DE10250325A1 - Process for controlling the operation of plants in a refinery or petrochemical complex - Google Patents

Abstract

The invention relates to a method and a tool for controlling the operation of the systems provided in a refinery or a petrochemical industrial complex. This is divided into a planning level, a scheduling level and a process control level. The data generated in the scheduling level, which are made available to the process control level as reference variables, are based on a decomposition of an equipment system in at least two individual systems and a numerical optimization of variables carried out in each individual system. The optimization is also coordinated by using factors determined in one of the individual systems for optimization in the other individual system.

Description

The invention relates to a method to control the operation of a refinery or a petrochemical Industrial complex provided facilities.

Refineries are Multi-product plants, which are made up of different starting materials of products. The continuous operation of the Process plants are generally discontinuous in so-called campaigns delivery or delivery the starting materials and products.

Task of production planning and - Control in refineries is timely purchasing and selection suitable educts as well as the determination of the operating modes of the process plants with the aim of maximizing the profit achieved in the factory. Farther this also requires timely delivery to customers be guaranteed with specification-compliant products.

Due to the complexity of the resulting Optimization problems and emerging uncertainties in the The data required for production planning exist Refineries generally have multiple hierarchical levels for production planning and control.

One belongs to these levels Planning level. The main goal of this planning level is Profitable creation of rough specifications for the starting materials to be purchased and product quantities to be produced. This is essentially what is needed from the outside information about Educt and product prices as well as possible sales of certain products estimated by marketing. The duration of the planning periods considered typically varies from about two weeks to a year. Often there will be another finer subdivision in the sense of short-term planning (two to four Weeks) and long-term planning (one to six months).

As part of the planning level generally considered the entire refinery. The whole complexity of this The work can be done with a numerical solution at the planning level occurring optimization problems are not taken into account. In the Illustration of equipment and the procedural system, major simplifications are made. This applies in particular to the modeling of non-linear systems and Blending behavior and the modeling of tanks.

Exact times and quantities for certain Purchases and sales are common at the time of planning not known. For this reason, the relevant information estimated and agglomerates over the planning period or over larger sections of this planning period. The agglomerated view This data is a characteristic indicator of the planning level.

To formulate and solve the There are methods of optimization problems that exist at the planning level linear programming (LP) usage.

The generated at the planning level rough guidelines for purchasing and production are too imprecise and insufficient, to ensure a smooth operational production process. Therefore a scheduling level is also planned. This is used for Generation of for a smooth operational process with sufficiently precise timing courses all the key performance indicators required at the process control level. In particular the assignment of orders to equipment takes place at the same time The presence of several suitable pieces of equipment. Also at the duling level an attempt is made to generate an optimal schedule. Indeed often occurs here also a consideration other criteria such as the robustness of a generated one Schedules.

The duration of the schedule considered periods is generally between three and fourteen days. Here too there is a division into short-term scheduling and long-term scheduling, however, an exact assignment to the duration of the planning periods considered is inaccurate.

If the decisions about delivery times and delivery quantities of starting materials such as crude oils not within the refinery to be made in consultation with the supplier becomes scheduling an exact quantity and schedule structure for educts is required. Otherwise be for the scheduling the quantities of starting materials determined at the planning level needed. Exact delivery times and delivery quantities are determined while of the scheduling process. Furthermore, an exact quantity and scheduling framework of the products to be delivered to customers.

The accuracy of that at the scheduling level used models is greatly improved compared to the planning level, so that a sufficiently precise specification of the reference variables for the process control level guaranteed can be. For example, nonlinear aspects of the considered subsystems and blending. Still acts these are usually not rigorous models. Mostly find it simplified models such as using empirical models.

In contrast to the higher-level planning level, the resolution of time is also aimed at, which enables the generation of time profiles of the control parameters for the process control level that are sufficiently precise for a smooth operational process. In contrast to the higher-level planning level, quantity data is agglomerated only to a very limited extent. The required, sensible resolution generally depends on the dynamic behavior of the systems under consideration. to Most of the optimization problems that arise at the scheduling level only exist on simulation-based software tools that support the personnel concerned in generating schedules based on manual specifications.

