Jiamei Deng

An Explicit Model Predictive Control Framework for Turbocharged Diesel Engines

The turbocharged diesel engine is a typical multi-input multioutput system with strong couplings, actuator constraints, and fast dynamics. This paper addresses the exhaust emission regulation in turbocharged diesel engines using an explicit model predictive control (EMPC) approach, which allows tracking of the time-varying setpoint values generated by the supervisory level controller while satisfying the actuator constraints. The proposed EMPC framework consists of calibration, engine model identification, controller formulation, and state observer design. The proposed EMPC approach has a low computation requirement and is suitable for implementation in the engine control unit on board. The experimental results on a turbocharged Cat C6.6 diesel engine demonstrate that the EMPC controller significantly improves the tracking performance of the exhaust emission variables in comparison with the decoupled single-input single-output control methods.

Input Constraints Handling in an MPC/Feedback Linearization Scheme

Input Constraints Handling in an MPC/Feedback Linearization Scheme The combination of model predictive control based on linear models (MPC) with feedback linearization (FL) has attracted interest for a number of years, giving rise to MPC+FL control schemes. An important advantage of such schemes is that feedback linearizable plants can be controlled with a linear predictive controller with a fixed model. Handling input constraints within such schemes is difficult since simple bound contraints on the input become state dependent because of the nonlinear transformation introduced by feedback linearization. This paper introduces a technique for handling input constraints within a real time MPC/FL scheme, where the plant model employed is a class of dynamic neural networks. The technique is based on a simple affine transformation of the feasible area. A simulated case study is presented to illustrate the use and benefits of the technique.

Diesel engine emissions prediction using parallel neural networks

Emission legislation has forced the pace of development of engine management functions. Legislation that will be applied to diesel engines during the period 2010–2020 continue to put great emphasis on both nitrogen oxides NOx and particulate matter (PM). With the increasing effort to reduce emissions and maintain fuel economy manufacturers are focussing on engine control. Engine control requires data acquisition and acquisition requires sensors, but hardware in the form of sensors adds further cost to the production. As a result, so called virtual sensors are introduced. These are estimators that predict the required data, which is costly to measure or simply incapable of measurement. In this paper a parallel neural network structure is built. It consists of three Non-linear autoregressive exogenous input (NLARX) neural network models used to predict the smoke emissions of a diesel engine operated in a Non-Road-Transient Cycle. Existing resources from Matlab toolboxes are used in order to monitor both the cost and computational expenses of analysis. The data is re-ordered into training and validation sets and processed. To overcome the weakness of the neural network approach in respect of high frequency signals, the data is divided into layers to split up the frequencies and cut high amplitudes. Three horizontal layers of the signal are processed in parallel through independent NLARX-models and their performances are added to give an overall result.

Particulate matter prediction in both steady state and transient operation of diesel engines

Diesel engines produce a variety of particles generically classified as diesel particulate matter (PM) owing to incomplete combustion. The increasingly stringent emissions regulations require that engine manufacturers must continue to reduce the PM. The ability to predict the PM emissions is one of the key technologies that could be used in a PM reduction strategy. This paper describes a predictive technique that can be used as a virtual sensor for monitoring PM emissions in both steady and transient states for a medium- or heavy-duty diesel engine. The predictive structure is stable over a broad range of engine operation points. The input parameters are chosen on the basis of the PM formation mechanism, physical knowledge of the process, and an insight into the underlying physics. Principal-component analysis (PCA) is used to reduce the dimensionality of the inputs of a non-linear autoregressive model with exogenous inputs (NLARX) from nine inputs to five inputs. PCA not only reduces the input number but also improves the performance of the prediction model. The results show that the NLARX model could predict the particulate matter successfully with an R2 value above 0.99 with only five inputs.

Explicit model predictive control on the air path of turbocharged diesel engines

The turbocharged diesel engine is a typical multi-input multi-output (MIMO) system with strong couplings, actuator constraints, and fast dynamics. This paper addresses the air path regulation in turbocharged diesel engines using an explicit model predictive control (EMPC) approach, which allows tracking of the time-varying setpoint values generated by the supervisory level controller while satisfying the actuator constraints. The proposed EMPC framework consists of calibration, engine model identification, controller formulation, and state observer design. The proposed EMPC approach has a low computation requirement and is suitable for implementation in the engine control unit (ECU) on board. The experimental results on a turbocharged Cat® C6.6 diesel engine illustrate that the EMPC controller significantly improves the tracking performance of the exhaust emission variables against the decentralized single-input single-output (SISO) control method.

Application of constrained predictive control on a 3D crane system

This paper describes the SIMULINK implementation of a constrained predictive control algorithm based on quadratic programming and linear state space models, and its application to a laboratory-scale 3D crane system. The algorithm is compatible with Real Time Windows Target and, in the case of the crane system, it can be executed with a sampling period of 0.01 s and a prediction horizon of up to 300 samples, using a linear state space model with 3 inputs, 5 outputs and 13 states.

Predictive computed-torque control of a puma 560 manipulator robot

This paper describes the integration of constrained predictive control and computed-torque control, and its application on a six degree-of-freedom PUMA 560 manipulator arm. The real-time implementation was based on SIMULINK, with the predictive controller and the computed-torque control law implemented in the C programming language. The constrained predictive controller solved a quadratic programming problem at every sampling interval, which was as short as 10 ms, using a prediction horizon of 150 steps and an 18th order state space model.

A review of driver modelling

Increasingly accurate vehicle simulations are required by automotive manufacturers and researchers in order to develop new ideas while minimising the use of costly prototype vehicles. The development in vehicle simulations has been the focus of much research; however the progress of driver modelling development for use in these simulations has not been as rapid. In areas such as fuel economy it has been shown that the behaviour of the driver plays a significant role and as such a vehicle simulation used to investigate fuel economy should include a driver model that can mimic different driver behaviour. Historically fuel economy simulations have been undertaken with a reference speed profile followed accurately with classical closed loop control and this approach will neglect some of the characteristics of real world driving that come from having a human driver. In this paper the development of driver modelling will be summarised up to the current state of the art and further applications for driver models are detailed.

Fuel path control of a diesel engine at the operating point of the low load and medium speed

Current control implementations for engines are proving unwieldy for emerging emissions standards and fuel economy demand. Calibration is becoming progressively more complex as the number of controlled variables increases. The issues are acute with diesels. We describe a project in which a detailed investigation of the fuel path dynamics in a modern engine is made. This is an initial work about diesel engine fuel path control. In order to facilitate the control development, a medium speed (1550 rpm) and low torque (250Nm) point is chosen to develop the control strategy as it is believed that this is the safe point to start with the fuel path work. The development of fuek path control is difficult as the nature of fuel injection parameters affects the whole engine performance significantly and quickly. This paper demonstrates a closed-loop control and an architecture of controllable injection for C6.6 engine based on predictive control that could control exhaust temperature, of nitrogen oxides (NOx) and particular matters (PM) without changing the fuel quantity at medium speed and low load point.

Combined hybrid clustering techniques and neural fuzzy networks to predict diesel engine emissions

This paper presents a neural fuzzy modeling approach based on hybrid clustering technique to predict a diesel engines NOx emissions. A hybrid clustering algorithm is provided. Since the combustion process is very complicated, therefore, it is almost impossible to find a simple and accurate first principle model to predict diesel emissions. Black-box models implementing Artificial Intelligent Techniques must be developed. Fuzzy modeling seems to be one of the most suitable approach for modeling diesel emissions with big oscillations and high frequency. Clustering is used with fuzzy modeling approach for determining fuzzy if-then rules, so that a fuzzy network, trained with back propagation, adjusts the centers and widths of the membership function. This paper uses hybrid clustering techniques to build a neural fuzzy model successfully. The results show that the model has very good accuracy in predicting diesel engines NOx emissions.