The Future of Predictive Maintenance

Process equipment represents an enormous investment for manufacturers, with medium to large plants having thousands of assets that support the supply chain, production and distribution. All equipment must be managed efficiently to ensure reliability and productivity. This is a priority for industrial plants as they seek to move from reactive and preventive maintenance approaches to predictive. According to ARC Advisory Group, the growing number of plant assets coupled with the reduction in skilled labor means there are 500 to 2,000 process control loops per technical staff member. Maintenance personnel are responsible for the associated equipment for an individual loop, including instruments, valves and pumps, making it increasingly difficult to manage plant assets using in-house staff alone. With baby boomers retiring at the rate of 10,000 per day, industrial plants are relying on automation solutions to manage assets. This dependence is also driven by aging infrastructure, competitive market pressures, environmental regulations and shrinking profit margins.

Predictive Maintenance

The growing use of predictive analytic software, wireless sensors and cloud computing, as well as other information technology advances, is a key enabler. These technologies broaden the scope of real-time performance monitoring and machine-to-machine communication, which, in turn, supports the evolution of predictive maintenance approaches. In some cases, equipment can also be self-repairing. Accurately predicting remaining useful life helps streamline the supply chain with the right parts arriving for scheduled replacement. Unscheduled downtime and inventory costs are minimized. Suppliers can better support equipment using remote diagnostics. In the future, predictive maintenance systems will collect real-time and historical data, apply analytics, and assess the health of the equipment and the entire production process. The systems will prioritize maintenance and operator actions based on criticality and recommend corrective actions. These expanded capabilities will enhance maintenance and operating decisions while providing critical equipment information to process control and business systems. Plant management and engineering can make more informed decisions related to capital budgets and long-term modernization plans.

Intelligent Pump Control

A common application of intelligent pump control is a parallel pumping system such as a cooling tower. Parallel systems are designed to provide incremental flow by turning multiple pumps on or off to meet changing system demand. Cooling tower systems are common across manufacturing facilities that use heat exchangers. These applications offer an opportunity for industrial plants to reduce energy consumption and improve reliability. For example, multi-pump application of intelligent variable frequency drives (VFDs) represents the convergence of pump and automation technologies for predictive maintenance. The pump intelligence embedded in the drive's microprocessor allows the pump to identify conditions such as dry running, dead-heading and cavitation in real time. Conditions can be communicated via a digital bus back to the distributed control system (DCS) for display and alarming. The operator can quickly take corrective action either by making adjustments to the system at the DCS console or by automatically generating a high-priority work order requesting system maintenance. The individual drives are interlinked and can be configured to automatically change the lead and lag pump(s) at regular intervals to even mechanical wear over time. If one pump fails, the system can automatically synchronize and adjust individual pump speeds to ensure that flow demand is maintained. Multi-pump control maintains stable process conditions, which optimizes the number and speed of pumps needed and offers smooth startup and shutdowns because there is no need for a separate control logic in the DCS. The drive's intelligence also operates the pumps in the most energy-efficient manner, allows redundancy and provides the capability to mix pumps of different sizes and powers (although the ideal system configuration has identical pumps). While the multi-pump intelligence is embedded in the VFD microprocessor, the intelligence could also be in the DCS microprocessor or on a server in the cloud. While this is a special case of using embedded intelligence to control a multi-pump system, the future will see broader use of data to empower predictive analytic software for plant-wide asset management. Facilities will be able to implement mechanical systems that self-diagnose and automatically adapt to equipment failure and process upsets. Compared with fixed-speed systems, the broader use of variable speed pump systems requires smaller motors, allows removal of the control valves at the pump discharge and prevents the need for supporting infrastructure such as pneumatic lines and control wires. Wide application of variable speed technology will dematerialize the process and reduce maintenance and operating costs. In these scenarios, facilities can continuously monitor assets and make system adjustments over a wireless or wired network. Service technicians, whether on or off site, will eventually be able to use 3-D printing to produce standard or custom parts and ship them in a timely manner. These advancements and others will allow plants to work within budgetary constraints and overcome barriers caused by limited resources.