Nov-2023

Optimising fouled distillation units

Case studies discuss optimisation strategies and enhanced distillation unit performance against fouled conditions.

Soun Ho Lee
Valero Energy Corporation

Article Summary

Fouling is a critical issue in distillation unit operations. Fouled conditions not only downgrade unit performance but also reduce unit run length. Online equipment cleaning techniques can be applied while a unit is running. However, the effectiveness of online cleaning is limited, and fouling remains in most cases.

Thorough distillation equipment cleaning and/or upgrading can resolve or mitigate fouling. However, these activities require a distillation unit outage, resulting in a loss of production. What if a unit outage is not feasible and a distillation unit continues operating while handicapped? Is there a way to reinstate distillation unit performance without a unit outage?

Case study 1: Background
The crude distillation unit (CDU) under discussion belongs to a refinery that processes blended feedstocks among light, medium, and heavy crude slates. Charged crudes are heated through two parallel preheat trains and furnaces and then introduced into the crude atmospheric tower. This tower separates crudes into intermediate products:
υ Unstabilised light naphtha
ϖ Heavy naphtha
ω Kerosene
ξ Diesel
ψ Atmospheric tower bottoms (ATB).

The existing flow scheme was designed in the early 1990s to process medium-gravity Middle East crude slates. Current crude slates consist of light domestic crude and heavy imported crude slates. Heavy naphtha streams were withdrawn from two different locations and introduced to a single heavy naphtha side stripper.
It was believed that a single nozzle was not big enough to achieve targeted heavy naphtha draw rate, and two side- draw nozzles could help relieve hydraulic limitations. Seven fractionating trays were arranged between the top naphtha pumparound and the heavy naphtha lower draw nozzle, and eight trays were also positioned between the heavy naphtha lower draw and kerosene pumparound section. The kerosene pumparound and product had different draw locations to maximise kerosene-diesel fractionation. The crude atmospheric tower configuration is shown in Figure 1.

Case study 1: Problem description
This bespoke CDU faced multiple challenges in recent operations. Heavy naphtha-kerosene fractionation section flooding resulted in poor heavy naphtha quality. Pressure drop across the section was increasing through the run. Downgraded crude preheat train performance reduced the heater inlet temperature. Moreover, operation instability was experienced due to inconsistent kerosene pumparound heat removal. Crude preheat train heat exchanger bundles were pulled. It was found that severe foulants were built up in multiple heat exchangers. A fouled crude/kerosene pumparound cross heat exchanger photo is shown in Figure 2.

Case study 1: Troubleshooting
The aforementioned fractionation section pressure drop trend and heat exchanger fouling supported fouling in the heavy naphtha-kerosene fractionation and kerosene pumparound sections. Foulant material samples were obtained from the pulled heat exchanger. Laboratory testing of foulant material detected a significant amount of phosphorus. It indicated a root cause of fouling was phosphorus.
A crude atmospheric tower scan was also arranged to filter limited sections in detail. Scan results are translated in Figure 3. The scans disclosed some interesting behaviour. Seven fractionating trays above the heavy naphtha lower draw were severely underloaded (‘near dry’ conditions), while eight trays underneath the heavy naphtha lower draw nozzle were normally loaded. How could trays be more loaded after the liquid draw?

In addition, simulation modelling indicated less than one theoretical stage for the section above the heavy naphtha lower draw and three theoretical stages for the section underneath the heavy naphtha lower draw. Fouling could erode tray efficiency. Nevertheless, less than one theoretical stage out of seven actual tray counts was an extreme case. It was suspected that tray fouling was not a root cause of poor tray efficiency alone.

A field survey, including local temperature and pressure measurements, was conducted to identify the root causes of the unusual tray traffic profile. Temperature survey results highlighted in Figure 3 indicate that measured temperatures at the heavy naphtha upper draw, lower draw, and combined draw were the same. If heavy naphtha was withdrawn through the lower draw nozzle, the field temperature survey should indicate three different temperature levels.

Scrutinising the heavy naphtha draw circuit revealed another root cause. The heavy naphtha draw circuit geometries and pressure measurement values are summarised in Figure 4. The hydraulic balance was constructed based on the draw geometries and measured pressure values. The simplified hydraulic balance is also illustrated in Figure 4. Heavy naphtha flow to the side stripper was controlled by the side stripper level. The static inlet pressure of the level control valve was 35 psig. If the upper draw was full, pressure at the upper draw location would be 10 psi higher. A static pressure of 45 psig was ample to maintain driving force.

Meanwhile, measured pressure at the heavy naphtha lower draw was not matched to possible head pressure at the tie-in point unless the upper draw was completely empty. In this hydraulic balance, the lower draw would be pushed back into the crude atmospheric tower by the upper draw liquid. This undesired reverse flow caused seven underloaded fractionation trays and downgraded fractionation performance. Liquid from the heavy naphtha upper draw bypassed seven fractionating trays. It resulted in low tray efficiency.

Kerosene pumparound instability was also reviewed. The kerosene pumparound configuration was previously depicted in Figure 1. Three different draw nozzles were available. Operators thought a single nozzle was not large enough to draw both product and pumparound streams. Kerosene product and pumparound streams were withdrawn at different trays. The configuration was also believed to maximise fractionation between kerosene and diesel.