The company among others provides market with such goods as: high quality lead-free petrol, diesel oil, heating oil, aviation fuel, lubricating oil, asphalt, paraffin wax and LPG. It sells and distributes its products in the local markets in Poland, as well as exports them to international customers in global markets.
Grupa LOTOS owns petroleum refinery, based in the major industrial city of Gdansk, in the North of Poland. The refinery’s output capacity reaches 6milliontonnes of crude oil per year. Industrial plant refines three different types of crude oil – Ural, Kuwait and Rozewie from the Baltic Sea. The refinery processes light and heavy crude oil in blocked out operations. The run time for each type of crude depends upon availability of crude oil from different sources.
Grupa LOTOS engaged Honeywell to implement advanced process control (APC) solutions in the crude and vacuum distillation complex to maximise the crude processing capacity constraints to hardware limitations like furnace heat load. This project was split in to two phases so that APC in the crude distillation unit could be implemented after a revamp of crude furnace. The first stage concerned the advanced process control of the flash column, naphtha splitter and the vacuum distillation unit. The second phase saw the advanced process control of the crude distillation unit. The principal aim of crude unit implementation was to get the unit to increase its throughput from the existing 495tonnes per hour to its maximum capacity. A further objective was to enable the existing furnace to continue in full operation until the company brought in a new replacement furnace in a scheduled turnaround period some two years later, 2002.
In the first phase, the Honeywell APC implementation was a great success as it enabled Grupa LOTOS to push furnace heat duty to its hardware constraints and allow it to increase daily throughput until the next turnaround – a period that lasted over two years. The objective of APC implementation on the naphtha splitter was to maintain the benzene content of the gasoline pool. The naphtha splitter produces light naphtha, sent to an isomerization unit, and heavy naphtha which it processes in the continuous-catalyst-regeneration (CCR) unit to make high octane gasoline. The profit controller implemented with controlled variables as a pressure compensated top and bottom temperatures helped to minimise benzene precursors in to light naphtha so that when heavy naphtha is processed in CCR the high octane gasoline should have minimum benzene content.
P hase two of the project was started after a revamp of the crude distillation unit furnace. The commissioning of the new crude furnace had shifted feed maximisation constraints from the crude unit to the vacuum unit. Now the crude unit furnace could handle a higher feed rate than the vacuum unit furnace. The challenge for this phase of the implementation was to process the maximum feed rate in the crude unit with minimum production of atmospheric residue which means no slippage of the diesel component into the vacuum unit.
Two Profit Controllers were implemented on the crude distillation unit, one on the furnace and another for the preflash and crude unit fractionator. The two Profit Controller applications were designed and installed as opposed to one single large application in order to maximize the flexibility and minimize the complexity of the advanced control system. These applications were designed to maximise product recovery in crude distillation unit and minimize atmospheric residue. The main objectives of these applications are:
o Minimise atmospheric residue (maximise product slate value).
o Optimise fractionation between key component streams.
o Minimise disturbances during crude switch.
o Minimise give-away of product and quality.
o Manage main column hydraulic loading within constraints.
The main challenge of this project was to recover the maximum diesel component from the atmospheric residue to minimise the total amount of atmospheric residue, so that the feed to the crude unit can be maximised. The quality control of other side streams was equally important. Although the lab analysis of products was completed two to three times a week, variations in crude quality make it difficult for operators to control product quality near the required specifications. Another problem was that the heavy gas oil draw-off tray above the flash zone section often ran dry which resulted in coloring of the diesel product. Before the implementation of inferential control, these difficulties caused the quality of products to vary widely from desired specifications.
The main objective of the inferentials installation was to improve product quality control during normal operation and better control of fractionator product qualities during crude switchovers.
Honeywell Process Solutions used its proprietary crude switching and cutpoint technology in order to minimise the time between crude switches. This is accomplished by calculating a true boiling point (TBP) curve on-line every minute. As the column conditions change during a crude switch and the crude unit transitions to a new steady state, the TBP slope changes depending on the volatility of the crude being processed. Any change in crude quality can be instantly predicted by a change in the TBP slope or volatility. This scheme gives a benefit of increasing or decreasing side products as crude volatility changes without waiting for the slow feedback from either analysers or laboratory readings (analyses).
The calculated variables used in the inferred properties react directly and instantaneously to changes in the crude column operation and different crude types or changes due to stratification in the tanks of the same crude type. This reaction precedes feedback from the laboratory or from on-line analysers (where available and reliable) thus allowing tighter control of these properties using the Profit Controller.
All inferential models are configured using a hybrid approach of crude fractionator cutpoint calculations and Honeywell's Profit Sensor Pro application. The basis of the inferential modelling is a minimalist approach using the minimum number of inputs that are required for a good model. This is based on Honeywell's experience of inferential modelling.
The naphtha 90percent point is economically important because it affects the production of gasoline. The methodology used to calculate naphtha 90percent is to calculate the naphtha cutpoint from a first principles model. The naphtha cutpoint model relies on the measurement of all parameters, which affects the partial pressure of naphtha at the top of the fractionator. The graph below represents the performance of inferentials for the naphtha 90percent point for the preflash column and main fractionator.
The LGO90% and MGO90% was also inferred using TBP cut point. These inferred properties were used to maximise diesel recovery. These properties were controlled variables in the profit controller with an objective to maximise 90percent distillation point of these products.
Grupa LOTOS uses the feedback from the laboratory through Honeywell’s LIMS laboratory information management system to improve the inferential predictions. The lab update is bundled with the Profit Sensor Pro software. This allows the laboratory values to trim the inferred model of the final product composition. The model bias is updated every time a new (and validated) laboratory sample is passed back to the DCS through the plant historian (Honeywell’s Uniformance PHD) or manually. The laboratory reading goes through a series of tests (typically bad value, range checking etc.) to ensure that the reading is valid before it is used in the correction of the bias and even then the bias is filtered to ensure that only underlying smooth process changes are fed back in the inferential models.
Conclusion
Through the implementation of Honeywell’s advance process control solution, the Grupa LOTOS refinery has been able to significantly improve process operations, with increases in throughput, product quality and overall plant profitability. The profit controller applications with inferentials were able to reduce atmospheric residue production by 1.3percent. The application of Honeywell advanced control technology to the crude unit has resulted in significant savings of approximately $0.27USD/bbl of crude only by upgrading lower value products. This doesn't include feed maximisation benefits.
In conclusion, using advanced process control technology has significantly improved production throughput and have a huge impact on the Gdansk refinery's profitability.o
Mariusz Zienkiewicz is Operations Lead, and Krzysztof Wrzesien is Plant Manager at Grupa LOTOS, Gdansk, Poland. Bhupinder Bikram and Phillip Allen are Principals Consultant at Honeywell, Bracknell, UK. www.honeywell.com
