Why Maximum Product Yield Is Not a Property of the Refinery

       One of the most common questions in refinery production planning and refinery production optimization appears deceptively simple:

From a business perspective, the expected answer usually includes:
        - the minimum, base-case, and maximum production of each major product;
        - which operational decisions increase or decrease product yields;
        - which products must be sacrificed to produce more of another product;
        - the transition time between operating modes;
        - how the achievable production range changes with feedstock composition and throughput.
       In other words, planners are interested not only in independent minimum and maximum production levels but also in the feasible product slate—the complete set of product combinations that can be achieved under the specified operating conditions. A typical question might be:

Example
Suppose a refinery currently produces

       Management asks: Increase diesel production by 300 t/day. The natural assumption is that this requires changing only diesel production. In reality, the refinery may instead produce

No single product changes independently. The refinery moves to another feasible operating state.

       At first glance, this appears to be a straightforward optimization problem. If the feedstock composition is fixed and all material must be processed, it seems reasonable to expect that the technological limits of the refinery can be calculated once and then used for production planning. Unfortunately, this assumption is fundamentally incorrect. The concept of a single "maximum product yield" does not exist independently of the operating conditions. Before any optimization can begin, it is necessary to answer a more fundamental question: Which operating scenario are we trying to optimize? Only after this question has been clearly defined does it become meaningful to discuss the feasible operating region or the attainable product slate.

       A common misconception in refinery planning is that every refinery has a single maximum production capacity for gasoline, diesel, jet fuel, or any other product. In reality, such values do not exist as intrinsic properties of the refinery itself. A refinery is not a static production system. It is a highly configurable network of interconnected process units, storage facilities, and logistics constraints whose behavior depends on numerous operating decisions.
       Consequently, the maximum achievable production (maximum product yield) of any refinery product depends on the selected refinery operating scenario rather than on the refinery itself. It is determined by a specific combination of operating conditions. For example, the maximum diesel production obtained during winter operation may differ significantly from the maximum achievable during summer operation. Likewise, a refinery configured to maximize gasoline production will have a completely different set of attainable product yields than the same refinery configured to maximize jet fuel production.
       Therefore, asking

is incomplete. A technically correct question is

       Only after the operating scenario has been fully defined does the concept of a maximum product yield become meaningful. This distinction is fundamental. A product yield is not an absolute characteristic of the refinery. It is a result of the interaction between the refinery configuration, feedstock properties, operational objectives, and technological constraints.
       Ignoring this fact leads to one of the most common mistakes in refinery optimization—treating calculated production limits as universal technological maxima when, in reality, they are valid only for the particular conditions under which they were obtained. As a result, optimization based on such limits may recommend production planning and production scheduling decisions that cannot be reproduced under different operating conditions.

       Even when processing the same amount of feedstock with an identical composition, a refinery may produce very different product slates under different operating conditions. This is because refinery operation is constrained by numerous technological and operational factors that govern material flows and process unit utilization.
       Key scenario-defining factors include:

• seasonal operating conditions;
• production strategy (e.g., gasoline, diesel, or jet fuel maximization);
• feedstock characteristics;
• equipment availability and maintenance schedules;
• initial tank farm inventories;
• operating modes of key process units.

Together, these factors define the search space for refinery production optimization. In practical refinery digital twins, inventories, operating modes, and process constraints are represented by specialized simulation components such as Tank Farm (RpAccumulative), Flowing Tank Farm (RpFlowing), and Source.
       Each combination of these factors defines its own feasible operating region for refinery production planning.. Consequently, there is no single product slate that represents the refinery's technological capability. For example, one study may conclude that the refinery can produce 1,200 t/day of diesel, while another reports 1,050 t/day under the same feedstock conditions. Both results may be correct—they simply correspond to different operating scenarios. Therefore, product yields cannot be interpreted without explicitly defining the operating scenario. Only after the scenario has been fixed does it become meaningful to evaluate production limits, analyze product trade-offs, or search for an optimal operating point.

       The individual workflow stages are implemented by specialized refinery simulation components (Source, Plant, Tank Farm,
Product Mixer, Shipment Node) and optimization models.

