The major operational cost of the refinery is contributed by the price of the crude oil, an estimated by 80-90% of cash flow. Reducing the cost of the crude feedstock, without changing the range and volumes of high valued distillates, increases the refining margin. Refinery profits are a direct outcome of the strategy applied by the refinery to purchase low cost crudes and to produce distillates with a high market value.

In the past, refineries were constructed to distill conventional light crude oils. Current economics, variations in the price of crude oils and shifting demand for distillates have forced refineries to reduce the cost of their distillation feedstock. Commonly, this is achieved by blending high-value light crude oils with heavy (unconventional) crude oils of inferior quality, or by buying ready-made blends.

Low quality crudes include heavy crudes from known locations, as well as opportunity crudes that are brought on the market by traders worldwide. These crudes, of lower quality, can be purchased at low cost. Blending of these with costly crudes is inevitable to produce crude blends that bear optimal properties to be processed, and at minimum cost.

More recently and especially in the US and in Europe, demand for fuels has shifted from gasoline towards diesel fuels. This means that while in the past predominately light crudes were distilled, today refineries must be able to distil heavier crude oils to increase the amount of middle and heavier distillates.

Refining margin for many refineries which were not able to adapt to the changing situations decreased. Technological limitations caused many refineries to buy expensive light crudes that do not produce specifically those distillates that are most needed in the market. For  many refineries, the losses were too large. Many closed or changed their activities from distilling toward blending.

Nowadays, crude blending is performed either by blenders or by refineries themselves which buy various types of low cost crude oils. They upgrade their chemical and physical properties to produce a synthetic crude oil at lowest cost, which can be processed in refinery equipment and will yield high value distillate.

Quality properties determine the market value of each type of crude. The most important quality characteristics are the density, the total acid number (TAN) and the sulphur content. For crude distillation unit optimization there is also very helpful to measure distillation curve (true bulling point) and SARA (Saturate, Aromatic, Resin and Asphaltene).

To increase refining margin and remain competitive, refineries are obliged to minimize the cost of their crude feed, without affecting their capacity to produce high value distillates. As heavier crudes are more difficult to process, and with the increase of consumption of diesel oil as compared to gasoline, light sweet crude oils are marketed at a higher price than heavy crudes. Reducing the cost of the crude input, without changing the range and volumes of high valued distillates increases the refining margin.

The strategy of crude oil blending includes several parameters. Each of them contributes to the overall final cost of crude oil entering the crude distillation unit as well as the refining margin:

  • The engineering limitations of crude distillation units to refine any type of crude oil
  • Cost differences of crude oils according to their location of origin, and their chemical and physical properties. An increased ability to process unconventional crudes leads to improved refinery margin
  • Product shifting in the market from gasoline towards diesel fuels. Increased demand for diesel fuels in the European market caused refineries to increase diesel yield over naphtha yield
  • High viscosity, especially in heavier crude oils, affects the flow properties of crude during transportation. Blending these types of crude oils with diluents or conventional crudes may be required to reduce viscosity and to improve flow properties.

In contrast to tank blending, in-line blending is performed by simultaneously transferring different crude oils through an on-line static mixing device to the final blend tank. The predetermined flow ratio between the different crudes will provide a blend of the required quality. In-line blending enables on-line correction of the quality of the blend, by changing the ratio between feeds. The blend is produced instantaneously and no stirred ‘blending tanks’ are required.

Blending different crudes, especially when unconventional crudes are involved, may cause precipitation of asphaltenes, which causes fouling in the pipes and process units. Asphaltenes are soluble in polar aromatics, such as toluene, but insoluble in paraffinic non-  polar solvents. On-line analyses of  the SARA content (saturates, resins, aromatics and asphaltenes) can be a potential tool for on-line determination of quantitative ratio between different crudes to be blended, or between crude oils and polar solvents, without  causing  asphaltenes to precipitate.

Efficient in-line blending includes a complex of different components. A major strategy in reducing the cost of the crude input, without changing the range and volumes of high valued distillates is to involve different technologies of hardware and software (HW/SW) solutions.

Hardware solutions are given by crude oil on-line analyzers, which provide real time measurements of physical properties of feedstocks. Software solutions to increase the refinery’s financial performance include optimization software, which includes the linkage of process streams between different refinery units, the logistics behind transportation of crude oils and final products.

The purpose of Modcon-AI HW/SW crude oil blending package is to provide refinery with set of modern optimization tools, which enables connectivity, validation and prediction of main KPIs, to take the correct decisions to maintain and improve effective industrial processes management. The implemented in this solution simulation “digital twin” is based on deep reinforcement learning (DRL) dynamic modeling, which enables to predict physical properties and chemical compositions for different process streams, and proposes required set points, that will accomplish the calculated predictions.

Being a powerful tool, which requires no historical data for reinforcement learning, DRL has one significant weakness, which makes it more difficult to implement for hydrocarbons processing with wide ranges of operation. DRL is likely to improve performance only where the pre-trained parameters are already close to yielding the correct process steams quality. The observed gains may be due to effects unrelated to the training signal, but rather from changes in the shape of the distribution curve. Therefore, there is a need in real-time monitoring of process yields, rather than only their prediction using the pre-trained models. This can be reached using the on-line analyzers installed in the process to determine the chemical composition or physical properties of substances involved in hydrocarbons processing.

The newly redesigned MOD-4100 crude analyzer represents a breakthrough in crude oil on-line analysis. It is a single analyzer, that performs on-line a variety of different critical crude oil measurements to provide real time analytical data, which is highly important for optimized operation processing of crude oils. The new design of the MOD 4100 analyzer system is based on a “Modular Package concept”. It is inspected and tested by the factory, and ready for immediate installation on-site.

The following crude oil critical parameters can be measured on-line and correlated to ASTM:

  • Salt Concentration (D3230)
  • Distillation (D2892, D86)
  • SARA (IP-143 and D893-69)
  • Emulsion stability (F3045, D4007 and D3707)
  • Hydrogen Sulfide content (D5705)
  • RVP (D6377 and D323)
  • Viscosity (D445 and D2501)
  • Water Content (D4928 and D4006)
  • Sulfur content (D2622 and D4294)
  • Density (D4928 and D1250)

Different blending options exist to upgrade unconventional crude oil into synthetic crudes of higher values. An automatic crude blending station integrates deep reinforcement learning models with MOD-4100 on-line crude analytics. It can be used either by traders who offer blending services, or directly by refiners. Cost, market value, availability and choice of technology are the main factors to be considered in planning a configuration to be used for upgrading unconventional crude oil.

Highest blending optimization can be achieved only by updating the simulation “digital twin” with real time analytical data for crude oil and blend quality. The HW/SW crude oil blending technology provides real time data and information about the physical and chemical properties of the blend in process. On-line adjustments and changes between blend components can be performed accordingly until the required physical properties are achieved.