On January 1st, 2020, the global shipping industry will undergo a radical change, with all ships having to reduce the sulfur content within marine fuels from 3.5% to 0.5%, as mandated by the International Maritime Organization (IMO). As with all radical changes, winners and losers await, meaning there is significant opportunity everywhere.
In this paper, DrillingInfo uses a refinery optimization model to look at the potential effects of this regulation for a single complex refinery with a delayed coking unit.
What does the change mean for refineries?
Bunker fuel provides a valuable outlet for refineries as it has the highest sulfur content specification out of the major refinery products. The heavy bottom of the barrel fractions that have too much sulfur content to become diesel (0.0015% maximum sulfur content), No.2 oil (0.05%), or light sulfur fuel oil (1%) can still generate revenue as long as they meet the 3.5% specification for marine fuel in non-emission control areas (ECAs). Furthermore, the markets for light sulfur fuel oil and other relatively high-sulfur products are small compared to that of 3.5% bunker fuel (Figure 1) so as the new specification takes effect there will not be another “release valve” for high-sulfur fractions above the 0.5% threshold.
With the new IMO specification, the residual fractions that used to go into marine fuels will largely become waste unless they can be further processed and desulfurized. Refineries with delayed coking units stand to benefit from this change as they have the capability to process these fractions into other products, which most refiners lack.
What does a single refinery model look like?
The most standard refinery processing unit is an atmospheric distillation unit (ADU), which separates crude oil into different fractions depending on their volatility. This is the process which splits crude oil into the gases, naphthas, kerosene, gas oils, and residues that it consists of. The yields of these fractions, expressed as a percentage of the crude oil input, and their physical properties, including sulfur content among others, can vary greatly between different crude oil assays. In this case we are particularly interested in a heavy assay that produces a lot of residue, the type that is often refined in the US Gulf Coast, because the ability to process this type of crude gives complex coking refineries their advantage in the market. The yields of a standard Western Canadian Select (WCS) assay are illustrated in Figure 2.
After the fractions exit the ADU, their path through various refinery units is anything but standard. The flows of these fractions through the refinery depend not only on the available processing units and their capacities, but also on product specifications and prices as the refiner will decide how much of each fraction to send into each unit in order to maximize profit. This type of problem – maximizing a profit objective subject to operational constraints – is the perfect application for linear and nonlinear programming.
In this paper we utilize the OptiFlo-Crude model to look at a complex refinery based on a Gulf Coast facility with some modifications made for illustrative purposes. Our refinery has an ADU capacity of 335 MBbl/d, and it has the following units: vacuum distillation unit (VDU), hydrotreaters for naphtha (NHT), kerosene (KHT), and gas oils (GOHT), light naphtha isomerization (LNIS), hydrocracking unit (HCU), fluidized catalytic converter (FCC), reformer (RFO), alkylation unit (ALK), and delayed coking unit (DCU).
A typical refinery blends different assays together to maximize profit while considering the physical properties of the assays together with the state of the transportation network (bottlenecks, cost of transport, crude availability, etc.). While this blending is very important on a macro scale and is being modelled at DrillingInfo, for this single refinery case we consider only the WCS assay.
How does the refinery operate with current specifications and prices?
In the base case scenario, the selected refinery’s operations are optimized given today’s 3.5% maximum sulfur specification for bunker fuel. The past 12-month average prices for end products are used as inputs. Figure 3 illustrates the basic flow diagram for this refinery, from the ADU into the other units.
The optimal product outputs for this refinery are displayed in Table 1. On a volume basis, the outputs are 42% diesel, 35% gasoline, 8% jet fuel, and a combined 15% for all other products. Some demand constraints were imposed on the model to avoid overproduction that would flood a low-demand market with supply. For example, the 3.5 MBbl/d of light sulfur fuel oil produced at this refinery already makes up more than 6% of US production of this product (Figure 1) so the light sulfur fuel oil output was limited to 1% of the refinery’s production.
What happens when IMO regulations are implemented and the price of bunker fuel rises?
In scenario 2, the IMO 2020 regulations are implemented and the price of the new bunker fuel is set accordingly. Since the maximum sulfur specification of the new bunker fuel (0.5%) lies between that of No.2 oil (0.05%) and light sulfur fuel oil (1%), its price is set using an average of these two prices weighted by the strictness of the specification. The resulting price is $74.07/Bbl compared to $85.32 for diesel, $81.98 for No.2 oil, $65.29 for light sulfur fuel oil, and $62.11 for the current 3.5% sulfur bunker fuel.
The optimal product outputs for this refinery are displayed in Table 2. Interestingly, this refinery increases bunker fuel production from 7.2 MBbl/d in the base case to 23.2 MBbl/d with the new specification and price as it reduces the production of diesel by 17.4 MBbl/d.
The spec change and price increase in bunker fuel caused the refinery to shift hydrotreated lower sulfur fractions away from the diesel pool and into bunker fuel. The composition of bunker fuel (Figure 4) changed from mostly ADU/VDU fractions in the base case to mostly DCU/GOHT fractions in scenario 2. The light gas oil (1.4% wt. sulfur), heavy gas oil (2.3%), light vacuum gas oil (3%), heavy vacuum gas oil (3.7%), and heavy cycle gas oil (2%) could be blended together to meet the old 3.5% specification, but the new specification required hydrotreated heavy vacuum gas oil (close to 0%) blended with light cycle (1.1%) and heavy cycle (2%) gas oils coming out of the DCU.
Despite higher pricing, the estimated revenue of this refinery decreases from $27.44 million in the base case to $27.23 million in scenario 2. If we impose a maximum limit of 7.2 MBbl/d on bunker fuel production based on the base case, this refinery still reduces diesel production and increases the production of waste from zero to 0.5 Mkg/d, further reducing revenue to $27.18 million. This happens because the refinery does not have enough capacity to desulfurize the residual gas oils enough to blend them into a product.
What happens when distillate prices respond?
If refiners do start to shift production away from diesel, its price is expected to rise providing incentive to meet market demand. Given the opportunity cost of producing the new bunker fuel, the price of diesel must rise by $13/Bbl for this particular refinery to return diesel production to roughly equal to the base case. To avoid other readjustments of fraction flows, the prices of jet fuel and gasoline must also rise slightly, by less than $1/Bbl. The optimal outputs of this scenario are displayed in Table 3.
Although most products are back to roughly the same level of production as the base case, the output of new bunker fuel is significantly reduced as waste production is increased. It’s interesting to note that while volume of jet fuel increases, the mass of jet fuel is the same as the base case and the volume swell is entirely due to a change in specific gravity caused by changes in its fraction composition. The expected revenue for this refinery rises from $27.44 million in the base case to $29.13 million in scenario 3.
These results illustrate why the price of diesel must rise in response to the IMO 2020 sulfur regulations and why the change is expected to benefit complex coking refineries. The $13.00 (or 15%) rise in diesel price is specific to this refinery with this crude oil assay and will vary with other refinery and assay configurations.
Refinery modelling on a macro scale
DrillingInfo OptiFlo-Crude is a nonlinear programming model that considers the entire US refining network on a macro scale, linking production (upstream) to refining demand and exports (downstream) via all the transportation paths (midstream) across the country. By understanding the demand for specific barrels by specific refineries based on the crude oil composition and refinery operations, we can predict how crude will flow, what price a certain barrel will command in the market, what pipelines and routes will be under/over utilized, where additional infrastructure is needed, where bottlenecks will emerge, and what slate of refined products will be produced in the US.
Contact us today for an introduction!