The process control level is used for Generation of suitable tax parameters for the tax and control systems of the plants in the refinery. objective this is where possible exact tracking the system operating point based on the one specified by the scheduling level Leaders at simultaneous warranty safe plant operation. Other considerations such as An energy-saving system operation can remain within the scope Degrees of freedom taken into account are, but are only of minor importance. The one on this Level considered periods are in the range of seconds to minutes. In addition to the through the Scheduling level predefined temporal courses of management variables are at this level different metrics the ongoing system operation and used. At the process control level, simplified, robust plant models that generally control and regulation the system in real time. As part of the control and regulation of process engineering systems exist a big one Number of methods and established commercial systems, for example for advanced process control.

Because of the complexity and complexity of the scheduling problems encountered in a refinery are typical several people entrusted with the scheduling. Often they are Functional areas. For example, is known the following breakdown:

• - scheduling the ships unloading or loading at a pier,
• - scheduling the use of different types of crude oil and the relevant ones crude oil tanks as well as the processing of the atmospheric residual and the associated tanks,
• - scheduling of the liquid gas production line and the relevant product tanks and scheduling of the gasoline production line and the relevant component and product tanks and
• - scheduling of the kerosene production line and the relevant components and product tanks.

Depending on the structure and organizational structure Within a refinery, another division of the Tasks are carried out The individual areas are scheduled generally not automated, but based on daily experience won heuristics. In many cases the scheduler runs through supports a software tool based on spreadsheets.

With this, for example, tank levels for certain Depending on the times of a certain planning period of operating parameters specified by the scheduler for the considered Attachments determined. The plant and mixture models used in this tool are great Part created and can with big Inaccuracies. With the help of this tool the Scheduler now a feasible one Schedule to create, for example, no fuel level violations having. This activity can be just in situations where a large part already the existing planning freedom is exploited to be very time consuming.

In most cases, this type of is missing Creation of schedules any criterion to the quality of a created feasible To be able to compare schedules with someone else's. Quality criteria like that when performing the Profit schedules or the flexibility of the schedules are indeed intuitively recorded, but can only be quantified in the rarest of cases and involve in decision making. Continue to arise independently problems with the type of division of tasks coordination of the different areas of scheduling, because they influence each other. Will be given in the above Example of the schedule for the occupancy of the pier changed, because, for example, a tanker loaded with crude oil is late, this changes also the requirements for the scheduling of crude oil and thus also the schedules for gas, petrol and kerosene production. Often the changed Information too late communicated to the responsible persons. A central attitude the for The data required for scheduling is generally not provided. Due to insufficient or different evaluation criteria of the respective scheduler for the goodness the schedules you create is also a decision about one Improvement of goodness of a schedule at the expense of goodness another schedule is usually not possible.

As already shown above results ultimately for a schedule to apply for the entire refinery several iteration steps developed almost independently of each other Plans of the single scheduler. The resulting overall schedule is mostly as part of a daily or a layered discussion with those responsible for the system leaders discussed. Problems that may become visible here can be a problem another change of the schedule before it is finally implemented.

A disadvantage for the current situation is accordingly that decisions are made in the local area of responsibility the respective decision-makers upstream the situation or insufficient knowledge of downstream units. Currently use decision makers generally my Excel sheets for their analysis. The necessary global decision is made in one Meeting.

Known first simulation-based implementations, for example the SIPP-II solution from Chemtech or the Schedule ++ from OR Soft work with simple models.

Based on this state of the art is the object of the invention, the data determination in the scheduling level improve so that the control and regulating units of the plants the refinery improved management data to disposal stand.

This task is accomplished through a process solved with the features specified in claim 1. Advantageous configurations and further developments of the invention are specified in the dependent claims. The Expectations 13 to 21 relate to a tool for controlling the mode of operation that in a refinery or a petrochemical industrial complex provided facilities.

