties in urea injection amount control lies in the accurate measurement/estimation of the urea injection mass. To effectively address this issue, this paper defined a correction factor for under-dosing or overdosing detection and correction and proposed two methods to identify the correction factor. The first method is based on urea pump model and line pressure. Through frequency analysis, the relation between the urea pump speed and power spectrum characteristics of the line pressure by using FFT method was revealed.

This paper presents a compact thermal management solution for a high-power traction inverter. The proposed design utilizes a stacked cooling system that enables heat extraction from two of the largest heat sources in a power inverter: the power module and the DC-link capacitor. The base plate of the power module has circular pin fins while the capacitor comes with a flat surface which must be placed on a cold plate to provide the adequate heat dissipation. Incorporating individual cooling mechanisms for the DC-link capacitor and the power module would increase the weight, complexity and overall volume of the inverter housing. The proposed cooling system mitigates these problems by integrating the cooling mechanisms of the power module and the DC-link capacitor within a single cooling system. The cooling mechanism is designed to provide a uniform coolant flow with minimal pressure drop across the heat sink of the power module and DC-link capacitor.

Using standard tensile testing methods, the material properties of AKDQ and DP600 steels tubes along the axial direction were determined. A novel in-situ optical strain mapping system ARAMIS® was utilized to evaluate the strain distribution during tensile testing along the axial direction. Microstructural and damage characterization was carried out using microscopy and image analysis techniques to compare the damage evolution and formability of both materials. Failure in both steels was observed to occur via a ductile failure mode. AKDQ was found to be the more formable material as it can achieve higher strains, total elongations and thinning prior to failure than the higher strength DP600.

Lithium-ion battery (LIB) terminal voltage models are investigated using two modelling approaches. The first model is a third-order Thevenin equivalent circuit model (ECM), which consists of an open-circuit voltage in series with a nonlinear resistance and three parallel RC pairs. The parameters of the ECM are obtained by fitting the model to hybrid pulse power characterization (HPPC) test data. The parametrization of the ECM is performed through quadratic-based programming. The second is a novel modelling approach based on long short-term memory (LSTM) recurrent neural networks to estimate the battery terminal voltage. The LSTM is trained on multiple vehicle drive cycles at six different temperatures, including −20°C, without the necessity of battery characterization tests. The performance of both models is evaluated with four automotive drive cycles at each temperature. The results show that both models achieve acceptable performance at all temperatures.

The purpose of this work was to initiate a comparative evaluation of the aqueous corrosion resistance of ferritic stainless steels currently used to fabricate automotive exhaust systems. Both acid condensate and double loop electrochemical potentiokinetic reactivation (DL-EPR) testing using both as-received and heat treated test coupons prepared from Types 409, 409Al, 436 and 439 stainless steel was conducted for this purpose. A truncated version of an in-house acid condensate testing protocol revealed that Type 409Al stainless steel was the most resistant to corrosion of the four ferritic stainless steels examined, whereas Type 409 stainless steel was the least resistance to corrosion.