Anti Jerk Control Of Hybrid Electric Vehicle Engineering Essay

The report details the review done in the area of anti-jerk control of vehicles. It explains the phenomenon of driveline oscillations excited due to torsion of driveline at high engine torques causing unwanted longitudinal oscillations in vehicle body.

Report starts with a broad introduction to the topic and lays down the aims and objectives of the research project.

The aims and objectives of the project are to design an active feedback control for an electric motor of a hybrid electric vehicle to damp these undesirable oscillations.

A summary review of previous research work done in this area is presented which defines the starting point of the project. Different techniques used for anti-jerk control in conventional vehicle are discussed. Basic approach to model based control design is presented in the report. An introduction to hybrid electric vehicle driveline and description of its components different from conventional vehicle is shown.

The report also lays down the procedure of the work to be followed for the research project with a project plan.

INTRODUCTION

The increasing pressure of protecting the natural environment and to decrease the dependence on non-renewable sources of energy has encouraged vehicle manufacturers to develop more clean and efficient power systems. As a result Hybrid vehicles are hot topic for research among vehicle manufacturers. Efforts are being made to make the vehicles more fuel efficient with high performance and minimum CO2 emissions. Diesel hybrids are considered as next big thing among ecological designs since diesel engines are 20 % – 30% more fuel efficient than their petrol counterparts.

Nowadays noise, vibration and harshness (NVH) issues have become more important for a customer such that there are very stringent requirements for noise or vibration in a vehicle. Driveability of a vehicle is majorly impacted by NVH problems and is strongly felt as vehicle’s performance during various manoeuvres such as acceleration, braking, gear change etc. Thus vehicle manufacturers and OEM’s have realized the importance of NVH issues and target to identify the root cause of vibration and incorporate the correcting measures at early stages of manufacturing with a main goal to refine the ride and driveability of a vehicle.

These days’ efforts are being made to increase the performance of vehicles and to make them more fuel efficient by using different techniques and one of the techniques is electrification of various systems in a vehicle. Electronic power assist steering system, electric active roll control, electronic differential, brake by wire are some of the examples of electrification of vehicle systems. This kind of methodology becomes more evident in case of hybrid electric vehicles where electric power components including power electronics are integral part of the powertrain. It has been observed that using these electrification techniques not only helps in increasing the performance of the vehicle but also makes them more fuel efficient. Thus the electrification in a diesel hybrid powertrain to improve the driveability of vehicle forms the main platform for this research project.

Problem Statement

One of the major NVH issues in automobiles is torsional vibrations of driveline. The torsion of drivetrain at high engine torques causes oscillations in driveline. These oscillations are generally excited by various sources such as rapid or sudden load/torque change, gear change and unevenness of road surface. These oscillations cause longitudinal acceleration in vehicle structure and passenger compartment of the vehicle, which reduces the comfort and driveability of the vehicle and are experienced as unwanted jerking by the driver and passengers. These type of driveline oscillations are termed as Jerking, which is the main focus of this research.

The high engine torques at low engine speeds makes this phenomenon more common in modern diesel engines. A sudden/abrupt pressing or releasing of accelerator pedal by the driver causing a rapid increase or decrease in engine torque called as tip-in and tip-out respectively, which gives rise to oscillations between frequency range of 2 – 10 HZ. To reduce these oscillations and to increase the comfort and improve the driveability of vehicle some kind of ‘anti-jerk control’ is required. In conventional diesel engine vehicles, active control of engine fuelling and driver’s torque demands and passive utilization of mechanical isolators such as dual mass flywheel (DMF) are used to permit feed forward active and passive damping of these oscillations.

Aims and Objectives

The aims of the research project are to investigate the phenomenon of jerking in a diesel hybrid electric vehicle and to electrify the anti-jerk control using the electric motor of hybrid vehicle to damp these oscillations. The aims of the project would be realized through following objectives:

A feed forward approach in modeling the vehicle driveline to investigate driveline jerking using SIMULATION-X. The response of vehicle during tip-in, tip-out, idle and coast conditions would be studied.

