The Plant Services Best Practices Awards recognize management techniques, work processes, and product and service implementations that exemplify the definition of a best practice, which the Society of Maintenance and Reliability Professionals (SMRP) defines as: “a process, technique or innovative use of resources that has a proven record of success in providing significant improvement in cost, schedule, quality, performance, safety, environment or other measurable factors that impact the health of an organization.”
Entries must demonstrate how to implement a best practice, show the potential payoffs in both qualitative and quantitative terms, and provide inspiration for those who must overcome cultural inertia and make effective changes. Entries may be submitted by plant personnel, vendors, engineering firms, consultants or anyone who is familiar with the application and has permission to make it public knowledge. Our 2010 categories also include Equipment, Management and Energy Efficiency, but this round’s focus is on Reliability.
Every contender offered an impressive reliability practice that can increase productivity, improve efficiency or reduce costs. Judging criteria included percentage reductions or cost savings, return on investment and broadness of applicability, with recognition given for innovation and creativity.
The winning practice was submitted by Jayesh Patel, reliability manager, Valero Refinery in Paulsboro, New Jersey. By managing its equipment below the alert level, the refinery is able to be proactive in its machinery management, allowing Valero to mitigate reactive work and the associated process interuptions. The results of this shift to proactive maintenance are improved product quality, improved machinery availability and increased profits.
Condition monitoring is combined with decision-support capabilities that utilize prewritten rules, as well as additional customized rules set by Valero.
The combination allowed the refinery to schedule maintenance without the additional pressure of emergency conditions, and Valero’s successful implementation won the votes of our judges to become this round’s Best Practice in Reliability.
Winner: Decision-support system lets rules dictate maintenance
Refinery uses protection systems in conjunction with condition-monitoring and decision-support software
Valero’s Paulsboro Refinery has a capacity of 195,000 barrels per day and employs nearly 550 individuals. Condition-based maintenance is used extensively and employs a mix of permanent and portable technologies, depending on asset criticality. Low-criticality assets are addressed by a portable data collection system. High- and mid-criticality assets are addressed by online systems. For its most critical assets, Paulsboro uses Bently Nevada continuous machinery protection systems in conjunction with System 1 software. These assets include gas turbines, steam-driven and motor-driven centrifugal compressors, hydrogen reciprocating compressors, utility air compressors and liquid ring compressors for flare gas recovery. Mid-criticality assets in the refinery’s coker unit are addressed by the Bently Nevada Trendmaster system, a permanently wired “sensor bus” architecture that monitors conditions several times per hour. Both the continuous monitoring systems and the Trendmaster architecture are tied into System 1 software for a unified online condition-monitoring environment.
One of the keys to Paulsboro’s success with condition-based maintenance is its practice of managing machinery “below the alert level.” Alarms set to notify machinery specialists of impending problems allow uninterrupted operation while appropriate actions, such as scheduling maintenance, planning an outage or recommending changes to operating or process conditions, are taken. This proactive maintenance drives the Paulsboro refinery to the far left limits of the P-F curve, resulting in higher product quality, improved asset availability and increased operating profits.
The P-F curve shows qualitative time relationship between potential failure (P) and functional failure (F). The further to the left (closer to P0) one can operate for any given asset, the easier it is to plan maintenance and lower the likelihood of surprise functional failures.
Managing too many alarm levels can become onerous, and a balance must be found in the quest to move farther to the left on the P-F curve. One way to achieve this is by relying not only on level-type alarms, but also on technologies that automate the data analysis and anomaly detection processes that human experts would use if manually reviewing data. Paulsboro has used the System 1 software’s decision-support capabilities to embed subject-matter expertise for a particular asset or class of assets and detect asset problems automatically. While many users employ the decision-support module to detect anomalies with the rotating machinery monitored by System 1 software, what has set the Paulsboro facility apart is its use of the system on non-rotating assets, as well. By bringing process data from the plant’s distributed control system (DCS), turbine control systems and process historian into the System 1 database, Paulsboro is able to apply the decision-support engine in analyzing and detecting anomalies on assets for which only process measurements are available, addressing applications outside of conventional condition monitoring and detecting problems in non-rotating portions of turbomachinery.
