U.S. Refinery Vulnerability Model: Collapse to Recovery Trajectory

 Journey to a Faraway Beatville

Introduction

The complex refining capacity theory begins from a structural observation: the United States can produce large volumes of crude oil while remaining acutely vulnerable to the sudden loss of a small number of large refineries. Crude oil is not a finished fuel. It must be converted into gasoline, diesel, jet fuel, and other products through a capital-intensive, continuous-flow refining system that has become highly concentrated and tightly optimized. A handful of Gulf Coast facilities account for a substantial fraction of national distillation capacity and routinely operate at utilization rates near 95 percent, leaving limited spare capacity or physical buffers.

Under these conditions, the simultaneous or near-simultaneous loss of the largest plants is modeled as initiating a cascading, multi-sector decline rather than a simple proportional reduction in fuel supply. Because the system is organized as a just-in-time chain—upstream production feeding pipelines, pipelines feeding refineries, refineries feeding product pipelines and terminals, and terminals feeding retail and freight—disruption propagates with characteristic time lags. Upstream extraction faces rising back-pressure and storage constraints; midstream pipelines lose pressure and throughput; retail inventories are rapidly drawn down, especially under panic buying; and diesel-dependent freight, agriculture, and logistics encounter progressive rationing. Each sector is described as approaching its own asymptotic operating floor on a different timescale.

These dynamics can be expressed mathematically as first-order processes of the form \(\frac{dx}{dt} = -k(x - L)\), where (x) is the operating level of a given sector, (L) is its effective floor, and the rate constant (k) aggregates inventory cover, utilization slack, import flexibility, behavioral responses, and the strength of inter-sector coupling. Larger (k) or lower (L) produces faster decay; thicker inventories, demand destruction, or the ability to import refined products slows the approach to the floor. Recovery, when major refining capacity returns, follows the inverse process with additional lags, as downstream sectors cannot fully reconstitute until product flows resume.

The theory therefore treats refining capacity not as a static volume but as the critical, low-redundancy node in a continuous-flow energy system. It is a hypothetical model of simultaneous major disruption intended to illustrate systemic fragility and the time-dependent, multivariate character of cascading failure, rather than a base-case forecast or an observed historical outcome.

• The X post by @jameslangellepromotes a blog article warning that the US refining sector's extreme concentration and 95% utilization rate leave it vulnerable to rapid collapse, with the top 5 Gulf Coast refineries alone representing over 17% of national capacity.
• Accompanying charts detail the largest US refineries by throughput, major capacity losses from recent closures and accidents (e.g., PES Philadelphia at 335k bpd, Shell Convent at 240k bpd), and a projected 60-day timeline of escalating shortages in fuel, pipelines, and freight.
• The piece connects refining fragility to broader events like potential diesel market shocks from Russia waivers and SPR limitations, arguing a major disruption would trigger immediate upstream shut-ins, retail panic, and macroeconomic contraction without refined product buffers.

map the multivariate cascading effect across all energy sectors as a function of time

Hypothetical 60-day cascading collapse (from the blog linked in the X post) after simultaneous loss of the top five U.S. refineries (~3.08 million bpd, ~17.5% of capacity). Refineries normally run near ~95% utilization with minimal buffers, so the system has little spare capacity to absorb the shock.

Capacity / Operating Levels by Sector Over Time

Sector	Day 0–1	Day 3	Day 7	Day 14	Day 30	Day 60
Refining (Top 5)	Immediate → 0%	0%	0%	0%	0%	0% (crisis floor)
Upstream Crude Extraction	100%	40%	~25%	15%	~15%	Long-term damage risk
Midstream Pipelines	Backpressure begins	10%	0%	0%	0%	0%
Retail Fuel Availability	Normal (3–5 day buffer)	50%	5%	5%	5%	Persistent shortages
Freight / Logistics	Normal	Impacted	70%	Worsening	25%	Gridlock

Key propagation points from the blog:

• Hours 1–24: Pipeline backpressure; wells continue pumping into storage.
• Days 2–5: Colonial Pipeline (feeds ~45% of East Coast liquid fuels) loses pressure; aviation fuel rationing begins.
• Day 3: Upstream extraction falls as storage fills and force majeure hits; retail fuel drops amid panic buying.
• Day 7: Pipelines unpressurized; retail at ~5% stock; freight capacity already declining.
• Days 15–30: Diesel shortages create trucking/rail/port/agriculture quotas; freight bottoms near 25%.
• Days 31–60: Broader macroeconomic contraction (stagflationary effects on manufacturing, plastics, fertilizers, etc.). SPR crude releases provide little near-term relief because refined-product capacity is offline (export/re-import lag of 30–45 days).

