Last updated on October 4th, 2023
USE CASE
Optimizing new EV part production
Objectives
- Get new EV component manufacturing lines up to stable production quickly
- Reduce early risk of scrap and rework
Challenges
- Pressure is high to install new EV part manufacturing lines
- Troubleshooting issues for new EV components is complex since new products lack pre-existing standards and methods
Key Results
Automated root cause analysis with AI helps to solve problems quickly
Stable production achieved quickly and major quality spills avoided
Background
Between budgeting, design, equipment selection, and workforce planning, launching a new production line for the manufacture of electric vehicle components is a complex process, whether this be for a greenfield site or a retrofit in an existing facility. Tier-1s and OEMs are undergoing this process globally as they prepare to manufacture newly designed components like EV motors, battery packs, inverters, and charging ports.
The pressure is high for these new lines to reach stable production and meet the demand backlog for new electric vehicles. Manufacturers who produce parts efficiently and with a high degree of quality will have a competitive advantage to meet this demand.
Setting up a new manufacturing line for EV components comes with unique challenges. With newly designed products, there is limited knowledge about the most efficient way to produce parts and what defects may arise. In these situations, insights from the analysis of the manufacturing data are especially valuable in providing clues and context to help improve quality and First Time Through (FTT) yield.
Collaborating with System Integrators
When a production line is designed by a system integrator, a series of acceptance tests is required to ensure that the system is capable of producing good parts, and that it meets both functional and visual requirements. A small run of parts is produced to ensure quality.
However, the teardown, shipping, and setup of the system onsite can introduce several variations. It may be challenging to replicate the initial production conditions in a new location, particularly one with larger production volumes. Once the production line has been set-up onsite and further testing completed, there is aways the risk of things operating differently than expected, and defective parts being produced.
To ease the transition from design to onsite setup, manufacturing data from the initial runs and acceptance tests at the integrator can be analyzed in LinePulse, alongside the onsite runs and tests, to identify any variations in the process. If there are inconsistencies between the two setups, LinePulse can quickly identify which signals are indicating a difference.
Since LinePulse is a cloud-based software, the integrator and manufacturer can easily add their respective data without the limitations of on-premises systems. Previously, an incongruity between these two tests could be challenging to solve, with both the integrator and manufacturer reluctant to take responsibility for the issue.
Instant predictive insights into manufacturing data
Setting up LinePulse with a net new EV manufacturing line allows easy visualization of manufacturing data and a deeper understanding of the relationship between different parts of the process, even before defects or problems occur.
Several different modules in LinePulse can provide valuable information:
- Single-variate anomaly detection can show which signals have significant variation, and which ones are changing over time.
- Multi-variate anomaly detection can be set up to monitor if a group of signals are affecting each other, or if one process is affecting another downstream.
- Capability metrics can be automatically generated for any time period to measure overall performance and capability of the process.
- SPC charts can also be viewed from the LinePulse platform to monitor signals against fixed control limits.
Troubleshooting production problems
LinePulse can be used to accelerate root cause analysis when a test has failed. The root cause analysis module in the software instantly generates a list of statistically probable contributing factors to the specified failure mode, based on the available signal data. This prioritized list gives engineers a head start on their investigation to find the root cause of the problem and solve it.
Engineers can also use their domain knowledge to eliminate high-correlation but low-causation signals from the list to further increase analysis accuracy.
With this function, any initial failures that may occur on the newly set up line can be investigated quickly and thoroughly, helping the process reach stable production as soon as possible.
Results
By using LinePulse on a new EV manufacturing line, engineers are empowered with instant insights into their manufacturing data. Initial problems in setting up lines or finding the root cause of defects can be solved quickly. Manufacturers can rely on the data to guide them towards efficient, stable production, even when dealing with newly designed products.