To address this mix of conventional rotating machinery, fixed equipment and process-related applications, Paulsboro uses both GE’s machinery expertise in the form of pre-configured RulePaks, and their own expertise in the form of custom rules written by the resident subject-matter experts. In this way, Valero has decision-support capabilities tailored specifically to the needs of its Paulsboro operations.
“System 1 not only has predefined rules based upon Bently Nevada’s 50 years of machinery diagnostics experience, but also allows us to write our own rules,” says Jayesh Patel, reliability engineer at the Paulsboro facility. “These rules are what allows us to automate the diagnostic process. Further, we’re able to test our rules on historical data to make sure that they fire when, and only when, we want them to.”
For example, the feed filter for the naptha hydrotreater has a direct effect on product quality and operational problems, and as such it is critical to production. Before implementation of the decision-support system, the filters were replaced every 45 days to 90 days, but more than a third (35%) were replaced under an emergency work order. In the past two years, the filters have been replaced 10 times and not one was under emergency conditions.[pullquote]
Also, after nine months of operation, the recycle gas compressor needed to be shutdown for a water wash because of high vibration levels. Rules were developed to give advanced warning to this particular condition. Knowing that the compressor will need a wash allows it to be scheduled on Valero’s time, rather than the machine’s.
The 10 burner cans on the gas turbine provided another opportunity for operating belowing the alarm level. One of the rules defines the maximum difference between combustion thermocouple readings. After an outage, insulation problems were diagnosed, and information from System 1 provided the confidence to operate the turbine at full load.
With the facility’s asphalt heater at the decoking plant, overfiring can lead to loss of creep strength in the tubes, resulting in failure. Clogging also can limit unit performance. Rules compare tube temperature with empirically derived maximum values, and severity levels are assigned based upon temperature limits. Temperatures are compared to average temperatures from the previous week. Increases above certain limits triggers an event for a process engineer to analyze inlet conditions and feed compositions, saving both time and money.
Rupture disks protect pressure relief valves (PRVs) by separating them from corrosive process conditions. When they burst, as they’re designed to do, they need to be replaced. A pressure transducer is installed in the spool piece between the rupture disk and the PRV, and when pressure is detected a work order is written to have the rupture disk replaced. Regulatory paperwork also is filled out at this time. Finally, it’s important to Valero that the stack emmissions analyzer results agree with those obtained when the emissions are analyzed by a lab. System 1 rules are used to make sure these reports are in agreement, thus avoiding penalties.
Electrical audit cuts costs
Chain drive to belt drive conversion
Dale Renner, a sales representative from Bearing Headquarters, a Des Moines distributor, suggested replacing the roller chain drive with a belt drive. The recommended belt length and sprocket sizes fit within the elevator superstructure. While the chain drive required three idlers to take up slack, the belt drive required only one, and its noise level was less than 10 dB.
The conversion produced two immediate benefits — noise and the contamination risk from the twice weekly need to lubricate and clean the chain — and one longer-term benefit — eliminating mechanical stress on startup.
Trenton upgraded its main process control system and power distribution system. Redundant servers running HMI software monitors the plant’s SCADA system wirelessly to four remote pumping stations, and now management can troubleshoot problems immediately and remotely.
A Rockwell Automation services and support contract gives the plant monthly on-site visits from a technician, software upgrades, and a fixed rate for emergency service.
When a main server meltdown blinded the plant, within five hours the service technician transferred plant operation to a redundant system, thus ending the risk of an overflow or process failure. Plant operators keep a checklist of items for the technician to address during the monthly visit. Employees now have a resource to answer questions that arise. Nuisance issues get fixed before they become larger problems. The technician provides ongoing training and software upgrades for plant employees.
Trenton now has one-point contact instead of many different service providers that blamed problems on another company’s product.