Scrap and root cause analysis with machine learning method for HDP -High Pressure Pump- in gasoline systems

The project endeavors to address inefficiencies in high-pressure pump production, which result in the generation of scrap products during assembly due to multifaceted factors. This phenomenon leads to the wastage of resources including energy, labor, and materials, thereby escalating costs. Conventional engineering strategies have proven inadequate in resolving these challenges, particularly when errors originate from the interaction of numerous parameters. By harnessing machine learning techniques, the project aims to pinpoint the fundamental causes of these errors and institute preemptive measures. The process will involve data collection, purification, and analysis through the utilization of Bosch's Manufacturing Analytics Solution (MAS) within the Microsoft Azure Databricks framework. Support vector machines and decision trees will be employed to detect error origins, with the objective of reducing scrap rates by 30%. This innovative methodology is anticipated to bolster research and development capabilities, enhance product quality, and curtail resource wastage, thereby fortifying the company's technological proficiency and adaptability.