A straightforward maintenance strategy yields improved plant performance
Moving from reactive to preventive maintenance helps Monsanto improve output and reduce costs at its Muscatine plant. Smart field devices and control valve positioners provide critical diagnostic data to reliability engineers, guiding their efforts.
In many situations, a straightforward and consistently applied maintenance strategy can have results that attract positive attention as production and plant availability increase while maintenance costs decrease. Such is the case at Monsanto’s Muscatine manufacturing facility. The combination of technical solutions with corresponding work processes has had measurable positive results and earned the facility the 2012 HART Plant of the Year Award.
The Monsanto Muscatine facility is located in southeast Iowa along the Mississippi River. It manufactures Roundup herbicide along with acetanilide select chemistry products, including Harness Xtra, Degree Xtra, and Warrant. The facility covers 150 acres and employs around 450 individuals to operate and manage eight individual process units. These include waste treatment and plant utilities along with the product manufacturing process areas. The majority of the plant operates continuously, year-round, and the oldest parts of the facility date back 50 years.
An inherited maintenance platform
The processing units are controlled by a mix of Emerson DeltaV and Provox DCS (distributed control system) platforms tied to more than 3,200 instrument assets, including transmitters and control valve positioners. Of this total installed base, there are currently over 600 HART and 125 Foundation fieldbus instruments within the facilities AMS Intelligent Device Manager. The outstanding instruments not yet in the AMS Intelligent Device Manager are in large part either not terminated to smart I/O or are simply non-smart devices. One of the pivotal changes in Muscatine’s operation happened in 2006 when Joel Holmes, site electrical reliability engineer, received a standalone Emerson AMS unit from another Monsanto plant. Having access to that unit enabled him to explore some of the capabilities of a more sophisticated maintenance strategy using diagnostic information gathered from field devices using HART Communication and Foundation fieldbus protocols.
Given the facility’s extensive population of maintenance-intensive valves and instrumentation, this was an area where Holmes felt that improvements in reliability would have the largest impact on overall plant performance. Part of the process of integrating his new maintenance tools with the work processes involved assigning a criticality value to the instrument assets in each process unit. Those that could impede production were given the highest priority for monitoring.
As Holmes described the process, “We utilize an A, B, or C letter designation to denote asset criticality. The highest criticality ranking is an A, with the lowest criticality ranking being a C. Keep in mind, A-ranked critical devices will receive the most attention since their performance is crucial to the on-stream time of the production system whereas C-ranked devices may, in many cases, be allowed to run till failure since their impact to the production system is minimal or they have in-line spares.”
As Monsanto gained experience with AMS, use within the plant became more sophisticated with improvements in overall effectiveness. The Alert Monitor became the primary predictive tool integrating the plant’s asset criticality measure with the Device Manager. The respective device group assignment in turn directly impacts its polling rate, or the frequency at which the Alert Monitor extracts the devices status, health, and diagnostic information from the field.
Over time, Holmes created three groups of devices: transmitters, control valves, and vapor sensors. This helped balance the loading of the Alert Monitor since having too many devices poll at the same frequency overloaded the system, resulting in extremely slow update rates or no response at all. Each device group has a set series of polling rate values dependent upon its criticality ranking.
Working in a legacy environment
While some of the process units have relatively new control systems with HART-enabled I/O, there are other areas where the system is still plain analog. These will be upgraded as the plant moves through a series of migrations to DeltaV control systems with CHARMS I/O, but in some areas that is still years away. The work-around that the plant has developed to gather diagnostic information involves either WirelessHART THUM communication for critical tags, or HART/Foundation fieldbus handheld communicators for devices that require less frequent monitoring.
The WirelessHART deployment is growing with about 25 devices currently in operation. Holmes expects that number to grow, and each process unit will soon have its own wireless gateway.
A case in point
As the Glyphosate Technicals unit was anticipating a scheduled shutdown for catalyst regeneration, Holmes and his team addressed a planned and scheduled deficiency work order with a two-inch eccentric ball control valve that had been identified as suspect through a prior control valve reliability PM route. The AMS Intelligent Device Manager along with its ValveLink Snap-On application had measured increasing friction as the valve moved through its full span of travel. This indicator suggested a problem was developing and triggered a failure when performing the diagnostic scans on the control valve assembly. While the valve was not exhibiting any issues that would be visible to the operators, the valve signature indicated that the amount of force required to open and close the valve had increased substantially. It was still able to perform its control function for the moment, but comparisons with historical signatures revealed that a more drastic situation was developing.
Holmes recommended that the electrical reliability technicians change out the valve during the outage; therefore, the reliability technicians proceeded to remove it from service during that time. When they pulled the valve body free from the flanges, they found a 1/4 x 4 in. bolt stuck in the valve that had apparently come off one of the clamps holding a filter element designed to capture the catalyst. It had been carried down the pipe by the flowing process stream. Had that bolt remained in the valve, at some point it could have initiated flow control issues up to and including the valve refusing to close which would have caused additional problems.
The bolt had damaged the valve's ceramic plug and seat, so once the valve body was removed and separated from the actuator, it was sent out for rebuilding. A new valve body was mounted on the actuator in the shop. Prior to reinstalling the assembly, it was calibrated and then tested with the use of the ValveLink Snap-On application to make sure the valve signature and additional diagnostic scans indicated normal operating parameters using HART. All scans including the valve signature were recorded in AMS so that results can be compared to future analysis in trending for degradation. When approved, the valve was returned to service without delaying the scheduled startup of the process unit.