Inferential sensing

Engineers propose an onboard electronics hardware and software architecture that can deliver condition-based maintenance solutions in a low-cost platform.

OFF-HIGHWAY EQUIPMENT OPERATORS traditionally have had a dual role: safely operate the machine and diagnose problems. In the past, the operator was closely connected to the machine and had a fundamental understanding of the components to be able to infer the machine's health and condition. Their “open station” position gave them an intimate connection to all the machine's noises, smells, vibrations, sights, and sounds. However, as machine design has evolved, the operator is being isolated in comfort controlled cabs and active isolation seats. New systems are needed to replace the operator's monitoring and intuition skills to fill the void for determining the health and condition of the machine.

Today's machines include nearly 100 sensors for onboard electronic controls. Few of them are used primarily for diagnostics. Diagnostic sensors typically require higher accuracy and faster response time than sensors used for control. There are very few ways to directly sense the onset of problems or the conditions leading up to problems. Due to the added “diagnostic-only” cost and the limited capability to sense the onset of problems, these electronic diagnostics offerings are seldom offered on standard equipment. The conditions, and potential problems, are best inferred from other direct measurements and data-processing methods to avoid the added cost and complexity of dedicated diagnostic sensors. This is the premise of inferential sensing.