Process units and process models

The plant model in AVEVA Unified Supply Chain is composed of many linked process units. Process units represent the physical assets in the plant which are used to transform materials. Process units are generally one of two types, either a distillation unit, which separates material based on its boiling point, or a transformation unit which transforms material using a chemical process.

In AVEVA Unified Supply Chain distillation units are modeled using the cut points and fractionation efficiency of the unit. These parameters can be calculated by monitoring the unit performance and deriving the values from test run data. The parameters can be calculated in the Assay component of AVEVA Unified Supply Chain.

Transformation units are usually modeled using Base + Δ models. Base + Δ models represent how the unit transforms the input feeds into the output products. The models respond to changes in the input feed qualities to produce a proportional linear change in an output quality. In order to better represent the changes across a range of feed qualities, sometimes it is necessary to break the model into smaller pieces that are used for particular ranges of feed quality. These are known as bases.

The performance of a process unit is described by a process model. The process model normally has input feeds and output products. The feed properties are used by the model to calculate the qualities of the output products. The model may also require operational parameters, which represent physical settings of the equipment.

The material data is carried into and out of the model through model connections. Usually the number of input and output model connections equals the number of physical input and output connections of the process unit.

Pooling model outputs

For some process models it may be useful to track individual streams which do not appear as actual separate materials in the plant. This is most often used for calculating the yield of light hydrocarbons and tracking them through the plant. However, the process unit does not output these materials as separate streams, as they are usually carried in a single gas stream. That is, the process model outputs more connections than there are pipes in process unit. In this situation it is necessary to map several model connections to the same pipe. For example, a VGO Hydrocracker model may output the yield of methane, ethane, propane. These are carried out of the unit in a single saturated gas stream. Therefore within the process unit it is necessary to map the model connections for the light ends to the saturated gas flowsheet pipe.

Handling multiple process models in single process units

A process unit may handle several different types of material, each of which responds in a different way within the unit. It may not be possible to accurately represent the unit's performance with a single Base + Δ model, even one which includes many bases. For these units it is necessary to use a process unit which contains several process models, each of which acts on the different types of feed to the unit. Each model may have different inputs and outputs, which must be mapped to the available pipes on the physical unit.

Example: A naphtha hydrotreater may process light naphtha, heavy and cracked naphtha. These respond differently in the unit and so require different models. Each model outputs H2S, so this can be mapped to the unit H2S pipe. Each model also produces several light ends which can be pooled and mapped to the Sat Gas pipe. The light naphtha model produces light naphtha and is sent to the desulfurized light naphtha. The heavy naphtha model produces heavy naphtha which is sent to the desulfurized heavy naphtha pipe. The cracked naphtha model produces both light and heavy naphtha, so these separate streams are sent to the relevant pipes to be output with the products of the other models.

Where a process unit has operating parameters, the same value for this parameter is used across every process model, although a particular process model may or may not actually use this parameter to drive itself.

Depooling streams

Pipes on the flowsheet may carry several substreams of material. Each of these substreams is tracked separately, so its yields and qualities are preserved. Normally, when the pipe reaches its destination process unit, the substreams are pooled to make a single input feed.

However, it is possible to feed these different substreams into different models within a process unit. This process is called depooling. It allows you to recreate the real plant topology by representing real pipes on the flowsheet, while increasing model accuracy by using different process models for different feed components. Depooling the pipe components allows each specific substream in the pipe to be fed to a process model which best describes the reactions of that material within the process unit.

Fixed material associations

The output streams from process models can be associated with fixed materials. Fixed materials are a set of properties associated with a particular material. When a stream is associated with the fixed material, it gains all of its properties.

Fixed materials are useful when multiple streams in the supply chain model share a bulk set of property values. For example, many process units might produce hydrogen sulfide. Rather than having to enter the properties of hydrogen sulfide for every process unit, you can define this as a fixed material, and then associate this fixed material with the relevant streams.

Fixed materials are global objects and can be shared by different AVEVA Unified Supply Chain users. The properties of the fixed material are added to the properties predicted by a process model, but do not replace properties already predicted by the model.