By John Mayes
In our past blog series, “It Ain’t Heavy, It’s My Bunker,” we looked at the impacts that the upcoming International Maritime Organization (IMO) regulations on the future of bunker fuel sulfur will have. The pivotal meeting is scheduled for next week (October 24-28) and will dictate the timing of lowering sulfur specifications for bunker fuel to 0.5%. A major input to that decision is the conclusion of the CE Delft study on the likely availability of low sulfur bunker fuel in 2020, titled, “Assessment of Fuel Oil Availability.” We have reviewed this report and have determined the tone is very much, “Don’t Worry, Be Happy.” In analyzing the report closer, though, there is a lot to be worried about, especially for those that produce significant amounts of bunker fuel. In most cases, this means simpler refiners in Europe and Asia. At the same time, complex facilities in the U.S. and elsewhere will “Be Happy” as distillate prices will be positively impacted. Additionally, the report over-optimizes and over-simplifies the refining market in their model that draws these conclusions, as well as taking advantage of quantities of desulfurization equipment that does not exist. In this week’s blog, we walk through the report, highlighting some of the key points that could have serious impacts for both refiners and the bunker markets very soon.
The findings of this report are important in that if it determines that the supply of 0.5% sulfur bunker in 2020 is insufficient to meet global demand, the IMO has the authority to delay the implementation date from January 2020 to January 2025. One of the conditions placed on the study was that all refining output must have a demand source. This caveat would seem to eliminate the potential solution of diverting large volumes of distillate into the bunker pool and backing out high sulfur fuel oil with no alternate disposition. The Delft study concluded the required volume of low sulfur bunker fuel could be produced in 2020 without a significant diversion of distillate fuel.
Delft Study Basis and Results
The primary tool of the Delft study was a linear programming (LP) model. This tool is a very important asset for refiners and offers the ability to optimize a refinery’s operation utilizing tens of thousands of calculations. The LP model is prepared with specified inputs and computes the optimum solution within those constraints. The model, however, cannot recognize an illogical result, and only responds with those which are mathematically correct. It is incumbent upon the user to vet the output to ensure the logic and integrity of the results.
To establish a baseline, the study calculated a global balance for 2012. Many of the results shown in the Delft study are highly questionable. Table 6 on page 19 of the study for instance, states that the calculated crude unit utilization rates in North America (the U.S., Canada, Puerto Rico and the U.S. Virgin Islands) in 2012 were only at 64% (based on crude throughput divided by capacity in barrels per stream day).
The EIA indicates the U.S. 2012 crude unit utilization was 88.7% on a barrels per calendar day basis, which would equate to around 83.3% using barrels per stream day. In its Statistical Review of the World, BP reports the 2012 rates for the U.S. and Canada were 84.2% and 87.9% respectively on a barrels per calendar day basis which would equate to 79.1% and 82.5% respectively using barrels per stream day. Even adjusting for the Hovensa shutdown in early 2012, the Delft utilization for North America in 2012 is low by 12-15%. In spite of this, the final sentence on page 18 states, “The utilization rates are plausible and indicate that the model was appropriately calibrated.” Many might take issue with that statement.
Other results of the study are even more questionable. To achieve compliance with the low sulfur bunker requirements (in addition to the existing and future requirements for other products like gasoline and diesel), the model reoptimized the global refining system for 2020. This reoptimization produced many illogical results. The model for instance, indicates that coker utilization rates in North America would decline from 88% in 2012 to only 70% in 2020 (see Table 35 on pages 48 and 49 of the report). In Europe, coker rates would fall to only 48%, while in the former Soviet Union countries, coker rates would plunge to 38% of capacity. Even more severe results are seen for Isomerization rates, which would decline to 11% in the Middle East and 13% in the FSU. While mathematically feasible, the results are simply illogical. A sample of these is shown in Table 1.
Image may be NSFW.
Clik here to view.
An additional issue of the model output is in the repurposing of existing assets. Table 96 on page 117 of the report indicates that 1.2 million BPD of treated atmospheric resid will be blended into the bunker pool in Asia. A quick review of the current Oil & Gas Journal Worldwide Refining Survey (the stated source for most of the refining capacity data for the Delft study) indicates there are currently 17 resid hydrotreating units in Asia with a combined capacity of one million BPD. Table 60 on page 96 of the report indicates an additional 255 MBPD is planned to be in operation before June of 2019. This would seem to validate the potential to produce this volume of treated atmospheric resid. The problem, of course, is that the existing units are already in alternate service, generally to produce low sulfur fuel for electricity generation. Are these utility plants supposed to close? Again, mathematically correct, but illogical.
An Improved Analysis
The basic methodology of the Delft study was flawed. The IMO has no control over global refining assets and cannot mandate the refining operating parameters as outlined in the Delft study. Neither can the IMO repurpose other assets (like utility operations) to free up refining capacities in order to assist in the production of low sulfur bunker fuel. A study dependent on these actions is of little value.
A better methodology would have been to start with current global bunker parameters and then determine what needs to occur to produce 0.5% sulfur bunker fuel. Estimates for the average global sulfur level in bunker fuel (excluding MGO and MDO) vary, but are generally around 2.5% on a 2012 volume of 4.2 million BPD (228 million tonnes per year). A more realistic assessment should have been conducted on the assumption that global refiners would continue to optimize their refining assets based on their needs and finances rather than the IMO’s objective of producing low sulfur bunker fuel, particularly coking refineries which generally do not even produce residual fuel (with the possible exception of slurry oil).
Back to Reality
Regardless of the conclusions of the Delft study, if the IMO adheres to the January 2020 compliance date, a significant volume of the world’s low sulfur bunker fuel will be produced by blending large volumes of distillate fuels. As discussed in the blog two weeks ago, “It Ain’t Heavy, It’s My Bunker – Part 4: What Do Impending New Bunker Specs Mean for Refiners,” this will likely result in substantively higher global distillate prices and much lower high sulfur fuel prices for the surplus resid which has been backed out of the bunker pool. Because of the long lead time to construct new cokers, these pricing gyrations could last up to five years.
While the Delft study indicates that low sulfur bunker fuel can be produced without distillate diversions, the inability to control the operating parameters of the global refining industry renders the results meaningless. Ultimately we disagree with the study’s methodology and its worthiness in determining the implementation date to produce low sulfur bunker fuel.
The results of the IMO decision will have significant impacts on the industry. These impacts are discussed and analyzed as part of our 2016 Crude and Refined Products Outlook. For more information on our Outlook product or for any other services we may be able to provide, please send us an email or give us a call.