To design an active feedback control for electric motor of the hybrid vehicle to permit active damping of the investigated driveline oscillations.

To analyse its effects on fuel consumption, comfort and sportiness of the vehicle and comparison with conventional anti-jerk control methods.

To study the impact of passive damping of these oscillations using DMF in the driveline.

Due to the diverse nature of the project the work has been divided between two students. It was decided that modelling of the driveline will be done by Pavan Mukkamala and Kamal Shamnani will take care of control design part of the project.

Report Overview

As it can be seen before, the report begins with the introduction to the current scenario of hybrid vehicles, driveability and electrification of vehicle systems. Then problem statement was defined followed by aims and objectives.

The next chapter would summarize the undertaken review to understand the research work beginning with analyses of jerking. The literature review also serves as the basis to provide the expected response characteristics. It also explains the different control methods used on a conventional diesel engine vehicle.

Chapter 3 lays down the outline of the procedure to be followed for the research project along with project plan. Conclusions based the literature review have been made in chapter 4.

LITERATURE REVIEW

Driveline Oscillations

Different type of powertrain configurations (front wheel drive or rear wheel drive), engine type and dynamic response, engine torque at different engine speeds and vehicle mass excites different kind of oscillations in vehicle driveline. These oscillations have been summarized below [1]:

Torsion of shaft causes jerking and pitching type oscillations in the driveline. Jerking then causes longitudinal acceleration of car body in a frequency range of 2 – 5 Hz. The pitch motion frequency is about 1 – 2.5 Hz.

Engine movement and engine mount characteristic excites oscillations in a frequency range of 15 – 200 Hz. Gear box housing cause oscillations in the frequency range between 50 and 80 Hz.

Combustion process, movement of pistons inside the cylinder and engine speed causes oscillations depending upon type of engine, number of cylinders and engine speed. Inline-four cylinder engine types engine exhibit large cyclic fluctuations of torque at idle speed.

Driveline Jerking in a Diesel Engine Vehicle

The fast response and high torque of modern direct injection diesel engines results in sharp and large changes in driveline torque. Moreover smaller and more powerful modern diesel engines with reduced flywheel inertias and reduced mass of powertrain components reduce the initial response to transient throttle demands by the driver. But on the other hand faster response and large changes in driveline torque make them more susceptible towards generation of driveline oscillations. These types of oscillations are commonly referred as ‘driveline jerking’, ‘surging’, ‘driveline shuffle’, ‘driveline judder’, ‘jerk oscillations’ or ‘bonanza effect’. [2], [3] and [4].

These oscillations are generally excited due to the torsional vibrations of driveline when accompanied with sudden torque change. The sudden driveline torque change can be described as tip-in and tip-out manoeuvres. The tip-in and tip-out behaviour are described in figure-1 as a step input and after few seconds a step back to zero [2]. The tip-in and tip-out behaviour causes oscillations to engine speed. The difference between engine speed and wheel speed is generally used to describe the torsional oscillations of driveline shown as torsional speed in figure-2-1. These oscillations are transmitted to the vehicle body from wheel and tire and cause oscillations in vehicle longitudinal acceleration. The oscillations in longitudinal acceleration of vehicle body are shown in figure-2-2. The acceleration of vehicle body oscillates between the frequency range of 2 and 5 Hz. The observed frequency of oscillation is dominated by driveline’s first natural frequency. This frequency lies between the resonance frequencies of various sensitive human body parts such as shoulders (4 – 6 Hz), stomach (4- 8 Hz) and trunk (3 – 6 Hz) [5]. So, these different body parts may resonate due to this jerking and hence very unpleasant for drivers and passengers.

Figure 2: Effect of Tip-in and Tip-out on Engine and wheel speed [2]

Figure 2: Body acceleration during Tip-in and Tip-out [2].

The phenomenon of jerking is more common where the manual clutch is not present, as a human driver uses manual clutch to reduce these vibrations. Whereas in automatic clutch units the design philosophy is to reduce the clutch engagement and disengagement time to prevail smooth torque transfer which gives rise to driveline oscillations.

These oscillations differ at different gear ratios. It has been shown that oscillation frequency increases shifting from first gear to higher gears due to change in damping co-efficient associated with each gear and reflected engine inertia [6]. Generally driveline shuffle is also associated with phenomenon of clonk or gear rattle. Clonk is basically a noise which is a by-product of backlash present in gears. Backlash is allowed to incorporate operating clearances and manufacturing tolerances. The clonk can be reduced by increasing the driveline compliance but side effect of increasing compliance is the increased driveline oscillations hence jerking.

Anti-Jerk Control of Conventional Vehicles

The main parameters which govern the severity of jerking response of vehicle can be termed as engine torque rise rate and driveline compliance. Anti-jerk control acts as the balance between comfort and sportiness of the vehicle. The reduction in driveline oscillations is possible by controlling the engine torque rise rate, which can be done by smoothening of driver’s torque demands. In a conventional diesel engine vehicle it is done by electronic fuelling control and in a petrol engine vehicle it is done by using spark advance and electronic throttle control. Various manufacturers use the filtering of driver’s torque request but this becomes of no use when oscillations are excited by ground roughness.

In practice, anti-jerk controller parameters are experimentally optimized to combine comfort and sportiness of the vehicle [2]. The parameters are tuned and accessed, to include non-linearities present in the system, by using various test-runs to find optimal controller parameters. All these parameters are stored in form of different maps and relationship between different engine parameters which are then used to obtain a feed-forward controller for the vehicle.

The reductions in driveline oscillations in conventional drivetrain are possible by various active and passive methods. These include:

Optimizing driveline parameters such as compliance of clutch, driveshaft and transmission, gear ratios and inertia of driveline components.

Passive reduction of oscillations by adding damping to the system.

Active control of engine torque.

Model Based Anti-Jerk Control

In past these parameters were obtained by using trial and error methods which may not be fully optimized and this strategy takes a lot of time depending upon the knowledge of automotive engineer [2]. These days as explained above that optimal anti-jerk control parameters are obtained experimentally by systematic and standardized computer aided test runs. This procedure still relies on time-consuming test runs. Whereas use of simulation can help in analytical assessment of a large number of performance, driveability and comfort parameters with reduced costs and less risk of testing. Simulation can determine highly optimized and accurate system parameters out of which only few parameters are required to be checked by real tests which save a lot of testing time. Due to this a lot of emphasis is given on use of simulation to obtain the optimal controller parameters and also if possible integration of model based controller in the vehicle is also being considered. A review of model based controller design approach is shown below.

Modelling techniques

Driveline of a conventional diesel engine vehicle is made up of an internal combustion engine which is connected to a gearbox through clutch whose output is connected to driveshafts which transmits torque to the wheels. In model based controller design gray-box approach is used to model the drivetrain. According to which driveline can be modelled as lumped parameter model which is made up of rotating bodies with lumped inertias, compliances, damping losses, input torque and load torque through environmental forces [6]. Even a complete linear state space model based upon this relatively simple transformation of driveline is of 14th order which is very complex and requires very extensive computational resources to run on a real time engine ECU and is therefore very expensive and cannot be implemented for mass production [4]. Structure of a complex rear wheel drivetrain is shown in figure -3 [4].

Figure 2: A rear wheel drive vehicle driveline structure [4]

Thus it is required to define a reliable and low order linear representation of overall driveline dynamics. Since the purpose of anti-jerk controller is to damp the oscillations at driveline’s first resonance mode, some approximations can be made to reduce the model complexity. To simplify the model it can be assumed that driveline oscillations at first resonance mode are caused due to torsion of drive shafts which can be explained as they are subjected to relatively highest torque since the torque from engine is amplified by the transmission and final drive [1], [7]. Clutch due to its higher stiffness can be considered as not contributing to first resonance mode. All these assumptions can be justified by measurements [1], [7]. This reduces the driveline model to a 2nd, 5th or 7th order depending upon the type of drive configuration and flywheel.

