An Innovative Approach of How to Maximize the Efficiency of your Plant Assets

Poorly performing assets - a compressor, a turbine - drive production losses, quality problems and, in the worst cases, safety incidents. This whitepaper sets out an approach to plant asset performance management that combines out-of-the-box process knowledge, automated root cause analysis, predictive models and dynamically adjusted performance targets in a single integrated platform.

It begins with the four challenges facility owners in process manufacturing face today: access to the expert who can diagnose an asset's behaviour quickly; long diagnosis cycle times that let degradation worsen; identifying the optimum shutdown window for maintenance; producing accurate diagnoses when the real fault sits in upstream or downstream equipment; and consolidating asset information scattered across maintenance management, performance management and alarm management applications.

The paper's first answer is packaged process knowledge. Smart Equipment modules, built on Integration Objects' KnowledgeNet® (KNet) platform, are ready-to-use components covering compressors, turbines, furnaces, columns, boilers, reboilers, reactors, heat exchangers, pumps, valves and vessels, each shipping with performance metrics, expert rules for complex event detection and root cause analysis models. The Smart Compressor module is examined in detail, including its defined metrics (compressor efficiency, polytropic head, compressor power) and a worked complex event rule that predicts surging by calculating real-time distance to surge from the centrifugal compressor's surge line equation and performance curves, replacing the manual chart-reading process mechanical and process engineers otherwise perform.

The predictive section explains fault propagation modelling: building cause-and-effect graphs where progression upstream identifies root causes and progression downstream predicts impacts, illustrated with a Low Compressor Performance fault model. It then covers predictive model construction, selecting a baseline dataset that excludes shutdown periods and outliers, building a baseline model from correlation between a KPI and its primary drivers, and handling non-linearity by clustering scatter data into separate operating modes with a best-fit line per cluster. Rules validate model inputs, decide when models apply based on operating mode and equipment status, and sanity-check prediction results.

Later sections address the integrated asset management framework: online monitoring, performance reporting, predictive analytics, abnormal conditions management and integrated workflow management, and dynamic performance targets that adjust to detected operating mode, process scenario and feed rather than relying on static manual setpoints.

Aimed at reliability and maintenance engineers, process engineers, operations managers and plant managers in refining, petrochemicals and process manufacturing.