       Once it becomes clear that product yields depend on operating scenarios, an obvious approach is to evaluate all relevant scenarios and combine the results into a single feasible operating region. However, this quickly becomes impractical. Even a simplified refinery model must account for production strategy, feedstock characteristics, seasonal operation, equipment availability, inventory levels, and process unit configurations. As these factors are combined, the number of possible scenarios grows exponentially. Even with a coarse discretization, the number of operating scenarios can easily reach thousands. In practice, the challenge is not computational—modern simulation software can execute large scenario studies efficiently. The real question is which scenarios should be included. Every additional assumption changes the feasible operating region. Omitting a relevant scenario may exclude physically achievable operating states, while including unrealistic combinations may produce infeasible results. Consequently, the quality of the final feasible product slate depends far more on the correctness of the scenario definition than on the number of simulations performed. Ultimately, the objective is not to calculate thousands of independent production plans, but to understand the structure of the refinery's feasible operating region and the constraints that define it. This naturally shifts the problem from exhaustive scenario enumeration to the development of an integrated optimization model.

       The discussion so far leads to an important conclusion. The primary objective of refinery production optimization is not to determine the maximum production of gasoline, diesel, jet fuel, or any other individual product. Instead, the objective is to characterize the entire feasible product slate. A feasible product slate is the set of all product combinations that can be achieved while satisfying every technological, operational, and logistical constraint imposed on the refinery. Each point within this multidimensional feasible operating space represents a physically attainable refinery operating state. Some operating points may maximize diesel production. Others may favor gasoline or jet fuel. Most practical operating conditions, however, lie somewhere between these extremes. This distinction fundamentally changes the optimization problem. Instead of asking:

the relevant engineering question becomes:

       Only after the feasible operating region has been identified does it become meaningful to search for an optimal production plan. However, increasing the production of one product almost always requires reducing the production of others or changing refinery operating conditions. For example, increasing diesel production may require:

• lower gasoline production;
• reduced jet fuel output;
• different blending strategies;
• changes in downstream process unit utilization.

       Refinery planning is therefore not a collection of independent optimization problems, but the allocation of limited technological flexibility among competing production objectives.
The value of a feasible product slate lies not in its individual production limits, but in the trade-offs it reveals. Knowing that gasoline production can vary within a certain range is far less useful than understanding how changes in gasoline production affect diesel, jet fuel, LPG, stable condensate, and other refinery products.

       The feasible product slate includes all physically achievable combinations of refinery products. However, only a subset of these operating points represents the best possible trade-offs between competing production objectives. In multi-objective optimization, this boundary is known as the Pareto Front (or Pareto Frontier). Every point on the Pareto Front is Pareto-optimal: increasing the production of one product is possible only by reducing the production of at least one other product. For example, maximizing diesel production may require sacrificing gasoline output, while increasing jet fuel production may reduce both gasoline and diesel yields. The objective of refinery optimization is therefore not to identify a single maximum product yield, but to determine which Pareto-optimal operating point best satisfies the current business objectives and operational constraints. The Pareto Front thus provides a far more meaningful representation of refinery operational flexibility than independent minimum and maximum product yields. Practical refinery optimization requires mathematical optimization techniques capable of exploring feasible operating regions.

       Identifying the feasible product slate is only the first step. For refinery production planning, it is often more important to understand how movement within the feasible region occurs than simply knowing its boundaries. Every change in refinery operation redistributes material through interconnected process units. As a result, increasing the production of one product inevitably affects the production of others. These relationships are rarely linear and almost never independent.
For example, a planner may ask:

       From an engineering perspective, this is not the right question. The more informative question is:

       The answer may involve simultaneous reductions in gasoline production, changes in jet fuel output, increased production of intermediate streams, or different blending strategies. In many cases, several process units must be operated under different conditions, making the relationship highly nonlinear. Consequently, refinery flexibility cannot be described by a collection of independent production limits. Instead, it must be described as a multidimensional system of product interactions.

       Operational decision variables define how the refinery moves through the feasible operating region. The transition between feasible product slates is achieved by adjusting the refinery's operational decision variables, rather than the product yields themselves.
       Typical decision variables include:

• process unit throughput;
• process temperatures and pressures;
• distillation cut points;
• reflux ratios;
• recycle flow rates;
• blending ratios;
• routing decisions between process units;
• equipment loading and operating modes.