For a reliable and accurate design of a model-based controller identification of model parameters is done on basis of measurement data. To incorporate the non-linearities present in driveline the values of parameters depending upon engine speed can be stored in form of look-up tables resulting in time-variant model which is more accurate and reliable [1], [7]. Different techniques such as least square optimization and extended kalman-filter can be used to optimize the non-linearities present in the system to reduce the error. The model can be validated by comparing the simulation results with real behaviour of the car.

Controller Design

In a model based anti-jerk controller design the controller is used in a feedback path. Here also different methodologies have been used to design the control strategy. One of them is to use the model to obtain the controller parameters. Thus in this case model based controller design approach is used to assist the engineer and to save time on testing [2]. This type of control topology is shown in figure 2-4 [2].

Figure 2: Model based designed controller structure [2]

In this type of control architecture a two step procedure is used to damp the driveline oscillations first the requested driveline torque Tin is filtered to prevent jerking generation and then the feed-back controller minimizes the jerking further. State of the vehicle can be the feedback to the controller which can be engine speed, wheel speed, and vehicle speed etc, based on which variable controller parameters will be chosen from predefined maps. Advantage of this type of control structure is that it can be used to compensate for the disturbances from the ground roughness. The other advantage of this type of control strategy is that it requires only controller variable parameters to be stored in the ECU of the engine which will be extracted by two performance criteria i.e. sportiness of vehicle demanded by the driver and level of comfort required through no longitudinal oscillations [2]. The disadvantage of this type of controller is that it needs to store different parameter values for different gears and operating points which require the experience and knowledge of automotive engineer to adjust parameter maps. Also it may require some feedback of vehicle states from the vehicle which may be difficult to measure in a vehicle. Also the adjusted parameters values for comfort and sportiness will depend upon the subjective preferences of the engineer. A field of research in this case is the invention of tuneable factor between sportiness and comfort which enables customers to tune their car to their requirement [2].

Another approach in model based controller design is the predictive approach i.e. a reduced model of vehicle drivetrain will be included in vehicle ECU which will reproduce the occurrence of jerking before its actual occurrence and output of which is used as controller input which will then avoid jerking before its occurrence [1] , [7] and [4].

Figure 2: Model based predictive control structure [1]

The dead time or the time delay due to engine combustion events is separated from dynamics of process and modelled separately at the end of model which makes the response of the model faster than actual vehicle [1], [7] hence enables the accurate reproduction of oscillations before their occurrence. Structure of a predictive model based controller is shown in figure 2-5 [1]. These types of controllers are generally equipped with an observer which incorporates the external disturbances such as ground surface roughness and compensates the model inaccuracies by including a feedback between model output and actual process output [1] and [7]. The response of observer is faster than that of the process to make the steady state error zero after a short period of time due to this its design is done separately and values for different engine speeds and demanded torques are stored in form of look up tables. Root locus method is used for designing these kind of controllers. An advantage of using this type of controller is that less dependence on automotive engineer to adjust parameter maps. However an appropriate compromise between comfort and sportiness needs to be defined. One of the ways to increase sportiness is to deliberately delay the activation of controller until initial vehicle acceleration. This type of method can give the user an option of different modes for example, sporty, everyday and comfort by changing the controller values [4]. It can also predict system states which are difficult to measure on a real vehicle hence a more accurate system response can be obtained. At the same time this type of controller can be very expensive as it requires significant computational resources added on to vehicle driveline ECU.

Hybrid Vehicle Driveline

A hybrid vehicle is a vehicle in which powertrain has more than one source of power, a hybrid electric vehicle (HEV) has an internal combustion engine and an electric machine/motor. HEVs powertrain can be classified as series or parallel hybrid systems. A series hybrid system is one in which electric motor provides all the torque required for vehicle propulsion whereas a parallel hybrid system allows torque input from either internal combustion engine or electric motor or from both. Hybrid electric vehicles are the most important topic of research for the past few years.

