SAGA:prioritize maintenance and repair

Application Domain: Equipment maintenance

Purpose: To prioritize maintenance and repair actions.

Features: This customer has to repair hundreds of broken systems (mechanical and electronic equipment). The main problem is deciding the order in which the equipment should be repaired. Repairs are achieved along three different steps: failure diagnosis, subcontract of the repair and in-house repair. Three expert systems are scoring the broken systems, depending on their status, to rank them and decide which ones will be diagnosed, sub-contracted or repaired in-house. Rules are based on MTBF and MTTR, spare part availability, human resource and repair equipment availability and criticality of the need for the system in the field. An evaluation of SAGA’s performance demonstrated that its rankings were as good as those made by the best maintenance managers 85% of the time. The remaining 15% resulted from circumstances unknown to the system.

CallAid: real-time monitoring of automation

Application Domain: Real-time monitoring of an automated factory.

Purpose: Factories are usually automated through networks of PLCs and their associated computerized supervision. Generally such systems are excellent for normal operations – start-up and shut-down. However, they often work poorly in transient modes, lower speed operations, anticipating problems and monitoring transitions to normal operations.  CallAid solves this problem.

Features: Callaid is permanently scanning the real-time database built into the PLC network to diagnose the current context. When a problem is identified or anticipated, it is reported to the operators. A context dependent simulation model is run and a path to return to normal is proposed. When approved, the path is monitored closely and a difference to that path may generate a new alert, new diagnosis and a re-run of the simulation model. Extensive reports of actual events were recorded to improve the system. Callaid is used by a car manufacturer, for its highly automated paint shop. The client stated that the system’s cost was recovered within a few weeks.

AMIDEC: analyze causes and the consequences of failures

Application Domain: Failure Mode, Effects and Criticality Analysis (FMECA)

Purpose: To analyze the causes and the consequences of failures (dysfunction) in large, heterogeneous systems. Failures are described together with their possible causes and probabilities, their consequences and associated severity enabling System designers to reduce dysfunctions, develop a means of diagnosis for when they do occur and arrange appropriate system maintenance.

Features: Because failures can generate other failures a criticality factor is computed linking each cause with each possible consequence. A specific algorithm has been developed to compute the criticality of causes, failures and the whole systems. Such analysis can be made on existing systems to failure analysis and maintenance scheduling, or while designing a new system to improve its reliability and ease of maintenance.

Solveur: diagnose failures in industrial systems

Application Domain: Industrial systems

Purpose: To diagnose failures in industrial systems. Solveur provides one or more potential causes along with their associated probabilities.

Features: Solveur is also a learning system. First, users describe an initial set of failures and the contexts in which they occurred. Then they allow the system to perform an initial diagnosis. At this point Solveur is used in the actual context. When it fails (no solution or wrong solution provided) users improve the system by describing the new failure and its associated context; thus, allowing more accurate diagnoses in the future. While being used, the system becomes more and more accurate making a quick and precise diagnosis of the failure. Solveur has proven especially useful in integrating a maintenance team’s experience or recording the knowledge of senior personnel before a move or retirement. Solveur has also been tested in financial environments to select the best product combination for individual investors.

ADEQUAT: HR Management

Application Domain: Human Resource Management and Information Systems

Purpose: To assist the HR director in selecting internal managers to fill available positions.

Features: Many database systems can select potential candidates for a given position; however, ADEQUAT goes the extra mile and can choose the best candidate according to your criteria. ADEQUAT is able to score each selected individual according to customized rules covering qualifications for the position, mobility and career development. Then selected managers are ranked and pros & cons are explicitly displayed allowing the HR director to make the best possible decision.
ADEQUAT can usually be connected to the HR database or Corporate database for initial loading and updating purposes; however, it works in stand alone mode while selecting candidates.  It has been used by banks, large retail stores and large industrial companies.