File:CONDITIONS-BASED MAINTENANCE THROUGH AUTONOMOUS LOGISTICS (IA conditionsbasedm1094562712).pdf

Currently, operators and maintainers cull though numerous electronic reports, display boards, and historical maintenance records to determine and plan for maintenance activities for equipment. However, in recent years, emerging technologies such as the Internet-of-Things, big data analysis, and low-cost sensors and actuators have enabled applications that were not possible previously. From these developments, information that was once unavailable is now accessible through embedded sensors and actuators, providing real-time condition monitoring of complicated machinery. This thesis demonstrates the use of inexpensive COTS hardware devices and open-source software to develop an automated data collection architecture and a data processing framework to implement a preventative maintenance approach for the Marine Corps Medium Tactical Vehicle Replacement (MTVR). Data processing techniques were used to convert raw sensor data collected from on-board MTVR sensors into useable and measurable diagnostic data. Using statistical analysis based on a time series regression model, the diagnostic parameters that closely modeled engine operating conditions were chosen to predict engine usage characteristics of an MTVR engine. The thesis also describes a conditions-based maintenance policy that can be used to enhance preventative maintenance methods and decision support capabilities on Marine Corps ground equipment.

Subjects: data science; single-board computers; logistics; MTVRS; CBM; conditions-based maintenance; sensors; Raspberry Pi; J1939; SAE