A series hybrid vehicle driveline does not have a direct mechanical link between the engine and driveline. It is connected to a generator which in turn provides electric supply to power electronics. The electric machine/motor is connected generally through a single speed gearbox, final drive and drive shafts to wheels. Thus driveline oscillations in a series hybrid vehicle will be present, similar to a conventional vehicle. The main difference will be because of propulsion mechanism which is an electric motor. Regarding the anti-jerk control for a series hybrid vehicle, it can be done by active control of electric motor torque. It has been explained in a publication by Borodani and Ambrosio [3] in which an active robust control algorithm based on the H∞ control technique was designed for a Fiat Bravo 1600 cc for electric only mode.

In a parallel hybrid vehicle driveline both engine and motor are mechanically connected to the wheels through clutch, an automatic gearbox and drive shafts. The vibration modes of parallel hybrid driveline are similar to a conventional vehicle. A similar strategy can be used to minimize the jerking as used for conventional vehicles in engine only mode. Whereas an active control of electric motor torque can also be used to prevent jerking. As there is not enough evidence of work done regarding the investigation of jerking and development of anti-jerk control for hybrid vehicles, this will be a topic of this research.

Figure 2: Block diagram of hybrid drivetrain to be used

Electric motor in a hybrid vehicle driveline gives an opportunity to electrify the anti-jerk control which then can be compared with conventional methods of anti-jerk control for fuel consumption, degree of comfort and sportiness of the vehicle. A parallel four wheel driven diesel hybrid electric driveline with dual mass flywheel (DMF) of a sport utility vehicle will be used to investigate the driveline oscillation and then to design an active feedback control for electric motor to damp these oscillations. Vehicle drive configuration can be changed to front wheel drive by unlocking the central coupler. The effect electric real axle drive (ERAD) will not be considered for anti-jerk control design. The driveline is shown in figure 2-6. Some important components of driveline and their effect are explained below:

Integrated Starter Generator (ISG)

As name suggests ISG replaces both conventional starter and alternator/generator to a single electronically controlled device. An ISG can convert electrical energy to mechanical energy to start an engine and also mechanical energy to electrical energy to power all electrical systems in a vehicle such as lights, air conditioning etc and to charge battery. An ISG serves three important functions start-stop, electricity generation and power assistance [8]. It allows engine start-stop functionality to save fuel instead of idling when vehicle is not in motion. ISG generates electricity from spinning crankshaft of vehicle to charge the battery. In a hybrid vehicle an ISG can provide power assist to the engine during the boost-mode. Although it cannot propel a vehicle on its own but it can assist the engine and can be used to store energy while regenerative braking. The ISG is generally used in mild hybrids. An ISG can provide retarding force on crankshaft to generate electricity during braking; this feature of ISG can be used for anti-jerk control.

Dual Mass Flywheel

A dual mass flywheel [DMF] consists of two rotating flywheels [primary and secondary] connected by long travel arc-springs [7]. Figure -7 shows the structure of a basic DMF [4]. A DMF can be used to mechanically isolate the driveline oscillations arising from high speed engine oscillations by decoupling the transmission from engine. It filters out the engine irregularities completely. Due to the smooth operation of secondary flywheel and transmission input shaft gear rattle is significantly reduced. DMF can provide good vibration isolation at low engine speeds which in turn reduces the fuel consumption of a vehicle. Due to its excellent vibration isolation and damping attributes, DMF today is found in many vehicle drivelines including low budget cars [4]. But despite of its advantaged DMF increases the system complexity. DMF is basically a spring damper unit having highly non-linear characteristics. It can temporarily store energy due to which it applies a reaction torque on internal combustion engine and alters the engine speed signal which is used by ECU. It increases the probability of undesirable side effects such as jerking.

Figure 2: Dual Mass flywheel structure [4]

PROCEDURE

Procedure of work

Driveline vibration is mainly a problem of engine torque profile and compliances of components in the driveline. To study the effect of jerking in hybrid vehicles a model of driveline with all system complexities will be developed using a SIMULATION-X. The simulation-X model will be a complex model replicating the complete dynamic behaviour of the vehicle. As the frequency range of interest is 2-5 Hz, driveline’s first natural frequency the model will be reduced according to a final model to be implemented for the design on controller which can accurately predict the oscillations in the desired frequency range.