The advantages of the invention exist especially in that by decomposing an overall problem stable, comprehensible partial solutions allows and also the possibility including more complex, high quality models is created. For example, when modeling the Product properties along the value chain octane numbers and knock resistance is taken into account become. There is an optimization of sizes and not just a simulation.

Those according to the present invention holistic view leads towards a global Optimums for an entire facility. The results obtained are reproducible and documentable, the quality the results are improved.

Other advantageous properties the invention will emerge from its exemplary explanation based on of the figures. Show it

1 Diagrams to explain the principle of horizontal and vertical decomposition within refineries,

2 Diagrams for explaining a decomposition concept for numerical optimization,

3 1 shows a diagram which illustrates a generalization of the horizontal decomposition and a coordination of two individual systems,

4 1 shows a diagram which illustrates a decomposition into four individual systems and a coordination of these four individual systems,

5 1 shows a flowchart to illustrate the optimization process taking place at the scheduling level,

6 a diagram of an application example and the

7 - 10 Diagrams to illustrate determined and estimated target function values.

The present invention relates to a method of controlling the operation of a refinery or a petrochemical industrial complex. According to this At a planning level, procedures become rough specifications for the required raw materials and product quantities to be produced. The one in the planning level Rough specifications are determined in a scheduling level Implemented data that correspond to the timelines of performance indicators. The in the Data determined at the scheduling level are used as reference values in the systems of the Refinery provided control units. This generate taking into account of the reference variables mentioned, control variables for the respective Investment. This determination of the control variables goes beyond the Leaders too Measurands, which are obtained from the ongoing plant operation.

At planning level there is a refinery planning for periods from one to six months. At this level you will find essentially Linear programming (LP) methods use. extended strictly linear programming through recursive methods, which allow, for example, the representation of the mixture of streams. Non-linear behavior of systems is determined piece by piece within the so-called LP models linear modeling only shown to a limited extent. Within the Planning levels are essentially the types and quantities of those to be purchased crude oils and determines the quantities of end products to be produced from it. The goal is to maximize the profit for the entire refinery in the considered Period. Decision variables at this level of decomposition are for the considered planning period, the quantities to be purchased differ crude oils, the distribution of the material flows within the refinery, plant operations and distribution of the blending components on the end products. Supplementary conditions for the optimization problems in the planning level there are essentially upper and lower limits for the Amounts of purchases and sales, upper and lower limits for Throughputs through Sub-systems and product specifications. The one at the planning level generated specifications, such as the amount of purchased crude oils, are taken into account when scheduling the fuel line, among other things.

The subject of the present invention is, in particular, a systematic approach for optimizing the reference variables determined at the scheduling level. The starting point for the need for such optimization is generally short-term events or circumstances that make it necessary to change the way the systems are operated. An example of such an event is the short-term availability of particularly cheap crude oil on the spot market. In this case - as far as the storage and ver the refinery's work capacities allow this - the production volume can be increased in the short term in order to benefit from the currently cheap crude oil. Another example of such an event is the short-term failure of an aggregate, which is associated with a temporary reduction in the tank capacity. In this case, the production may have to be sensibly reduced in order to cope with the short-term event. As part of this adaptation of the operation of the systems to short-term events or conditions, technological, logistical and economic aspects must be taken into account.

A basic idea of the invention is in scheduling in the sense of a holistic view considering one if possible huge Balance room, for example the entire refinery. The invention allows one Inclusion of high quality process models, in particular a modeling of product properties along the value chain. The modeling can be non-linear. There is a mathematical Optimization of variables.

This is essentially because of this accomplished that first a decomposition of an equipment system takes place in at least two individual systems, each of these systems at least one for a numerical optimization has a suitable variable. Then will coordination was carried out in such a way that the variables were first optimized one of the individual systems takes place and then during this optimization determined factors when optimizing the variables in the other Single system can be used, and vice versa. Is one Coordination takes place in the horizontal direction, then in the sense an integration concept or a concept for vertical Coordination made a connection to other decision levels.

