Canadian bitumen production currently runs about 1 MMbpd, with some being sold as Synbit and Dilbit. Over the next 10-12 years output is expected to increase to 3.5 MMbpd and more refiners will begin investing to process it and come to depend on the Synbit and Dilbit for a significant part of their supply. Few today, however, have ever processed these feeds at high blend ratios, and are unaware that conventional process and equipment designs are not up to the job. Canadian oil sands to desalt, difficult to vaporize, thermally unstable, corrosive, and produce high di-olefin product from the coker. If you intend to lock into a long-term supply, therefore, it is imperative that you consider reliability and run length from a particular design.

Too low tube velocity in the vacuum heater tubes will lead to precipitation of asphaltenes. Too fast a flow rate will erode the tube bends. If coil layout, burner configuration and steam rate are not correct, run length will be measured in months, not years. Diluent recovery unit designs must take into account possible upsets from water slugs and other unpredictable situations that have damaged internals, resulting in diluent losses and high vacuum unit overhead condensable oil. Diluent is neither cheap nor plentiful, and high vacuum column operating pressure will reduce overall liquid volume yields. And if the design of the delayed coker fractionator is based on today’s experience with conventional heavy feedstocks you will be lucky to run six months.

What all this means is that special process and equipment designs are needed to satisfy the special demands of processing oil sands crudes. Such processes are not generated by computer based designers who have little or no experience and never leave the office. They are developed only by engineers with know-how who have real experience wearing Nomex® suits and measuring true unit performance in Northern Alberta. Shouldn’t this be kept in mind by those considering long term supply agreements?