The chart below approximates these trajectories (refining, upstream, pipelines, retail fuel, and freight) as a function of time:

Context notes:

• Current U.S. refining utilization is high (national ~93–94%, Gulf Coast often 95–98% in recent EIA weekly data), consistent with the blog’s premise of limited spare capacity.
• The scenario is a modeled “worst-case simultaneous failure” of the largest plants, not a base-case forecast. Real disruptions (hurricanes, accidents, maintenance) have historically been more localized or staggered, with imports and inventory drawdowns providing some mitigation.

This is a combined asymptotic collapse of the national energy sector.

The blog’s scenario models a rapid, multi-sector decline in the U.S. petroleum liquids system that approaches low operating floors rather than instantaneous total zero. Those floors are what produce the asymptotic shape: most curves fall steeply in the first 1–14 days, then flatten near residual levels set by remaining capacity, imports, inventory, or physical constraints.

Key asymptotic behaviors in the model:

• Refining (top five plants): Step drop to ~0% within ~1 day and stays there (hard capacity loss).
• Midstream pipelines: Exponential-style pressure collapse to 0% by roughly Day 7 (Colonial and similar product lines).
• Upstream crude extraction: Approaches a ~15% floor by Day 14 as storage fills and shut-ins occur (some residual production possible where takeaway or storage remains).
• Retail fuel availability: Approaches a ~5% floor by Day 7 after panic buying exhausts station and terminal buffers.
• Freight/logistics: Approaches a ~25% floor by Day 30 as diesel rationing and port/rail constraints bind.

An asymptotic approximation of these trajectories looks like this:

This is a collapse of the petroleum product chain (crude production → refining → product pipelines → retail fuels → diesel-dependent transport and agriculture), not of the entire national energy sector. Petroleum accounts for roughly 37–38% of U.S. primary energy consumption (36 quads of ~96 total in recent EIA data). Natural gas (36%), nuclear (8–9%), renewables (9%), and coal (~9%) are not directly modeled as failing in the same way; their primary risk is secondary (logistics, chemical feedstocks, and economic contraction). Electricity generation and natural-gas systems have different physical bottlenecks and larger inventory or ramping options in most cases.

The “combined asymptotic” character therefore comes from the tight coupling and lack of buffer in the continuous-flow oil system itself: once the largest refineries are offline, the upstream and downstream sectors each settle toward their own constrained floors on different timescales. The blog presents this as a hypothetical simultaneous-failure case, not an observed or base-case outcome.

now graph the corresponding graph as the major refineries are back online to full capacity

The original blog and earlier charts modeled only the collapse phase after the hypothetical simultaneous loss of the top five U.S. refineries. They do not provide a corresponding recovery timeline. The chart below is an illustrative recovery model (not from the source material) showing how the same sectors could ramp back once the major refineries begin restarting and return to full capacity.

Recovery is lagged rather than simultaneous:

• Refining capacity leads the recovery, rising toward 100% as units complete startup, testing, and product blending.
• Midstream pipelines re-pressurize and resume flow once refining output begins.
• Upstream crude extraction increases as takeaway capacity returns and storage pressure eases.
• Retail fuel availability rebuilds after product reaches terminals and stations.
• Freight/logistics recovers last, as diesel and other fuels become reliably available across the system.

In practice, real refinery restarts after major outages are staged, safety-constrained, and often take days to several weeks depending on the cause of the outage, feedstock availability, and downstream inventory positions. The curves above are smooth exponential ramps for visualization; actual recoveries are typically stepwise and uneven across units and regions.

explain the nature of the "smooth exponential graphs" in terms of a dx/dt

The smooth exponential curves in the earlier charts are solutions to first-order linear ordinary differential equations of the form

\[ \frac{dx}{dt} = k(L - x) \]

where

• ( x(t) ) is the operating capacity or availability of a given sector (as a percentage),
• ( L ) is the asymptotic limit (the floor the system approaches during collapse, or 100% during recovery),
• \( k > 0 \) is a rate constant that sets how quickly the sector moves toward ( L ).

Collapse phase

During the modeled collapse, each sector is driven toward a low floor ( L ). The governing equation is

\[ \frac{dx}{dt} = -k(x - L) \]

(the negative sign produces decay). Separating variables and integrating yields the explicit solution

\[ x(t) = L + (x_0 - L)\,e^{-kt} \]

where \( x_0 \) is the initial value (usually 100%).

