Various industry projects at ITWM deal with improving the collection, evaluation and simulation of vehicle usage by different customers. The methods under development use the full range of modern information and communication technology. This includes vehicle-independent sources like navigation maps and weather forecasts, but also vehicle-based measurements via dedicated sensors or on-board electronics (CAN-bus, electronic control units).

• A better understanding of loads and consumption can lead to improvements in many areas:
  • Development: new vehicle models can be better adapted to their intended usage already during the design stage.
  • Sales advice: customers can be advised more accurately as to which vehicle best fits their requirements.
  • Fleet management: customers with a fleet of vehicle can plan better, which machine should be assigned to which job.
  • Intelligent control units: vehicles can automatically optimise fuel consumption or warn the operator that certain components are wearing out.
  • Driver assistance: the vehicle can support the driver in working more efficiently or reducing component fatigue.

Projects

• Measurement box for on-line collection of vehicle data
• Improved efficiency for commercial vehicles
• Virtual measurement campaign
• Statistical evaluation of large data sets

It is often impossible to determine realistic customer loads from field measurements alone, as these are relatively short. Methods from the field of extreme value statistics and non-parametric statistics are required to extrapolate to the full service life and to more extreme events.

Another issue is the large number of possible applications, which means that the proportion of certain events in the measurement campaign does not always match customer usage. The data can still be used effectively if it is classified according to cause variables, as the measurements can then be recombined into different mission profiles.

In case such a classification scheme is based on external factors (environment, task, etc.), one obtains a system-independent description of customer usage that can be used to evaluate the behavior of variant systems under comparable conditions. This approach is not limited to load data analysis; another application is e.g. estimating fuel consumption.

Before it can be used for fatigue testing, the customer load distribution needs to be transformed into a reference load. Here, it is important to consider that one only has reliable information about the duty stress and component strength for typical cases, but is really interested in hedging against extreme events (component failure during service life).

Given some reference load, methods from the fields of optimization and statistics can be employed to translate it into a specification for a proving ground or test rig. The load needs to be replicated accurately, but the test should also be efficient in terms of low cost and short runtime.

Projects

• Derivation of gear damage from field data (GM)
• Load data analysis and simulation for excavator development (VOLVO)
• Assessment of customer usage profiles and cause variables
• Optimal track mixing
• CAROD: Analysis and simulation of time series

When planning a test, a typical issue is whether to test few samples for a long time or more samples for a shorter time. And a common question one faces when evaluating tests is how to handle censored data properly, in case a run was stopped before the component under investigation showed any signs of wear.

To aid in the planning and evaluation of durability engineering the software package Jurojin has been developed. It permits e.g. the balancing of number of samples vs. run length or the adjustment of a test plan during testing. More advanced applications include comparisons with customer loads and the assessment of parts retrieved from customers.

Products

• Software Jurojin