Analysis

A refinery name, a lubricant catalogue and a cluster of storage tanks answer different questions. None alone establishes what enters a plant, which transformation occurs there, or how much it can process. Energy intelligence becomes more useful when facility classification, operating status, ownership and capacity each retain their own evidence.

This engineering guide combines primary-source process definitions with an original accounting example. All example rates, water content and density are prescribed assumptions. They describe no named installation, measured production, typical industry yield or validated process design. The accompanying chart represents four assumed mass exits, not a process-flow diagram.

Start with feedstock and product, not the facility name

EIA describes crude refining through separation, conversion and treatment. EPA distinguishes several destinations for used oil, including re-refining into lubricants and processing into fuel oils. These descriptions establish that different feedstocks and product objectives can sit behind similar industrial vocabulary. They do not identify the process at a particular map location.

Ask whether the incoming material is crude, a refinery intermediate, collected used lubricant or purchased base oil. Then distinguish recovering a base stock from blending an already manufactured base stock with additives. A lubricant product list does not demonstrate crude distillation or used-oil re-refining. Likewise, selling fuel does not establish that the seller manufactures it. Preserve an unknown feedstock until a plant-specific document resolves it.

Read the process evidence at its actual scope

Appendix B of DOE’s December 2020 report describes used-oil pretreatment, vacuum separation and finishing options, including hydrogenation, solvent extraction and clay treatment. It also distinguishes base-stock production from treatment aimed at fuel products. The report documents possible routes; it is not an equipment inventory for every used-oil business.

A filtration reference therefore cannot establish a complete re-refinery. Nor does a generic hydrotreater description prove that a named site has hydrogen supply, a particular catalyst or an operating reactor. Retain document date, plant address and the exact process wording. Design announcements, commissioned equipment and current operating evidence should remain separate records.

Set a mass boundary before calculating yield

For a defined control volume, total mass conservation is dM/dt = Σṁin − Σṁout. M is contained mass and ṁ is mass flow. At assumed steady inventory, accumulation is zero and total incoming mass equals total outgoing mass. Chemical reactions rearrange material; they do not remove the need to account for it.

Choose whether the boundary covers a separation step, a production train or the whole site. Hydrogen, additives, solvents, purchased intermediates and returned material belong among the inputs whenever they cross that boundary. Water, residues, gaseous streams and material sent for treatment belong among the outputs. Internal recycle is not a second external feed. Align the accounting period before comparing totals.

Reproduce the hypothetical four-stream ledger

Assume one feed of 1000 kg/h and four grouped exits: base oil 700 kg/h, a secondary hydrocarbon stream 120 kg/h, water 100 kg/h and residue 80 kg/h. Prescribe zero accumulation and no other material crossing the example boundary. The sum is 700 + 120 + 100 + 80 = 1000 kg/h, so the accounting residual is zero.

This is a stipulated ledger, not a simulation showing that any technology achieves these splits. The grouped exits make no claim about phase, contaminant composition, discharge quality or waste classification. The zero residual follows from the chosen numbers; it does not demonstrate a zero-emission plant or prove that an actual meter closes. No hydrogen-consuming process is represented by this simplified example.

Keep wet-feed and water-free yield denominators distinct

On the total incoming-feed basis, the assumed base-oil mass yield is Y = 700/1000 × 100 = 70%. Now additionally prescribe that the feed contains 100 kg/h of water and that all of it exits in the water group. The water-free feed is then 1000 − 100 = 900 kg/h, giving Y* = 700/900 × 100 = 77.777…%, rounded to 77.78%.

The numerator did not increase. Only the denominator changed. Water-free does not mean pure hydrocarbon: other nonwater contaminants may remain. Neither percentage is a recovery fraction of chemically identified lubricant molecules. Such a component recovery would require feed and product composition. Report the basis beside the value rather than comparing these two percentages as different plant efficiencies.

Use an imbalance to investigate the ledger

In a separate fictional reporting exercise, retain 1000 kg/h incoming but suppose the reported exits total only 950 kg/h. The positive residual is 50 kg/h, or 5% of the reported feed. It is an unresolved accounting difference. Assigning the entire difference to evaporation, leakage or conversion loss would invent a physical explanation.

Check inventory movement, sampling periods, tank transfers, units, meter calibration and missing streams before interpretation. A batch inventory change cannot be compared directly with an hourly rate without a time basis. Retain original readings and corrections separately. A small arithmetic residual also cannot establish accurate individual meters: errors in different streams can offset.

Separate a mass rate from volume and capacity

For an additional illustration, prescribe base-oil density 0.88 kg/L at 20°C and take the 700 kg/h base-oil stream at that same reference condition. Q = ṁ/ρ = 700/0.88 = 795.4545… L/h, rounded to 795.45 L/h. This assumed density is not a measured property or a recommended value for a real product.

Without compatible density and temperature references, a reported tonnage cannot establish barrels per day. Even a valid conversion supplies only a rate on its declared basis. Capacity still needs to identify feed versus product, the particular unit or site, rated versus achieved throughput and the operating-time denominator. Storage volume is an inventory limit, not processing capacity. Leave an unsupported conversion unknown.

Follow a real record while preserving the example boundary

Alma Petroli’s Operations page describes atmospheric and vacuum fractionation of crude at its Ravenna plant, with bitumen among the products. This is a useful counterexample to classifying solely by the main product: bitumen specialization does not by itself exclude crude processing. The existing Refinery Atlas record osm-way-681248983 retains the corresponding classification evidence and source link.

Use the cited Atlas page and record API to examine that evidence, then follow the related mass-balance and density guides for the calculation methods. No Alma Petroli rate, yield, density or equipment performance was used in this example. The research link establishes an editorial pathway, not a used-oil feed relationship or a physical supply chain.

Keep quality and environmental claims outside unsupported arithmetic

DOE’s 2020 review discusses how system boundaries and assumptions affect energy and environmental comparisons. A closed mass ledger alone cannot establish energy savings, carbon intensity or economic advantage. A comparison needs an equivalent product function, material and utility inputs, residue treatment, transport and a stated allocation method.

Likewise, a base-oil yield does not establish viscosity, stability or a finished lubricant approval. Keep laboratory results, product specifications and commercial qualifications separate from throughput. Archive the exact source, access date, units, assumptions and calculation with each derived value. A later verification date updates the evidence review; it does not turn an old report into a current production measurement.