An active feed-back control for electric motor torque and for engine fuelling control will be designed by implementing the reduced model in MATLAB/SIMULINK. Then both controllers will be implemented in the actual complex model in SIMULATION-X to observe the system performance in terms of fuel consumption, level of comfort and sportiness.

Simulation tool

Simulation-X

Simulation-X has been chosen to investigate the phenomenon of jerking for the project due to its very user friendly graphical user interface. This interface provides a very intuitive and efficient lumped parameter modelling approach for modelling dynamic systems. It can model various physical systems (i.e. Electronics, Mechanics, Hydraulics, Pneumatics, Thermal etc) with much ease. It also allows for the signal and equation based modelling ideal for control systems. Another important feature of Simulation-X is the ease of parameterization i.e. it is very easy to define the system complexities and parameters depending upon the requirement. It has numerous possibilities for visualization, analysis and recording a large number of results associated with each element of a model. It is very beneficial for modal analysis as it can easily extract the dynamics of system by showing the modes of vibrations and the cause of these resonance modes. This feature is very handy and useful for this research project as it can be used to differentiate between the causes of oscillations in the driveline and thereby helping to reduce the model for the desired frequency range.

Project Plan

CONCLUSIONS

On the basis of literature review it can be stated that the high change in engine or load torque excites oscillations in vehicle driveline, known as jerking causing horizontal acceleration in vehicle body. The frequency of jerking lies in resonance frequency range of various sensitive body parts as a result it is entirely undesirable by the drivers. Comfort and driveability issues raise the need for anti-jerk control.

There has not been enough work done on anti-jerk control for hybrid electric vehicles since not many hybrid electric vehicles are available in market. Other than the conventional methods, electrification of anti-jerk control can be done in a hybrid electric vehicle which will be the main aim of the research project.

A model based controller design approach can be regarded as efficient and reliable. It can assist an automotive engineer in deciding control parameters for the controller, also it can be implemented directly to vehicle for anti-jerk control by reducing the order of driveline model.

Driveline oscillations can also be damped using passive damping methods such as using DMF which will be studied as a part of research project.

REFERENCES

[1] – Kiencke, U. and Nielsen, L. (2005), ‘Automotive Control Systems for Engine, Driveline and Vehicle’, Springer, Berlin, Germany.

[2] – Grotjahn, M., Quernheim, L. and Zemke, S. (2006), ‘Modeling & Identification of car driveline dynamics for anti – jerk controller design’, ICM 2006 – IEEE 3rd International Conference on Mechatronics , 1-4244-9713-4/06.

[3] – Borodani, P. and Ambrosio, C. (2002), ‘Robust control of active damping of a hybrid vehicle driveline shuffle’, MED2002 Conference.

[4] – Walter, A. Kiencke, U., Jones, S. and Winkler, T. (2008), ‘Anti-Jerk & Idle Speed Control with Integrated Sub-Harmonic Vibration Compensation for vehicles with Dual Mass Flywheels’, SAE International Powertrains, Fuels and Lubricants Congress Shanghai, China, 2008-01-1737.

[5] – McDevitt, I. (2005), ‘Torsional Vibrations of Vehicle Driveline’, MSc Thesis, Automotive Product Engineering, Cranfield University.

[6] – Lee, H.C. (2006), ‘Driveline vibration control of electric vehicle’, MSc Thesis, Automotive Product Engineering, Cranfield University.

[7] – Baumann, J., Torkzadeh, D., Ramstein, A., Kiencke, U. and Schlegl, T. (2005), ‘Model-based Predictive Anti-Jerk Control’, Control Engineering Practice 14 (2006) 259-266.

[8] – Viorel, I., Szabo, L., Lowenstein, L. and Stet, C (2004), ‘Integrated Starter-Generators for Automotive Applications’, Acta Electrotehnica, Volume 45 Number 3 2004.

[9] – Reik, W., Seebacher, R. and Kooy, A. (1998), ‘Dual Mass Fly-wheel’.