Jul-2017

Troubleshooting crude column constraints

Multiple limitations to operating capacity were identified and resolved during a column turnaround

JILL BROWN BURNS, KRISTEN BECHT and BRANDT MUELLER
Valero Energy

Article Summary

Troubleshooting operating problems is a daily part of the job description for a refinery process engineer, especially when charge rate to the refinery is inhibited. The successful identification of column operating problems becomes critical when the resolution includes modification to distillation tower internals that can only be executed during a turnaround outage. “Downtime is expensive. The cost of misdiagnosing a problem is equally enormous.”¹ And so is failing to diagnose the problem altogether.

Development of turnaround work scope challenges every refinery process engineer, especially as turnarounds seem to grow further and further apart throughout the life cycle of a refinery process unit. While this practice reduces major maintenance costs, extending turnaround cycles consequently minimises the opportunities to address unit operating problems that require downtime (without the enormous cost of an unplanned outage). Thus the engineer needs to effectively identify resolutions to operating problems, and develop the right scope to address them during the outage. The consequences of misjudging the fix can be significant, as the whole team is left scrambling in the midst of major maintenance work to retrace troubleshooting steps and to order equipment with a hefty emergency expediting price tag.

While many papers have been presented that highlight interesting and fascinating discoveries found inside a column during a turnaround, in this article the authors will describe a case study where all the critical discoveries were predicted and anticipated through good engineering practices. Troubleshooting fundamentals like energy balance and operating data analysis were employed along with utilising more rigorous tools such as isotope scanning to conclusively identify three separate and unrelated areas of limitation in the crude column. We will also discuss an unconventional hot tap bypass that allowed the unit to remain in operation until the planned turnaround, as well as outline the design changes that were implemented during the outage to prevent and address future limitations. The authors will finally present turnaround findings – ‘validations’ rather than ‘discoveries’ – that prove the conclusions drawn from operational and troubleshooting data analysis were correct, the appropriate equipment was ready and on-site, and the planned work scope was spot-on for addressing the tower problems.

Background
Valero Ardmore refinery operates a 90000 b/d crude and vacuum distillation unit, charging primarily mid-continent crude through the Cushing terminal (see Figure 1). The crude tower is a typical atmospheric operation (630°F [330°C], 40 psig flash zone) with 42 trays, a packed gasoil wash section, and a shrouded stripping section. Overhead light naphtha is routed to a gas plant, and atmospheric tower bottoms are routed partially to a gasoil hydrotreater (CFHT), with the balance to the vacuum tower. Five side draw products from the crude tower are further processed within the refinery. The tower has four pumparounds (heavy naphtha, kerosene, light diesel, and atmospheric gasoil) utilised to control product cutpoints according to downstream specifications.

This crude tower, like many atmospheric crude units, has a history of corrosion in the top of the tower. The top of the column was overlaid through Tray 4 with high alloy during a series of outages between 1998 and 2011. Overhead condensers experienced short run lengths, attributed to amine salt under-deposit corrosion. The overhead line was replaced, and the trays were renewed with high alloy during the 2011 crude unit turnaround. A chemistry change was made in 2012 to address the root cause of the corrosion, and to eliminate the high salt-point amine injection for neutralisation of hydrolysed chlorides in the overhead. The refinery expected the high alloy and revised chemistry to improve reliability in the top of the tower, but scheduled an isotope scan as a ‘health check’. Isotope scans (both active area chords and centre downcomers) in October 2012 and July 2013 showed normal tray loadings throughout the tower. A high froth height was observed on Tray 3 during the July 2013 scans, potentially signalling the early stages of column fouling/corrosion. However, operation remained normal.

In August 2013, several key heat exchangers in the crude unit preheat were cleaned. After this point, instability in the level of the kerosene side stripper was noticed (see Figure 2). The main column remained stable until November 2013, when the column overall pressure drop began to show intermittent periods of liquid accumulation (flooding).

Given the history of corrosion in the tower’s top section, it was hypothesised that the top naphtha section was flooding, holding up liquid and subsequently limiting liquid available to the kerosene draw.

When the instability occurred, the pressure drop in the tower increased by 3 psid, which has an equivalent head of ~10ft (3m) of naphtha-ranged hydrocarbon. With the pressure drop indication on the overall tower, the location of the accumulation must be determined by reviewing other data. The 10ft of liquid head could be localised to the top head of the tower or distributed amongst 20 trays. Further investigation was required.

Troubleshooting the crude tower
Engineers have a lot of tools available to diagnose column flooding. Isotope scanning is a powerful tool and was an integral part of troubleshooting this tower. However, the authors started with the fundamentals – reviewing data (field and historian) and evaluating the tower energy balance.

Using engineering fundamentals to troubleshoot tower flooding
Ardmore had altered the heat removal in several of the pumparound circuits by cleaning crude preheat exchangers. The authors recognised that redistributing the heat load in the tower impacted internal reflux. If a trayed section operates near a hydraulic limit, the change in internal loading could push that section into flood with resultant high pressure drop and instability observed at a product draw. Pumparounds, tower temperatures and product overlaps were analysed to understand the impact of exchanger cleaning on the tower’s internal traffic (see Figure 3).

Diesel pumparound duty had increased by over 200% (19 MMBTU/h to 50 MMBTU/h), significantly shifting heat removal lower in the tower. The kerosene-light diesel overlap increased with the higher diesel pumparound duty, indicating reduced internal reflux traffic below the kerosene draw (see Figure 4). The temperature indication in the vapour space of Tray 21 (directly below the kerosene draw sump) was observed to oscillate. This unstable behaviour preceded the exchanger cleaning by two months. After the instrument was checked in the field by a technician, it was hypothesised that this temperature instability may be symptomatic of flooding, where hold-up robs cooler condensed internal reflux from the trays below the kerosene draw.

The kerosene draw is configured with a centre downcomer draw sump. Pumparound and product are drawn from the sump and split in external piping where the pumparound stream is routed to a pump, and the product drains to the side stripper on stripper level control (see Figure 5). The liquid condensed by the pumparound plus the internal reflux from the trays above can either be drawn as product or overflow the sump onto Tray 21. The skirt panels of the Tray 20 centre downcomer were not submerged to provide a positive liquid seal at high draw rates, indicating that the original design intended for a large proportion of the liquid to overflow the sump and act as internal reflux for fractionation of kerosene from light diesel.