Reliability program improves refinery sulfur plant operations

John R. Braithwaite
American Western Refining LP
Lawrenceville, Ill.

The refining industry continues to live with the consequences of some 20 years of environmental laws, many of which have been implemented in ways that lack common sense.

This is particularly apparent as regulators strive to reduce remaining emissions by small increments at increasing costs.

Excessive environmental rule making will require refiners to improve efficiency by taking aggressive action to control excess emissions and maintenance costs.

In 1994, Indian Refining Co. implemented a reliability improvement program at its 86,000 b/d Lawrenceville, Ill., refinery. The goal of the program was to eliminate unplanned shutdowns of the refinery's sulfur recovery operations.

The reliability improvement program has reduced sulfur dioxide (SO₂) emissions dramatically. The program clearly has been effective in achieving environmental compliance and improved economic viability through increased equipment reliability.

Sulfur processing

In the U.S., sulfur dioxide emissions from petroleum operations are strictly regulated by the U.S. Environmental Protection Agency (EPA) and state environmental bodies. EPA's strategy to control sulfur dioxide emissions involves four programs:

• National Ambient Air Quality Standards (Naaqs) for sulfur dioxide

• New Source Performance Standards (NSPS), which limit air emissions from new stationary and mobile sources

• The New Source Review and Prevention of Significant Deterioration program, the goal of which is to prevent air quality from deteriorating in areas that meet Naaqs by requiring new, major SO₂ sources to conduct air quality analyses before receiving an operating permit

• The Acid Rain program, set forth in Title IV of the 1990 Clean Air Act Amendments.

To reduce the sulfur content of various fuels and comply with federal and state environmental regulations, petroleum refineries employ sulfur recovery plants. A typical refinery sulfur recovery operation, comprising Claus and Shell Claus Offgas Treatment (SCOT) units, is shown in Fig. 1 [55998 bytes].

Although the process is simple, problems do develop. Too much air leaves excess SO₂ in the tail gas; too little leaves some hydrogen sulfide (H₂S) unconverted.

Reliability program

Before the refinery started its reliability program, it experienced occasional unexpected malfunctions and shutdowns that interfered with environmental compliance.

Unexpected upsets and shutdowns of sulfur recovery operations cause excess SO₂ emissions and increase plant maintenance costs. Examples of resulting cost increases include: direct replacement costs for damaged equipment, labor costs for repairs, and increased costs for maintaining greater inventories of spare critical parts.

In an effort to reduce the environmental noncompliance and costly maintenance expenditures plaguing this NSPS-regulated sulfur recovery unit, Indian Refining implemented a reliability program.

The refinery established a central reliability group comprising one supervisor, four engineers, and four technicians. The group was charged with improving refinery reliability, and the sulfur recovery and SCOT units were given top priority.

The group's strategy was simple:

• To develop and implement a predictive/preventive maintenance program that would extend mean time between failures

• To develop measurement standards, or performance indicators

• To evaluate the effectiveness of the program.

Data evaluation

The reliability group began by reviewing the operating statistics for the sulfur recovery and SCOT units, including past shutdowns and equipment failure incidents. The group then analyzed the data in order to identify the causes of the failures.

The data revealed 34 shutdowns of the SCOT unit, 24 shutdowns of the sulfur recovery unit, 20 shutdowns caused by sour water stripper malfunctions, and 4 shutdowns caused by the amine recovery unit.

The group found four common, high-priority causes of unit failure (causes that fall in both the high probability and high consequence categories). Listed in decreasing order of frequency, these were:

• Fouling problems

• Electrical problems

• Instrumentation problems

• Human or procedural errors.

Fouling

The team determined that the fouling problems were caused by hydrocarbon carryover that induced formation of solids. The solids plugged instrumentation ports and other equipment, leading to loss of control, operational difficulties, excess SO₂ emissions and, ultimately, shutdowns.

The reliability group identified the main source of hydrocarbon carryover to be a hydrotreater. The liquid level in the hydrotreater's amine contactor was not controlled properly. As a result, low liquid levels caused the associated naphtha pumps to cavitate. Because of the cavitation, pressure was increased on the H₂S stripper tower, causing naphtha carryover into the amine contactor.

When examining the level control problem in the contactor, refinery engineers determined that the proper level could be maintained by controlling the pressure in the H₂S stripper accumulator. Once this scheme was implemented, contactor levels were maintained.

The maintenance of appropriate levels increased the head to the naphtha pumps so that they could overcome the resistance they were pumping against. As a result, pump cavitation and hydrocarbon carryover were reduced significantly.

In addition to the level control problem, there was a temperature differential problem in the amine contactor.

To prevent hydrocarbon from condensing inside the contactor, the amine temperature must be 25° F. greater than that of the inlet hydrocarbon stream. Typically, in this service, hydrocarbon temperature should be 95° F. and amine temperature, 122° F. At the Lawrenceville refinery, the amine was too cold, causing excessive hydrocarbon condensation in the contactor.

To solve this problem, temperature indicators were relocated so that engineers and unit operators could monitor them more easily. As a result of this change, proper amine temperature is maintained and hydrocarbon condensation has been eliminated.

Electrical problems

Electrical problems common to both the sulfur recovery and SCOT units were:

• Power failures caused by thunder storms and utility-provider interruptions

• Moisture buildup in electrical equipment

• Electronics failures in the distributive control system (DCS) computer.

Although power failures were plant wide, the effects were more severe in the sulfur recovery and SCOT units.

The reliability group could do nothing to control thunder storms, but were successful in eliminating unannounced power interruptions by the utility company. Once identified, failure problems in electrical equipment were eliminated through the preventive maintenance program.

The group further determined that electronic failures in the DCS were attributable to poor quality control of replacement electronics by suppliers. These problems were resolved and eliminated.

Instrumentation

Instrumentation problems were caused by the hydrocarbon carryover described previously, and also by unit design and equipment problems. One such problem occurred with the infrared scanners that monitor the pilot and burner flames on the sulfur recovery unit. These scanners are called "fire eyes."

Investigation revealed that the isolation valves on the fire eyes were of poor quality. A design problem was causing the valves to vibrate to the closed position, which caused the flame detectors to send a signal to the DCS indicating flame failure. This malfunction provoked activation of the DCS emergency shutdown system, causing excess SO₂ emissions each time.

In addition to the isolation valves, the fire eyes were a problem. Dust and moisture accumulation caused them to trip the units from time to time.

The refinery maintenance department placed the fire eyes on a rigorous inspection problem. The department also recommended replacement with a more-reliable valve (an appropriately sized gate valve) at the next opportunity.

In addition to the problems with the fire eyes, the team found that several sulfur recovery and SCOT unit level controllers, along with the air/acid-gas ratio controller and the air-demand control instrumentation, were giving false indications. These inaccuracies caused the unit operators to make unnecessary adjustments, which contributed to unit instability and, ultimately, caused shutdown.

To prevent shutdowns caused by false controller signals, the unit operators placed many of the control loops on manual to help stabilize unit operations. Over time, however, this only exacerbated the problem and did not improve operability.

Once the team determined that many of the control components and spares were defective, they were replaced with high-quality, reliable components, and the sulfur recovery operations were placed back on full DCS control.

Human error

The team found it difficult to identify common causes for the human errors investigated. Humans are much more difficult to predict than equipment performance.

Error rates for individuals can vary by a factor of ten, depending on mood, time of day, and emotional relationships off the job. Procedural errors, on the other hand, were addressed through improved training and revision of procedural manuals.

Training was improved by holding meetings between process engineers, reliability improvement team members, maintenance personnel, and unit operators. The goal of the meetings was to improve operational awareness.

Through additional training, operators and maintenance personnel were made accountable for individual areas of responsibility, thereby improving overall unit reliability.

Procedures were improved through the addition of:

• Troubleshooting guides indicating possible causes of equipment malfunction and failure

• Lists of equipment requiring vibration analysis and monitoring

• Recommended lubricants and lubrication intervals for specific equipment.

• Preventive maintenance task lists for critical equipment.

Preventive maintenance

A good predictive/preventive maintenance (PM) program includes equipment prioritization so that maximum effort is focused on the critical pieces of equipment, or those with significant impact on refinery operations and environmental compliance. The reliability group used the selection criteria in Table 1 [45918 bytes] to define the PM priorities used.

To ensure a high-quality, consistent PM program, detailed procedures were developed for the technicians responsible for PM work.

Inspection checklists with PM procedures-including lists of proper settings, log sheets, and tests specific to individual equipment-were made available to maintenance and operations personnel. In addition, a training and testing program was developed and implemented for personnel involved with PM work.

Performance

In high-performing refineries, failure of equipment is viewed as an exceptional event resulting from flawed focus. The reliability group focused on understanding causes of failure and using this information to improve operations, maintenance, and environmental performance.

By understanding these causes and taking appropriate action, the team made rapid improvement in extending the mean time between failures for the sulfur recovery and SCOT units.

The team used three process measurements to document the reliability improvements on the sulfur recovery and SCOT units: mean time between (MTB) excess SO₂ emissions, MTB sulfur recovery and SCOT unit upsets, and MTB sulfur recovery and SCOT unit shutdowns.

Excess SO₂ emissions were defined as any emission greater than ₂50 ppm SO₂, regardless of the duration of the emission. This conservative approach was taken so the reliability problems could be identified and corrected before a violation occurred. (The NSPS definition of an excess SO₂ emission is a 12-hr average of 250 ppm.)

Fig. 2 [13861 bytes] and Fig. 3 [15659 bytes] show the dramatic improvements in sulfur recovery and SCOT unit environmental performance since implementation of the reliability improvement program. Fig. 4 [12518 bytes]shows the reduction in the sulfur plant upsets and shutdowns that led to reduced SO₂ emissions.

These indicators provided a measure of the health of the sulfur recovery and SCOT units, and helped the reliability group set refinery performance targets.

Key maintenance indicators also were tracked. These included:

• Quantity of work requests per month

• Quantity of preventive maintenance work requests per month

• Work order priority trend

• Type of work requests

• Quantity of high-priority pump work orders and maintenance costs per month

• Monthly maintenance costs.

Of the maintenance indicators, the most significant was the maintenance cost indicator (Fig. 5 [10515 bytes]). Maintenance costs decreased dramatically-from a high of $260,000/month in May 1994 to well below $50,000/month in April 1995. This cost reduction clearly exceeded Indian Refining Co.'s target reduction of 10%.

The refinery was shut down in April 1995, and restarted for brief runs during April, May, and July 1995. After the July run, the sulfur recovery unit was not restarted again until late that year, when the refinery was purchased by American Western Refining LP.

The Author