In the context of horizontal or spatial decomposition, a superordinate optimization problem is divided according to functional criteria. For example, the entire refinery is divided into plant strings, which in turn are divided into individual plants. This is in the 1 illustrated, which shows a diagram on the left to explain the principle of horizontal decomposition within a refinery. Doing so with the reference number 1 the entire refinery, with the reference number 2 the individual production lines of the refinery and with the reference number 3 the process units or plants of the refinery. The arrow 4 symbolizes the increasing accuracy of the models used as the complexity of the optimization problems decreases.

In the context of vertical or temporal decomposition, a higher-level optimization problem is divided according to temporal aspects. In this way, a planning period can be divided into several shorter periods, which are individually optimized. This is on the right side of the 1 shown. Doing so with the reference number 6 the planning level, with the reference number 7 the scheduling level and with the reference number 8th illustrates the process control level, each plotted over the time axis. The arrow 5 symbolizes the increasing complexity of the optimization problems, the longer the period under consideration.

The 2 shows diagrams of a developed decomposition concept, which can be used for the numerical optimization of the production planning and control of the gasoline line of a refinery. The corresponding levels of decomposition are created by a combined application of horizontal and vertical decomposition. In the decomposition structure for numerical optimization, the one in the decomposition structure according to 1 existing scheduling level divided into two levels, namely a long-term scheduling level 7a and a short-term scheduling level 7b ,

Otherwise the structures are correct 1 and 2 match.

The following is a general one Form of the horizontal decomposition concept according to the invention presented.

Prerequisite for coordination after developed concept is the generation of Lagrangian factors or comparable information when solving the respective optimization problem. Optimization algorithms like random Search methods or algorithms based on heuristics are for the presented Unsuitable coordination strategy. An example of the Lagrangian factors comparable information is that in the context of the simplex process generated shadow prices.

Another prerequisite for coordination is the existence of market prices for those crossing the system boundaries Streams. Moreover have to Valid at the beginning of the optimization Initial values for the optimization variables are present. These valid starting values can based on existing empirical values, as part of a simulation be won or as part of an optimization process be determined.

The plants or equipment systems of the optimization problems considered in the process engineering, continuous production can generally be divided horizontally into two smaller systems. The resulting systems can also be divided up to a reasonable extent from a process engineering point of view into two further subsystems. In this way, an entire equipment system can be broken down into individual systems with sizes suitable for numerical optimization. The division of such a system into two individual systems with an arbitrary number of currents which cross the system boundaries and flow between the subsystems is shown in FIG 3 shown.

Doing so with the reference number 9 the overall system and with the reference numbers 10 and 11 denotes the two subsystems into which the overall system is divided. The in the 3 The arrows shown symbolize currents that flow into the system, leave it or flow between the subsystems. Each of these flows can be described by a flow vector, the components of which describe the mass or volume flow and the properties of the corresponding flow. Typical properties are, for example, the research octane number RON, the motor octane number MOZ and the density.

The procedure is as follows:
First, there is an optimization problem for the second system for a valid operating point of the overall system 11 solved where the current vectors for those in the system 10 returned currents remain fixed at the values specified by the starting point of the optimization. With this first solution of the optimization problem, Lagrangian factors are generated, from which weighting factors for those of the system 10 in the system 11 flowing currents can be determined at the starting operating point. The weighting factors determined in this way are in the objective function f _{1 of} the optimization problem for the system that is now to be solved 10 used. The objective function to be maximized is as follows:

Here ω denotes a vector of the weighting factors, S a current vector, P a process variable vector, B the operating costs and E the revenue. The indices i denote all streams by the system 10 in the system 11 flow. Additional conditions of the optimization problem to be solved here are the quantities and properties of those in the system 10 incoming flows, upper and lower quantity restrictions for intermediate products, specifications and prices of intermediate products, plant models for the system 10 and restrictions on step sizes.

