Friday, 26 July 2013

Technology focus - Future developments in On-Board Diagnostics

The latest generation of OBD is a very sophisticated and capable system for detecting emission related problems with the engine and powertrain. But, it relies on the fact that it is necessary to get the driver of the vehicle to do something about any problem that occurs. 

With respect to this factor, OBD2/EOBD is no improvement over OBD1 - as there must be some enforcement capability. Currently under consideration are plans for OBD3, which would take OBD2 a step further by adding the possibility of remote data transfer. This would involve using remote transmitter/transponder technology similar to that which is already being used for automatic electronic toll collection systems. An OBD3 equipped vehicle would be able to report emissions problems directly back to a regulatory agency. The transmitter/transponder would communicate the vehicle VIN (Vehicle Identification Number) and any diagnostic codes that have been logged. The system could be set up to automatically report an emissions problem the instant the MIL light comes on, or alternatively, the system could respond to answer a query from a regulator to its current emissions performance status.

What makes this approach so attractive is its efficiency, with remote monitoring via the onboard telemetry, the need for periodic inspections could be eliminated because only those vehicles that reported problems would have to be tested. The regulatory authorities could focus their efforts on vehicles and owners who are actually causing a violation rather than just random testing. It is clear that with a system like this, much more efficient use of available regulatory enforcement resources could be implemented, with a consequential improvement in the quality of our air.

An inevitable change that could come with OBD3 would be even closer scrutiny of vehicle emissions. The misfire detection algorithms currently required by OBD2 only watches for misfires during driving conditions that occur during the prescribed driving cycles. It does not monitor misfires during other engine operating modes like full load for example. More sophisticated methods of misfire detection will become common place which can feedback other information to the ECU about the combustion process, for example, the maximum cylinder pressure, detonation events or cylinder work done/balancing. This adds another dimension to the engine control system allowing greater efficiency and more power from any given engine design just via more sophisticated ECU control strategy.

Future OBD system will undoubtedly incorporate new developments in sensor technology. Currently the evaluation is done via sensors monitoring emissions indirectly. Clearly an improvement would be the ability to measure exhaust gas composition directly via on-board measurement systems (OBM). This is more in keeping with emission regulation philosophy and would overcome the inherent weakness of current OBD systems, that is, they fail to detect a number of minor faults that do not individually activate the MIL, or cause excessive emissions but whose combined effect is to cause the production of excess emissions.

The main barrier is the lack of availability of suitably durable and sensitive sensors for CO, NOx and HC. Some progress has been made with respect to this and some vehicles are now being fitted with NOx sensors. Currently there does appear to be void between the laboratory based sensors used in research and reliable mass produced units that could form the basis of an OBM (On Board Monitoring) system.





Fig 1 - NOx sensors are now in use! (Source: NGK)

Another development for future consideration is the further implementation of OBD for diesel engines. As diesel engine technology becomes more sophisticated, so does the requirement for OBD. In addition, emission legislation is driving more sophisticated requirements for after treatment of exhaust gas. All of these subsystems are to be subjected to checking via the OBD system and present their own specific challenges. For example, the monitoring of exhaust after treatment systems (particulate filters and catalysts) in addition to more complex EGR and air management systems.





Fig 2 - Current monitoring requirements for diesel engines

Rate based monitoring will be more significant for future systems which allows in-use performance ratio information to be logged. It is a standardised method of measuring monitoring frequency and filters out the affect of short trips, infrequent journeys etc. as factors which could affect the OBD logging and reactions. It is an essential part of the evaluation where driving habits or patterns are not known and it ensures that monitors run efficiently in use and detect faults in a timely and appropriate manner. It is defined as…

Minimum frequency = N/D

Where:
N = Number of times a monitor has run
D = Number of times vehicle has been operated

A significant factor in the development of future system will be the implementation of the latest technologies with respect to hardware and software development. Model based development and calibration of system will dramatically reduce the testing time by reducing the number of test iterations required. This technique is quite common for developing engine specific calibrations for ECUs during the engine development phase.

Virtual Development of OBD
Hardware-in-loop (HIL) simulation plays a significant part in rapid development of any ECU hardware and embedded system. New hardware can be tested and validated under a number of simulated conditions and its performance verified before it even goes near any prototype vehicle. The following tasks can be performed with this technology:

Full automation of testing for OBD functionality
Testing parameter extremes
Testing of experimental designs
Regression testing of new designs of software and hardware
Automatic documentation of results



Fig 3 - HiL environment for OBD testing

However, even in a HiL environment, an expensive target platform is needed (i.e. a development ECU). These are normally expensive, and in a typical development environment, they are scarce. In line with the general Industry trend to 'frontload' - it is now possible to have a complete virtual ECU and environment for testing of ECU functions, including OBD, running on a normal PC with a real time environment. The advantage is that no hardware is needed, but more importantly, simulation procedures (drive or test cycles) can be executed faster than real time - this means a 20 minute real time test cycle, can be executed in a few seconds - this has a significant benefit in the rapid prototype phase. See more information about virtual ECU development here: