30 credits - Measurement identification for test environment accuracy i Sodertalje

Thesis project at Scania is an excellent way of making contacts for your future working life. Many of our current employees started their career with a thesis project.

Background:
Testing software features of complex powertrain systems require different test environments, one of them being hardware-in-the-loop (HIL). A central part of the HIL environment is the models that simulate the outside world and process the inputs and outputs from the system under test (SUT); often an electronic control unit (ECU). There are often multiple models that are collaborating to produce a correct output to be read by the ECU and hence the complexity of the test environment can come close to that of the actual product it tries to simulate.

With growing complexity it becomes harder to establish the conformity of the test environment with the real component/system given that both use the same initial conditions.

Target:
Determine a set of measurement points (on both a model and on a real component/system) that are necessary for establishing conformity between a model and a real component/system without over- or underfitting the model.

In addition, a set of stimuli (e.g. step responses etc.) have to be decided upon, together with the measurement set, in order to excite both model and the real component/system in such a way that the measurements produce valuable information that captures the component’s/system’s dynamic behaviour/features.

For instance, in order to model the electric machine with an accuracy of 80%; torque, rotational speed, physical load, electrical current and voltage must be monitored. A step response from initial condition  with stimuli  is required to capture the desired dynamic behaviour with sufficient accuracy.

Assignment:

1. Investigate the state of the art of model design/implementation and -verification methods.
2. Answer how the measurement/sample space should be derived for real-time models.
3. Answer how the set of stimuli necessary to elicit the dynamic behaviour should be derived for real-time models.
4. For a limited set of models, design a method for calculating what and how many sampling points, according to step 2, are necessary to reach a certain model accuracy and what stimuli is necessary to verify the conformity of the model, according to step 3.
5. Verify method designed in step 4 on a real model against a real component/recorded data.

Education:
Specify education or specialisation: E, D, F, Y or similar faculty
Number of students: one or two students
Suitable start date: January 2019
Estimated time needed: 20 weeks

Contact persons and supervisors:
Milivoj Lärfars Persson, supervisor, +46 8-553 890 26,

Johan Kingstedt, group manager, +46 8-553 896 72,

Enclose CV, personal letter and school-leaving certificate

• Title: 30 credits - Measurement identification for test environment accuracy
• Business area: Research and Development
• Country: Sweden
• City: Sodertalje
• Last application date: 2018-11-30
• Job Id: 20182165