Engine control

Conventional engine control, founded on open loop actions based on static maps and associated with monovariable regulators, is now unable to adequately meet the needs of new gasoline or diesel combustion modes.

IFP Energies nouvelles's engine control offer is based on multivariable control techniques and observers (estimated from non-measured variables), making it possible to correctly and specifically manage transitions, i.e. both sudden variations in torque demand, for example, and slow changes linked to ageing or fouling of actuators and catalysts. The observers developed at IFP Energies nouvelles are preferentially based on simplified knowledge models, known as 0D models, which have been reworked to become totally real-time.

Engine control constraints are increasingly numerous. In order to reduce pollutant emissions, use is made of:

• new combustion techniques (HCCI, CAI, etc.)
• optimization of energy sources (e.g.: thermal/ electric);
• technological air supply solutions (turbochargers, re-injection of burned gases, etc.), mecatronic actuators (variable control valves, very high-pressure diesel injectors, etc.);
• technological after-treatment solutions (4-way catalyst, urea injection, etc.).
   

But the combinations and compromises are multiple among these different types of “remedies” (combustion/energy sources/technologies, etc.), hence the need for control techniques for overall management of the complex system.

More specifically, new “clean combustion” processes pose genuine challenges in terms of engine control if they are to work, since the objective is to obtain stable torque production with low pollutant emissions at source, via control, despite significant constraints (poor mixture, low temperature, variable air/fuel ratio, etc.).

In addition to solutions to reduce pollutant emissions, innovation in terms of powertrains involves the increasing addition of inter-dependent subsystems, each responding to different dynamics. The choice of control strategy takes into account transient loads, which are deliberately “discrete”, actuators that need to be accurately piloted, an increasing number of measurements, etc., at the same time ensuring the best possible drivability.

To be able to propose innovative control laws, system simulation becomes a crucial element. First of all, in the very methodology of the design, devising and testing the control unit on simulators, but also in its formalization, increasingly using models representing the dynamics of the system to be piloted within control units.