The current vectors resulting from the solution to this optimization problem are now specified in the constraints for the system's optimization problem to be solved 11 , To do this, the previous solution to the optimization problem for the system 10 the weighting factors of the system 11 in the system 10 flowing currents determined and in the objective function f ₂ for the optimization problem of the system 11 used. The objective function to be maximized is as follows:

The indices i denote all currents from the system 11 in the system 10 flow. The constraints on this optimization problem are the quantities and properties of those in the system 11 incoming flows, upper and lower quantity restrictions for intermediate products, specifications and prices of intermediate products, plant models for the system 11 and restrictions on step sizes.

In the following it is checked whether the termination criterion for the optimization is fulfilled. In order to ensure a convergence of the currents flowing between the systems, it is checked whether the change of all current vector components of these currents between two runs of the iteration fall below certain values. If none of the corresponding current vector components changes by a significant value between two runs, the optimization is terminated and the results are output. If the currents have not yet converged, they are used for the system 11 determined new optimal current vectors in the constraints for the system 10 is used and a new optimization is carried out.

The current vector can contain information about the mass flow, about compositions, about temperature and pressure, as well as about physical properties such as density and octane number th.

The procedure described above is described below using the 5 shown flowchart explained in more detail.

With the reference number 19 is the start of the optimization process of the overall system 9 designated. The optimization problem for the system is solved in accordance with step 20 11 with fixed current vectors. A determination of new weighting factors for the system then takes place in step 21 10 , Taking these new weighting factors into account, the optimization problem for the system is solved in step 22 10 , Thereafter, according to step 23, new current vectors for the system 11 specified. In the following step 24 new weighting factors for the system 11 determined. These new weighting factors are used in step 25 to solve the optimization problem for the system 11 used. In step 26, a query is made as to whether the abort criterion is met. If the termination criterion is not met, then new current vectors for the system are generated in step 29 10 predefined and there is a return to step 21. If, on the other hand, the termination criterion is met, the results are output in step 27. Step 28 denotes the end of the optimization of the overall system 9 ,

The 4 shows a diagram showing a decomposition of an overall system 12 in four individual systems 13 . 14 . 15 . 16 and illustrates a coordination of these four individual systems. This system is initially done in the same way as in connection with the 3 and 5 an optimization regarding the systems has been described 14 and 16 to a new system 17 to build. The two systems are then optimized 15 and 17 forming a new system 18 , Finally, the systems are optimized 13 and 18 , In the course of this coordination of the individual systems described above, it is necessary to re-optimize the previously formed new systems in each iteration step.

As is clear from the example shown the effort for coordination with the number of systems to be coordinated. It is therefore already to be considered in the context of a horizontal decomposition whether a division into a large one Number of small systems the resulting coordination effort justifies.

Using the procedure outlined above is in principle also the integration of the optimization of the gasoline line in an optimization of the entire refinery possible. For example, this could the LP model, which mostly exists in refineries, for that not for the gasoline train belonging equipment system be used. The required interface information to determine the change in objective function of the individual systems to be optimized depending on the change of constraints can from the shadow prices of the LP model or from the Lagrangian factors of the optimization problem for the gasoline train be won.

A way to coordinate of the various vertical planning levels is the default results from the optimization of the higher level as secondary conditions for the Optimization problem of the respective lower decomposition level.

As in the context of the developed Decomposition strategy possible, can, for example, on the parent Long-term scheduling a number of optimal blending recipes be determined. These recipes are within the constraints the lower level of scheduling used to be on this Level by determining an optimal recipe order optimize certain objective function.

Also the objective function of the subordinate Level can be used to track those on the upper level Intention to be considered at the lower decomposition level. So can on the lower planning level deviations from on the higher one Planning level created specifications are minimized. In the featured Can decomposition concept for example at the lower planning level of scheduling the deviations due to an optimal chronological arrangement of recipes to gas stations be minimized at the end of the period of the long-term scheduling optimization problem.

