By Herman Storey
Wouldn't it be great if we could truly optimize maintenance? How about if we only worked on instruments that actually needed attention, but did this work before they were in bad enough shape to have an impact on plant operation? We can!
But doing so requires careful planning and organizational discipline. If we want to optimize maintenance and lower our operating exposure to field-device failures, we have to do some basic implementation steps to properly present diagnostic information, and we need people and work processes to make use of the tools.
Asset Management Goals
Asset management programs have three basic goals:
- They must reduce unmanaged risk.
- They must assure sustainability over an asset life cycle. ;
- They should reduce operating and maintenance costs as a direct consequence of the above.
The minimum requirements for an asset management system include:
- A management system that provides for accountability and authority, measurements and auditing and reporting tools;
- Human resources (adequate qualified personnel);
- Deployed and integrated tools;
- Clearly defined processes, roles and responsibilities
Asset Management in the Oil Patch (and Other Places)
To apply the above to automation systems in mining and manufacturing industries, some basic diagnostic techniques associated with these systems are needed:
- Model-based process or equipment diagnostics.
- Control utilization and performance monitoring.
- Safety and cyber security programs, including tools for network or system monitoring.
- Dumb things that require manual tests. This is a primary target for periodic proof testing for safety devices or systems.
- Smart things with internal diagnostics. This is a much underused resource that can provide huge benefits.
This is the primary target of the remainder of this article.
Life-Cycle Work Processes
An optimized asset management process involves activities during all phases of the asset life cycle. Performing these activities at the correct phase of the life cycle is crucial to achieving optimum results. Some asset management activities required during a life cycle include:
- Front-end design (FED—also called front-end loading) is crucial. Criticality ranking is the first step. Use criticality determination to optimize tradeoffs between predictive maintenance frequency, redundancy or optional diagnostics capability. These tradeoffs are more expensive or completely impractical after FED.
Do vendor qualification, selection and templating of typical configurations as soon as possible after criticality and design requirements are set. Delay until after FED will result in sub-optimum cost or implementation choices and/or rework after bulk building.
- The detail design and factory acceptance test (FAT) phase of a project should include building and testing of all tools necessary for asset management during the rest of the life cycle. Begin training of field personnel during this phase.
- The install, commission and loop check phase represents the first time in the life cycle when asset management tools can actually give a return on investment. In fact, these tools can pay for themselves during this phase. This is also the best time to train field personnel on their use.
- The pre-startup safety review and actual plant start-up also provide for significant benefits from asset management tools. Many subtle or hard-to-detect issues can be found by a quick sweep of diagnostics during these phases.
- The run-and-maintain phase of the asset life cycle provides the biggest opportunity for diagnostic-based asset management tools and work processes. During this phase, you can detect asset deterioration at an early stage and plan and execute corrective action with minimal impact on operation. These tools minimize maintenance tasks.
- Turnaround management offers another opportunity for diagnostics-based asset management. Perform diagnostic screening immediately before asset shutdown, during shutdown and immediately after shutdown to identify problems not detected during normal operation. These screening activities can also identify candidates for repair-in-place or devices that need to be removed for shop repair or even replaced.
Asset Management Tools
Asset management tools generally have data management functionality along with other primary functions, including:
- Resource planning, tracking and accounting
- Design calculation, drawing and documentation
- Hosts that provide for configuration management of systems and attached devices, troubleshooting and human interface, including DCSs, handhelds or laptops.
Each has one or more databases and is designed to use its data as master, even though the data is often shared with another tool, and a different tool may have the master data. The actual master data location changes between tools during the asset life cycle. Keeping track of which data is master and which is a copy requires some planning and effort.
Tools, devices and systems must exchange data to keep the databases synchronized. Data import, export and reformat are needed to accomplish data exchange. Data comparison and reconciliation is necessary to assure integrity of exchange. Change tracking and history are needed to manage asset data changes.
In addition, all asset management tools, systems and devices need remote access, multiple simultaneous access and role-based security. A standardized security model would be nice.
Use of Asset Management Tools
When asset management tools are used in the field, most diagnostics go directly into a black hole and become "stranded" and useless. Instead they should go to engineering, maintenance and/or operations.
Alert priority is a function of severity of impact, not its likelihood. The severity and likelihood should both be determined and documented in the criticality ranking process.
Proper use of diagnostics in real time depends on a graphical user interface (GUI) that covers all system nodes, networks and modules, and attached devices. The GUI should allow for navigation from overview to detail in three mouse clicks.
Free format diagnostic reports are necessary for periodic surveillance by engineering and maintenance personnel. The reports should indicate historical activity over a time period.
Diagnostics must always flag data quality. This should be done in the manner specified in NAMUR recommendation NE 107. Data quality must propagate to all applications and graphics.
Most HART installations use the 4 to 20 mA signal to flag data quality. This allows flagging data bad for a failed transmitter, but does not support the other states in NE 107. Also, this mechanism does not report valve failures. There is also a lack of standardization between vendors for the transition between good and bad signals. The NAMUR NE 43 option for HART transmitters will avoid this.
The most common asset-management practice for field devices is to issue handheld communicators to technicians and hope for the best. Audits reveal that the best doesn't happen. In fact, configuration drift and random configuration errors are very common in field installations that rely on handheld communicators.
The only way to maintain configuration accuracy is to use an integrated host with data management features for all configuration tasks. Some users have even banned use of handheld configuration tools—even though they are convenient for some field tasks.
Good configuration management also requires good work practices. These include a template-management process for default and option settings for device configuration.
Some diagnostics problems are the same as those for configuration management. Most vendors have a portable tool that can do a good job as long the tool is used to monitor the device, but these tools do not scale up support to large installations and do not integrate well with hosts that could monitor and report diagnostics while unattended.
The scale-up problems come from use of polling for alerts and slow communication speed. These are inherent limitations of HART, but are avoided by Foundation fieldbus technology. Unfortunately, some vendors have chosen to use polling for Foundation fieldbus with resulting limitations in functionality, scalability and performance. This problem can be partially mitigated for HART subsystems by restricting these polling subsystems to 64 device scan groups with simultaneous parallel polling of each subsystem.
Herman Storey is the principal with Herman Storey Consulting, LLC