• Large ( k ) produces a steep early drop (pipelines, retail fuel).
• Small ( k ) produces a slower approach (freight).
• Once ( x ) is close to ( L ), \( \frac{dx}{dt} \) becomes small, so the curve flattens—the asymptotic behavior visible in the charts.

Recovery phase

When refining capacity begins to return, the same form is used but now \( L = 100\% \) (or a sector-specific ceiling) and the initial condition is the low floor left by the collapse. The equation is again

\[ \frac{dx}{dt} = k(L - x) \]

with solution

\[ x(t) = L - (L - x_0)\,e^{-kt} \]

(or an equivalent delayed version \( x(t) = L - (L - x_0)\,e^{-k(t-\tau)} \) for \( t > \tau \) to represent the lag before a downstream sector can respond).

The exponential term again forces \( \frac{dx}{dt} \to 0 \) as \( x \to L \), producing the smooth, concave-down approach to full capacity seen in the recovery plot. Different sectors simply use different values of ( k ) and different lags \( \tau \), reflecting that refining must lead, pipelines and retail must follow, and freight recovers last.

In short, every smooth exponential segment in the graphs is the integral curve of a linear restoring force \( k(L - x) \) that weakens as the variable nears its asymptote.

what might speed up or slow down the original "decay" equation?

The original decay equation used for the collapse charts was the first-order form

\[ \frac{dx}{dt} = -k(x - L) \]

where (x(t)) is the operating level of a sector, (L) is its asymptotic floor, and \(k > 0\) controls the speed of approach to that floor. Anything that increases the effective value of (k) (or lowers (L)) speeds the decay; anything that decreases (k) (or raises (L)) slows it.

Factors that speed the decay (larger effective (k))

• Thinner initial buffers. Low commercial product inventories, nearly empty terminal tanks, or stations already near minimum operating levels make the early drop steeper. The system reaches the “panic-buying / stockout” regime faster.
• Higher pre-existing utilization. When refineries, pipelines, and trucking are already near 95%+ capacity, there is less unused margin to absorb the loss of the top five plants. The same absolute capacity reduction produces a larger percentage shock.
• Stronger coupling and just-in-time behavior. Tight integration between refining, product pipelines (e.g., Colonial), and downstream distribution means a loss of pressure or product at one point propagates with little delay. Panic buying multiplies the effective demand rate and therefore (k).
• Concurrent stresses. A simultaneous weather event, cyber disruption, or labor/logistics failure adds an extra negative term, effectively increasing (k) or lowering (L) for multiple sectors at once.
• Demand inelasticity in the short run. Essential freight, agriculture, and emergency services continue to pull fuel until physical stocks are exhausted, accelerating the decline toward the floor.

Factors that slow the decay (smaller effective (k) or higher (L))

• Larger commercial and strategic inventories. Higher gasoline, diesel, and jet-fuel stocks at terminals and in the SPR (even if the SPR holds mostly crude) stretch the time before retail and freight hit their floors, reducing the observed (k).
• Ability to import refined products. Rapid cargo arrivals or increased pipeline/rail imports from Canada or overseas raise the effective floor (L) and slow the net rate of decline.
• Demand destruction. Sharp price spikes cause immediate cutbacks in discretionary driving, some industrial use, and non-essential freight. This lowers the draw rate on inventories and therefore reduces (k).
• Rationing, prioritization, or allocation rules. Government or industry measures that reserve fuel for critical sectors (military, emergency services, food transport) flatten the decline for those users and can raise the system-wide floor.
• Partial or staged capacity retention. If only some units at the major refineries go offline, or if smaller plants and remaining operable capacity can be maximized, both the initial drop and the subsequent decay rate are smaller.
• Alternative logistics or routing. Ability to shift product by rail, barge, or truck, or to reverse or reconfigure pipelines, adds temporary throughput and slows the approach to the pipeline and retail floors.

In short, (k) is not a fixed physical constant; it aggregates inventory cover, utilization slack, import flexibility, behavioral responses (panic vs. conservation), and the degree of system coupling. The steeper curves in the original model correspond to the low-buffer, high-utilization, tightly coupled case the blog emphasized; real-world conditions that increase buffers or reduce demand would produce visibly slower exponential decays.

Glossary of Names, Places, and Terms

Names

• James C. L’Angelle (@jameslangelle): Author of the X post and linked blog post that present the refining-capacity vulnerability model and the associated 60-day cascading-collapse scenario.
• Motiva Enterprises: Operator of the largest U.S. refinery (Port Arthur, Texas).
• Marathon Petroleum: Operator of major plants at Galveston Bay (Texas) and Garyville (Louisiana).
• ExxonMobil: Operator of large refineries at Beaumont, Baytown (Texas), and Baton Rouge (Louisiana).
• Colonial Pipeline: Major product pipeline system that historically supplies a large share of liquid fuels to the U.S. East Coast.