The following is a closer look at one Suitable concept for coordinating the decomposition Sub-problems in the context of production planning and production control of the petrol line.

For the system reformer blending system of a refinery is initially for one Period a coordination concept for the optimization problems shown, which is from a decomposition of the optimization problem for the Long-time scheduling level arise.

The overall system considered here 30 which in the 6 shown consists of the subsystems reformer 32 and blender 24 , In the balance area of this overall system 30 go a number of educt streams such as the reformer feed stream 31 and components 36 for gasoline blending. Furthermore, a number of streams leave the balance area. This is the case with the blending system 24 to products manufactured according to specific specifications 35 and in the case of the reformer plant 32 about intermediates 37 such as hydrogen. Within the system under consideration, it becomes the reformer 32 leaving reformate stream 33 as a component for blending 24 used.

Each individual stream within this overall system is described by a so-called stream vector S _i , the components S _{ij of} which contain the mass or volume flow and properties of the corresponding stream. The index i denotes a specific current, the index j denotes a specific component of the current vector in question. Typical properties are, for example, the research octane number RON, the motor octane number MOZ or the density ρ. The following applies: S i = f (P t )

Here, the vector P _t describes the plant operating mode set in the subsystem t. This vector contains all process variables to be optimized in the subsystem, such as pressures, temperatures or, in the case of blending, volume flows of individual components that are blinded for each end product.

As part of an optimization at the Long-term scheduling levels are optimal for the entire system Driving style of the reformer and optimal distribution of the accruing Determine blending components on the different end products. The profit achieved in the entire system is additive from the profits made in both systems.

For the coordination of the two systems now raises the question like a change of the optimization variables of a subsystem affect the target function of the other subsystem. In the presence of such quantitative Information an iterative approach can be developed that of a valid Starting point of the overall system, the operating points of the individual systems so changed that ultimately an improved operating condition for the overall system is achieved.

As will be shown below, the analysis of the Lagrangian factors at the optimum of a subsystem 2 Information about the partial change in the objective function f of this subsystem as a function of the changes in the current vector components S _{i, j of} the system 1 in system 2 win flowing current S _i . The vector of these weighting factors is referred to as i below:

Is the vector i known for a current S _i by the system 1 in the system 2 flows, can be in a certain range around the respective optimum of the system 2 the change in the objective function of this system depending on changes in the system 1 generated electricity and thus changed process variables of the system 1 predict. The calculation of the changed current vectors as a function of the changed process variables takes place here via the system model.

The optimization problems that arise at the long-term scheduling level can be described as: min f (x) taking into account the secondary conditions g in the form: gg (xx) = 0

Typical constraints of the multiblend optimization problem are, for example, volume balances and mixing rules to meet product specifications. Such a mixing rule reflects the functional relationship between the optimization variables of the multi-blending problem x, that is to say the quantities of component streams blended into a specific product, the properties of the component streams contained in the current vectors S i and a property of the product produced. These secondary conditions are generally formulated as follows, where the constant e _k corresponds to a value specified by specifications: g _k (x, S ₁ , S ₂ , ... S _n ) - e _k = 0

For optimization problems with constraints, the Lagrangian function L is defined with the vector λ of the Lagrangian factors λ _k belonging to the constraints as follows: L (x) = f (xx) + ∙ g (x)

The corresponding Lagrangian factors λ _{k are} generally determined using numerical optimization methods.

The following is an example on the calculation of the change ω of the optimal Value of a target function f depending on the change a current vector component S gone in only one only constraint g of the optimization problem occurs.

Let f * be the value of the target function at optimum xx *, which in turn depends on the current vector component S considered: f * = f (x * (S))

The following also applies to the secondary condition g * considered: g * = g (x * (S)) = 0 as well as the Kuhn-Tucker condition: ∇f * + λ * (S) ∙ ∇g * = 0

According to the general chain rule:

With the chain rule for the change of the target function depending on the current vector component S:

It follows:

For the change of the target function depending on the property S follows:

Since a current vector component cannot only appear in a constraint of an optimization problem, the above approach is extended to all n constraints of the optimization problem under consideration. For this follows:

Using this equation, we can use the Lagrangian factors of constraints of an optimization problem and over the partial derivatives of these constraints according to the current vector components S the weighting factor ω with respect to one current vector component occurring in these constraints become.