In refinery digital twins, these decision variables are implemented through process unit parameters, routing logic, blending components, and flow control objects.
       These variables represent the actual control actions available to refinery operators and optimization systems. By adjusting them, the refinery moves from one operating state to another while satisfying all technological constraints. Consequently, refinery production planning is not about maximizing individual product yields, but about selecting the combination of decision variables that delivers the most valuable product slate. For decision-makers, understanding how operational decisions redistribute production across multiple products is far more valuable than knowing isolated production limits. This information enables planners to evaluate alternative refinery operating strategies and optimize refinery production planning.

       This workflow illustrates the essence of refinery production optimization. Decision variables modify the refinery operating state, redistribute material flows, generate a new product slate, and ultimately determine business performance. Learn how refinery digital twins simulate operational decisions and evaluate their business impact.

       Instead of evaluating thousands of operating scenarios individually, refinery planning systems formulate the refinery as a single mathematical optimization problem. It is to construct an integrated mathematical model of the refinery that captures:

• material balances;
• technological constraints;
• equipment capacities;
• operating policies;
• production objectives;
• interactions between process units.

       Within such a model, the feasible operating region emerges naturally from the underlying physical and technological constraints. Optimization is then performed directly within this region rather than by repeatedly evaluating individual scenarios.
This is the approach adopted by modern refinery planning systems. Instead of exhaustively enumerating operating scenarios, they search directly for the operating point that best satisfies business objectives while respecting all refinery constraints.
The result is not simply faster computation, but a fundamentally different approach to refinery planning—one that enables engineers to analyze product trade-offs, evaluate alternative operating strategies, and identify economically optimal operating points for refinery production planning and production scheduling.

       Determining that two refinery product slates are physically feasible does not necessarily mean that the refinery can transition freely between them. This distinction is often overlooked but has significant practical implications. Every refinery operates as a dynamic system with its own refinery operating state. Changing from one operating state to another requires coordinated adjustments across multiple process units, storage facilities, blending operations, and logistics systems. Some transitions may take hours, others several days, and certain transitions may be impossible without violating operational constraints. Consequently, production planning involves more than identifying feasible operating points. It also requires understanding how the refinery can move between those operating points.

       This introduces a new class of engineering questions:

• Which decision variables must be adjusted to reach a new product slate?
• Which process units limit the transition?
• How long will the transition take?
• What intermediate operating states are required?
• What operational risks or costs are associated with the change?

       These questions cannot be answered by analyzing feasible product slates alone. They require understanding the reachability of operating states. Refinery flexibility therefore has two complementary dimensions:

• Feasibility — whether a product slate satisfies all technological and operational constraints.
• Reachability — whether the refinery can realistically move from its current operating state to the desired one.

       Reachability depends on transition time, available inventories, process inertia, equipment ramp rates, intermediate operating constraints, and operator actions. Transition analysis requires dynamic simulation rather than steady-state optimization. Consequently, two product slates may both be feasible while the transition between them is impractical, prohibitively expensive, or even impossible. Modern refinery planning should therefore characterize not only the feasible operating region, but also the transitions between feasible operating states, including the required control actions, transition time, and associated operational costs.

       The question "What is the maximum production of gasoline or diesel?" appears simple, but it is fundamentally incomplete.
Without specifying the operating scenario, there is no unique answer. Even after the operating scenario has been defined, independent minimum and maximum product yields provide only limited insight into refinery capabilities. They describe isolated operating points rather than the refinery's production flexibility. The real engineering challenge is to understand refinery operational flexibility through the feasible product slate—the complete set of physically attainable product combinations—and the trade-offs that govern movement within this region. Only after this feasible operating region has been characterized does optimization become a meaningful task. The objective is no longer to maximize the production of a single product but to identify the operating point that best satisfies economic objectives while respecting every technological and operational constraint. This shift fundamentally changes the way refinery production planning and optimization problems are formulated.. Instead of asking:

Engineers should ask:

       The difference may seem subtle, but it fundamentally changes the planning process. It replaces isolated production limits with a comprehensive understanding of refinery flexibility and transforms optimization from a search for individual maxima into a search for the best operating point within the refinery's feasible operating region.