Places

• Port Arthur, Texas: Site of Motiva’s flagship refinery and other large facilities; frequently cited as the single largest U.S. refining complex.
• Galveston Bay, Texas: Location of Marathon’s large coastal refinery.
• Garyville, Louisiana: Site of Marathon’s major Louisiana refinery.
• Beaumont and Baytown, Texas: Locations of two large ExxonMobil refineries.
• Baton Rouge, Louisiana: Site of a major ExxonMobil refinery.
• Gulf Coast (PADD 3): Petroleum Administration for Defense District that contains the majority of U.S. refining capacity, including the top-ranked plants in the model.
• Cushing, Oklahoma: Primary U.S. crude-oil storage and pricing hub referenced in discussions of upstream back-pressure and storage constraints.
• East Coast (PADD 1): Region heavily dependent on Colonial Pipeline and other product supply routes; highlighted for vulnerability once midstream flows decline.

Terms

• Refining (distillation) capacity: The volume of crude oil that can be processed into finished products (gasoline, diesel, jet fuel, etc.). Distinct from crude production; the theory treats it as the critical bottleneck.
• Utilization rate: Percentage of operable refining capacity actually in use. The model emphasizes sustained rates near 95%, implying limited spare capacity.
• Top-five refineries: The largest plants (Motiva Port Arthur, Marathon Galveston Bay, Marathon Garyville, ExxonMobil Beaumont, ExxonMobil Baytown) whose simultaneous loss is the modeled trigger (~3 million barrels per day).
• Cascading collapse: Sequential, time-lagged failure across linked sectors (upstream production → midstream pipelines → refining → retail fuel → freight/logistics) once major refining capacity is lost.
• Asymptotic floor ((L)): The low operating level toward which a sector declines and then stabilizes in the decay model (e.g., retail fuel near 5%, freight near 25%, upstream near 15%).
• Decay equation \(\frac{dx}{dt} = -k(x - L)\): First-order differential equation used to generate the smooth exponential approach of each sector toward its floor. (k) is the rate constant; larger (k) produces faster decay.
• Upstream crude extraction: Oil-well and field production. Modeled as declining once pipeline back-pressure and storage limits force shut-ins.
• Midstream pipelines: Crude and product transportation systems (including Colonial). Modeled as losing pressure and throughput within days of major refining losses.
• Retail fuel availability: Gasoline and diesel stocks at terminals and service stations. Modeled as collapsing rapidly under normal draws plus panic buying.
• Freight / logistics: Trucking, rail, ports, and agriculture that depend on diesel. Modeled as declining more slowly toward a reduced operating floor.
• Strategic Petroleum Reserve (SPR): U.S. government crude-oil stockpile. The theory notes that SPR releases supply crude, not refined products, and therefore provide limited near-term relief at the pump.
• Just-in-time / continuous-flow system: Operating model with minimal inventory buffers; disruption at one node quickly affects upstream and downstream nodes.
• Force majeure: Contractual declaration used by pipelines or producers when they can no longer accept or deliver product, triggering shut-ins.
• PADD: Petroleum Administration for Defense District; regional divisions used by the EIA to report refining and petroleum statistics (PADD 3 = Gulf Coast).
• Recovery trajectory: Illustrative inverse process in which sectors ramp back toward full capacity once major refining units restart, with characteristic lags between refining, pipelines, retail, and freight.

Grok’s role in this assessment was analytical and synthetic rather than originative. The core refining-capacity vulnerability model, the identification of the top-five Gulf Coast plants, the 60-day cascading timeline, and the sector-specific floors were drawn from the X post and linked blog by James C. L’Angelle.

Grok extracted and organized those claims, cross-referenced them against publicly reported EIA utilization and capacity data, and rendered the described dynamics into explicit mathematical form (the first-order decay equation \(\frac{dx}{dt} = -k(x - L)\) and its recovery counterpart). It then generated the illustrative multivariate and asymptotic charts via code execution, produced the recovery trajectory visualization, compiled the glossary of names, places, and terms, and drafted the introductory framing.

In short, Grok translated the source material into structured tables, differential-equation representations, and visual models that make the hypothesized time-dependent, multi-sector cascade easier to examine and discuss. The underlying scenario and its assumptions remain those of the original author.