The immediate validity the determined weighting factors for predicting the change in objective function with change the associated Constraint is only for infinitesimally small changes given the respective secondary conditions. This is justified as follows:

• - The appraisal of change a nonlinear objective function by a constant term corresponds to one Linearization of the target function and can therefore only in immediate Near the certain optimums his.
• - One Improvement of the optimum in a certain direction is under circumstances limited by other constraints. In this case too a new determination of the weighting factor is required.

The following is based on a typical Multiblend optimization problems checked to what extent the weighting factors obtained for predicting the change in objective function dependent on of change of constraints can be used.

The optimization problem considered here was created based on a real quantity and schedule structure for products and starting materials. The blending models used were also based on real operating data calibrated.

The sample optimization problem was first solved in the course of the following investigation. The optimum f * _{old found in this way} is in the 7 - 10 identified by the vertical axis.

At the optimum found, the weighting factors ω _{i, j} with respect to selected flow properties and the volume flow of the reformate were determined using the method described above.

Using the weighting factors obtained in this way, new optimal values of the objective function f * _{new, estimated} as a function of the change in a component of a current vector ΔS _{i, j, were} estimated:

These values obtained in this way were compared with the values f * _new , which resulted from a real change in the equation constraint by ΔS _{i, j} and then a new solution to the optimization problem with the modified constraint while at the same time keeping all other constraints constant. If the optimization problem could no longer be solved by changing a constraint, there was no corresponding entry for f * _new in the figures.

As components to be examined of a flow vector were the vapor pressure, RON, MON and the volume flow of the reformate. Nonlinear blending models were found to determine the considered current properties of the manufactured products use.

To compare the individual Optimization results despite different orders of magnitude the considered current vector components and the constraints belonging To ensure Lagrangian factors became the considered range of current vectors for the investigation limited to values, where the predicted change the target function not more than ten percent of the initial value scam.

In the 7 the estimated and determined values of the target function are displayed by the optimum f * _old when the vapor pressure changes. The estimated and determined values for the target function agree over a wide range of vapor pressure.

In the 8th the change in the objective function depending on a change in the research octane number is shown. Here, the predicted changes in a narrow range around the initial value agree well with the true changes, but the change is underestimated in the area of higher octane numbers. If the octane number of the reformate is greatly reduced to approximately 94 RON, no solution to the optimization problem can be achieved.

Also for those in 9 The dependency of the optimal target function value on the number of engine octets shows that for the majority of the values there is a good agreement between the predicted and the true value. However, a problem of nonlinear optimization becomes clear in this series of measurements: One of the values f * _new is significantly above the corresponding value of f * _{new, estimated} . The points f * _new to the right of this pair of values should have at least the same target function value, since a further increase in the number of engine octets, for which no upper limit is specified by specifications, should not result in a deterioration in the optimal target function value. Obviously, only a local optimum was achieved at these points due to the specification of unsuitable starting values.

Similar to that in the 9 are also in the 10 again, the difficulties with nonlinear programming are clear. While a good prediction accuracy is achieved in the area of increased volume flows, the optimized values in the area of a volume flow reduced compared to the base case are exclusively above the values obtained by estimation. Furthermore, a better target function value is achieved for these points than in the area of higher volume flows. Since the upper tank level limit for reformate was not limiting for the optimization problem under consideration, this behavior is again due to a local optimum that was found by the optimization algorithm on the basis of the specified start values.

In summary it can be said that in all considered cases the tendency to change the constraints on the target function were correctly predicted. In many cases is for the respective, albeit local optimum, a very good agreement reached. The developed method is basically suitable to change within a narrow range around the respective optima to predict the objective function. Because of this result the development of a step-by-step coordination strategy for those considered Optimization problems possible.