       Refinery production optimization is an essential tool for refinery planning, but every optimization model is an abstraction of the real refinery. To remain computationally efficient, models simplify refinery behavior by linearizing nonlinear processes, discretizing time, and omitting many operational constraints. As a result, optimization identifies the best operating point under the assumptions of the model, but it does not necessarily determine whether that solution is operationally achievable. Transition dynamics, equipment switching, inventory evolution, and process interactions are often beyond the scope of traditional optimization models. Simulation models and refinery digital twins complement optimization by verifying whether an optimal production plan can actually be implemented. In other words, optimization determines where the refinery should operate, while a digital twin determines whether and how it can get there. Learn how refinery digital twins complement optimization by validating production plans under realistic operating conditions.

       The central challenge of refinery optimization is not to maximize individual product yields, but to understand the feasible operating region, quantify product trade-offs, and identify operating states that are both economically optimal and operationally reachable.

1. What is a product slate in refinery planning?
A product slate is the complete set of marketable products produced by a refinery over a given planning period, together with their production volumes. Typical refinery product slates include gasoline, diesel fuel, jet fuel, LPG, stable condensate, fuel oil, and other petroleum products.

2. What is a feasible product slate?
A feasible product slate is any combination of product yields that satisfies all technological, operational, quality, and logistical constraints of the refinery. Every feasible product slate represents a physically achievable operating state.
See also: Material Balance, Tank Farm

3. Why can't a refinery have a single maximum product yield?
Because maximum product yield depends on the operating scenario. Feedstock properties, production objectives, equipment availability, seasonal operation, inventory levels, and process unit configurations all influence achievable product yields. Therefore, there is no universal maximum production value for any refinery product.

4. Why are independent minimum and maximum product yields insufficient?
Independent production limits ignore the interactions between refinery products. Increasing the production of one product usually requires reducing the production of one or more others. Production flexibility can only be understood by analyzing the complete feasible product slate rather than isolated product limits.

5. What is the difference between a feasible and a reachable product slate?
A feasible product slate satisfies all refinery constraints and is physically possible. A reachable product slate is a feasible operating state that can actually be attained from the current operating condition through realistic operational changes within acceptable time, cost, and technological limitations.

6. Why is reachability important in refinery production planning?
Production plans are executed from the refinery's current operating state, not from an arbitrary point in the feasible operating region. Some feasible product slates may require long transition periods, temporary production losses, or operating actions that are impractical or economically unjustified.

7. What are product trade-offs?
Product trade-offs describe how changes in the production of one refinery product affect the production of others. They represent the technological compromises that arise because refinery process units share feedstocks, intermediate streams, and operating capacities.

8. Why is scenario enumeration not an effective solution?
Although thousands of operating scenarios can be simulated, the main challenge is determining which scenarios accurately represent real refinery operation. Exhaustive enumeration often produces a large number of isolated operating points without providing a clear understanding of refinery flexibility or product interactions.

9. What factors define an operating scenario?
Typical operating scenarios are determined by feedstock composition, production strategy, seasonal operating conditions, equipment availability, maintenance schedules, storage inventories, process unit operating modes, and other technological or logistical constraints.

10. How do modern refinery production optimization systems solve this problem?
Modern refinery planning systems build an integrated mathematical optimization model of the refinery. Instead of evaluating thousands of independent scenarios, they represent material balances, technological constraints, equipment capacities, and operational policies within a single optimization framework. The feasible operating region emerges naturally from the mathematical model.

11. What is the real objective of refinery product slate optimization?
The goal is not simply to maximize the production of an individual product. The objective is to identify the economically optimal operating point within the feasible operating region while satisfying all technological, operational, and business constraints.

12. How does understanding the feasible product slate improve production planning and production scheduling?
Knowing the feasible product slate allows production planners to evaluate operational flexibility, understand product trade-offs, assess the consequences of changing production targets, compare alternative operating strategies, and identify production plans that maximize economic performance while remaining technically achievable.