Case Study

Case Study

Overview
KPIT helps an Automotive Tier 1 to reduce Engine ECU throughput by 12%
About the Client
The client is a Tier 1 committed to the design & production of hi-tech systems and components for the automotive sector. They have a global presence with 80+ production units and 10+ R&D centres supplying to most of the car makers in Europe, North and South America and Asia.
 
Business & Technology Context
An ECU functions properly if the CPU load is around 80-85% even at its peak usage (high speeds). When the CPU load nears 100% during normal speed, ECU would not be able to function properly at high speeds. While operating at high speeds, high CPU load can result in less time available for processing tasks leading to increase in throughput time. This would lead to queuing up of tasks and performance degradation of the ECU. Hence arises the need for throughput optimization.
The client was developing an Engine ECU for a production program of an Automotive OEM. The CPU load had crossed 90% when the ECU was about to get into production program. Therefore the client was looking for a partner who could help them to optimize the throughput while maintaining the RAM & flash memory usage within acceptable limits.
The client was facing stringent delivery timelines for the ECU and hence consulted KPIT for performing design optimization for their models that could reduce the execution time.
The client shared models for 9 modules and KPIT had to come up with a solution for optimizing the throughput based on these modules.
KPIT's Solution
This was a one-of-its-kind situation and KPIT successfully reduced the ECU throughput by 12% within a span of 3 months.
 
KPIT created a joint IP in HMI development framework- ‘Kwik framework’ for the client. Apart from the high level design & detailed HMI framework, the scope of the work included HMI (QML) & HMI Model development using (Qt/C++), business Logic development & middleware integration. ‘Kwik’ HMI was deployed to improve the overall reusability, productivity, maintainability and quality.
Key Highlights


 
Incremental Design Changes
Control Algorithm Re-design
Load Distribution Analysis
 
  • Simple Validation Cycle
  • Completed in isolation without much involvement of other groups
  • Architecture framework can be designed & deployed which will lead to discipline in design
  • Module design guideline can be standardized
  • Long term benefits from Maintenance & Re-usability standpoint
  • Good chances of overall Throughput improvement
  • Significant improvement in throughput with no development effort
  • Existing Validation methodology can be used to verify the changes
  • CPU Load is evenly balanced
 
 
Success Factors
The client appreciated KPIT's commitment and R&D investment in developing 3 approaches instead of 1 for throughput optimization
The client expected KPIT to assess the CPU load only at model level. For this, KPIT delivered the combined implementation of 'Incremental Design Changes' and 'Control Algorithm Re-design'. This combined implementation was able to deliver 12% throughput reduction in desired timeframe
KPIT also went one step further and performed scheduler level optimization using load distribution analysis and regression testing on HIL test bench. This helped in enhancing the throughput and balancing the CPU load across the execution frames.
Tags: powertrain, Engine ECU, throughput optimization, CPU load, throughput, load distribution analysis, incremental design, control algorithm, scheduler level optimization, regression testing, model design, throughput reduction

Terms of Use   Privacy Policy   Cookie Policy   Home   
Copyright © 2018 KPIT® Technologies Ltd.

CIN: L72200PN1990PLC059594