Manage Alarms With These Techniques
Ken Bannister | February 20, 2019
Lubrication alarms prevent catastrophic failure, but only if they trigger preventive maintenance.
Lubrication is an integral part of every rotating and linear-movement asset. A well-designed and maintained lubrication system will ensure long bearing life, machine accuracy, and energy conservation because it reduces friction by separating moving surfaces, cools bearing surfaces, washes away debris, and protects surfaces from outside contamination. Regularly checking and monitoring lubricant and lubrication-system integrity, followed by timely action when a potential or actual failure state occurs is the foundation to any proactive lubrication program. Taking advantage of condition-based and predictive strategies for managing lubrication potential failure and alarm states is essential to assure best-practice asset reliability and life-cycle longevity.
Failure can be broken down into three distinct stages. The first stage is potential failure. It occurs when a change in condition or performance is detected. For example, if we determine that the normal/designed lubricant operating pressure is between 80 and 100 psi, any deviation outside of that parameter is seen as a change in condition or performance, indicating an alarm level 1 potential failure condition. This condition can be caused by a problem in the pump, line integrity, seal integrity, or change in lubricant viscosity.
Subsequent readings are used to confirm a condition trend and the rate of decay toward alarm state 2, known as functional failure. A pending failure can be predicted and translated into a tangible window of time in which maintenance can troubleshoot and repair the problem with minor or no consequence to the machine or operational performance. This is represented by the standard P-F curve (Potential to Functional failure). Be aware, if this P-F warning is ignored, the equipment will continue to rapidly decay into the third stage, which is catastrophic failure.
Whereas a catastrophic failure is easily defined as a total non-operative state, functional failure is an interpretation based on the consequence to the person viewing the failure. For example, a seal leak at any rate is an immediate functional failure for any safety or environmental engineer if the fluid is allowed to leak unabated. The leak becomes a maintenance functional failure when the seal leaks at a rate faster than it is practical to keep the reservoir filled to the low-level limit. For most production engineers, functional failure occurs when minimum production and quality levels are breached, which may only be reached once significant internal equipment damage has been sustained. When managing alarm states, all of these functional failures must be taken into account.
To assess potential lubrication-related machine failure, the maintenance technician or reliability engineer must regularly capture, plot, and trend a variety of current lubrication-state indicator readings. Depending on the type of lubrication system and mode of failure, lubrication-system indicators are used to monitor six major conditions. These include lube pressure, flow, temperature, lubricant quality (can be partially assessed by online monitoring or by a third-party laboratory performing sample testing using oil-wear particle analysis), filter bypass condition, and/or system volume. A seventh system indicator, energy draw (amperage), can also be assessed online to indicate friction buildup due to a lack of lubricant at the bearing surface(s).
The method of current-state data capture and quality of trending will vary, depending on whether the online system-monitoring devices are passive or active. Effectiveness of any device or test method is dependent on the process that’s used to capture data, and the process used to take action when a potential or actual failure state has occurred.
Passive devices, also known as “eyes on” devices, rely on regular operator/maintainer interaction to perform visual checks and manually determine if a change in state has occurred. Typical passive-monitoring devices include in-line temperature gauges, pressure gauges, and reservoir fill-level windows. The quality of data and associated potential and actual failure control is dependent on reading consistency and frequency. Most passive gauges are not interactive, requiring the viewer to capture a reading every time and interpret that reading, based on experience. This creates a lot of data capture that results in excess administrative work and interpretive inconsistency when using operators or maintainers with differing skill levels.
Pro tip: Mark up all passive devices with engineered Hi-Lo limits and only capture readings when the needle displays outside of the safe operating window. This approach makes the device visually interactive, reduces the amount of administrative data capture, and can be consistently read and interpreted by all skill levels.
Active devices most often have some form of sensor control that, once triggered by an alarm condition, can generate an external electrical signal. Depending on the level of control required, the signal can be used to activate an audible or visual alarm such as a buzzer or warning light on an operator panel. Simple alarms however, still require an operator to understand and relay the problem to maintenance personnel in the form of a work request. The danger of any simple system is the alarm can be interpreted as a nuisance to the production process and temporarily disconnected at the source, bypassing any monitoring control if maintenance does not tend to the alarm in a timely manner.
As machine-control systems have become more sophisticated through use of computer control and SCADA (supervisory control and data acquisition) systems, alarm states are now able to sound a local alarm, if still required. Simultaneously, incidence data is gathered remotely, trended, and proactively reported to the maintenance or reliability engineering office for action. In many cases, the data-acquisition signal can also be connected directly to the CMMS (computerized maintenance management system) to print out an investigative work-order request.
The following techniques show how active-monitoring control devices can be used to automate maintenance actions at the machine level:
• For automated conveyors, the lubrication system can be set up to turn on and off using an ammeter-controlled switch connected to the drive motor. Whenever amperage draw hits a high threshold point, the lubricator is turned on and allowed to run for a predetermined number of chain revolutions. It turns off when amperage draw reaches a low-threshold point. This ensures the chain is properly lubricated based on condition and load while optimizing energy use.
• When large hydraulic oil-filtration systems use a bypass monitoring sensor alarm, instead of turning on a panel light and allowing the lubricant to internally bypass the filter element, the signal is used to activate a solenoid-operated valve placed before the filter to externally divert the lubricant through a “twinned” secondary oil-filter loop with no loss to production. The same signal is used to send a work order to change the original filter.
• In critical systems, an alarm state can be used to shut down the lube system and activate a backup lubrication system whenever a problem occurs. Once again, the signal can also be used to send a work order to change the blocked filter.
With increasing frequency, equipment installations that use controlled lubrication-monitoring devices are implementing active sensors that are IIoT (Industrial Internet of Things) and remote-control ready. However, the alarm state is only the first step. It is imperative that all lubrication-management programs have quality work flow and business processes in place that ensure that when potential failure is encountered, maintenance attends to the asset’s needs prior to functional or catastrophic failure occurring. EP