Search Results

The high-frequency whining noise produced by motors in modern electric vehicles causes a significant issue, leading to annoyance among passengers. This noise becomes even more noticeable due to the quiet nature of electric vehicles, which lack other noises to mask the high-frequency whining noise. To improve the noise caused by motors, it is essential to optimize various motor design parameters. However, this task requires expert knowledge and a considerable time investment. In this study, we explored the application of artificial intelligence to optimize the NVH performance of motors during the design phase. Firstly, we selected and modeled three benchmark motor types using Motor-CAD. Machine learning models were trained using Design of Experiment methods to simulate batch runs of Motor-CAD inputs and outputs.

OTA (over the air) updates help automotive manufacturers to reduce vehicle warranty and recall costs. Vehicle recall is expensive, and many automotive manufacturers have implemented OTA updates. Updating parameters for connected vehicles can be challenging when dealing with thousands of vehicles across different regions. For example, how does the manufacturer prioritise which vehicles need updating? Environmental and geographical factors affect degradation rates and vehicles in hotter regions or congested cities may degrade faster. For EVs, updating the BMS (battery management system) parameters requires careful analysis prior to the update being deployed, to maximise impact and reduce the likelihood of adverse behaviour being introduced. The analysis overhead increases with the number of vehicles. This is because it requires simulation and optimisation of the fleet BMS calibration in a digital twin environment.

With the introduction in Europe of drive cycles such as RDE and WLTC, transient emissions prediction is more challenging than before for passenger car applications. Transient predictions are used in the calibration optimization process to determine the cumulative cycle emissions for the purpose of meeting objectives and constraints. Predicting emissions such as soot accurately is the most difficult area, because soot emissions rise very steeply during certain transients. The method described in this paper is an evolution of prediction using a steady state global model. A dynamic model can provide the instantaneous prediction of boost and EGR that a static model cannot. Meanwhile, a static model is more accurate for steady state engine emissions. Combining these two model types allows more accurate prediction of emissions against time. A global dynamic model combines a dynamic model of the engine air path with a static DoE (Design of Experiment) emission model.

Model-based calibration (MBC) is a systematic method to calibrate an engine control unit (ECU) system. Due to the working principle of MBC, it is only being used for steady-state systems (time independent models). This limits the use of MBC; because an ECU contains statistical and dynamical systems. Due to the limitations of MBC, dynamical systems require manual tuning which may be time-consuming. With the increasing popularity in hybrid and electrical vehicle systems, most of them rely on dynamical systems. Therefore, MBC is about to be superseded by manual parameterization methods. Remarkably, MBC is not limited to the steady state systems. It can be achieved by separating the time factor of a system and extracting the statistical data from a time series measurement. Typically, MBC model is conceived as the representation of a system plant (i.e.: air path, fuel path, mean value engine model). As a matter of fact, MBC model is not limited to identification of system plant.

Design of Experiments (DOE) introduces a number of design types such as space filling design and optimal design. However, optimal design type is best for a system with high prior knowledge. Meanwhile, space-filling design is good for unknown systems, which is normal for engine calibration. It would be best to have a design that can support constructive model building, where a block of engine test is run for most of the day and followed by engine modeling at the end of the day. However, this needs separate space filling design for each day and separate design is susceptible to redundant test points. Among of the five space-filling design type, Sobol sequences and Halton sequences can support constructive model building due to the deterministic random sequence characteristic. When the model is good enough for system prediction, the remaining engine test can stop and proceed to model optimization.

A full calibration exercise of a diesel engine air path can take months to complete (depending on the number of variables). Model-based calibration approach can speed up the calibration process significantly. This paper discusses the overall calibration process of the air-path of the Cat® C7.1 engine using statistical machine learning tool. The standard Cat® C7.1 engine's twin-stage turbocharger was replaced by a VTG (Variable Turbine Geometry) as part of an evaluation of a novel air system. The changes made to the air-path system required a recalculation of the air path's boost set point and desired EGR set point maps. Statistical learning processes provided a firm basis to model and optimize the air path set point maps and allowed a healthy balance to be struck between the resources required for the exercise and the resulting data quality.