As has been shown, use the weighting factors to make changes to the optimal objective function value dependent on of changes of predicting current vectors. This information can be use to specifically control the driving style of the blending Modify systems so that through the so changed Properties and quantities of blending component flows in the overall system a higher one Profit is made.

In the first step, a specification is made of values for the current vectors of the reformate. These current vectors should be one potential Driving style of the reformer, for example at the time of planning current driving style.

In the second step, the respective Multi-glare optimization problem with the given current vector for the reformate current solved. In addition to determining optimal values for the optimization variables and the value of the target function at the optimum takes place within this Optimization of the calculation of the Lagrangian factors for all constraints the optimization problem.

Using this on the individual Lagrange factors related to constraints now become the weighting factors in terms of of the individual components of the current vector of the reformate.

Now the separate reformer optimization problem is solved. Here, the vector with the operating parameters of the reformer is to be changed in such a way that the greatest possible increase in the profit of the overall system consisting of reformer and blender is achieved compared to the current state. The following issues are considered:

• - changes of the profit made from the blending. The estimate is made here about the determined weighting factors relating to a component of the Reformatstromvektors;
• - Modification the operating costs of the reformer caused by a changed mode of operation were generated;
• - changes the proceeds, by change of sellable intermediate products produced in the reformer.

By summarizing the current vector related Weighting factors the objective function of the reformer optimization problem.

max f Ref  = Ref [S Ref (P Ref, new ) - p Ref (P Ref, old )] - [B Ref (P Ref, new ) - B Ref (P Ref, old ) + [E Ref (P Ref, new ) - E Ref , (P Ref, old )]

Supplementary conditions of the optimization problem to be solved here are the quantities and properties of those entering the system currents upper and lower quantity restrictions for intermediate products, specifications and prices of intermediates, the reformer model and limits for increments.

The reformer model provides here the relationship between the operating parameter vectors and the Current vector of the reformate. Furthermore, through the model also the electricity vectors of the intermediate products and the operating costs determined.

They have a special meaning Supplementary conditions for limiting increments between old ones and new set of operating parameters of individual iteration steps.

As already mentioned, the validity predicting changes in objective function limited to a narrow range around the respective optimum. Around The changes in the new operating status should not leave this area across from the old state to be limited to convergence problems in the iterative solution to avoid the two optimization problems. Possible strategies for defining the increments are:

• - a restriction of the absolute change of components of the process variable vectors from new to old state: | P new, I - P old, I | ≤ ε l Here, the value ε ₁ specifies an upper limit for the change in an operating parameter P _{new, 1 of} the reformer compared to the previous iteration step.
• - a restriction of the absolute change of current vector components from new to old state: | S new, i, j - p old, i, j | ≤ ε i, j
• The value e _{i, j} specifies an upper limit for the change in a component of a current vector compared to the previous iteration step
• - A restriction of the predicted change in the target function by changing individual current vector components: ω _{i, j} ∙ [S _{i, j} (P _new ) - S _{i, j} (P _old ) ≤ ε

The parameter e is one User-defined barrier to the upper limit of the change of the current vector component generated objective function change.

The first two approaches it is assumed that there is sufficient prediction accuracy the weighting factors in a certain, absolute range of Current vector or process variable vector components around the optimum determined. The third case is against assumed that the accuracy of the estimate using of weighting factors from the change depends on the objective function.

Furthermore, in contrast to the first two approaches to the last approach shown the weighting factors belonging to the current vector components are taken into account.

Weighting factors take part for current vector components very high values. These values are usually only in an extremely narrow range Range valid around the respective optimum. By limiting the Increment with constraints becomes the allowable increment for current vector components with high weighting factors compared to the step size for others Current vector components stronger limited.

The quality of each strategy as well as the width of the steps used in each case is still through optimization calculations to check. Perhaps A combination of individual strategies can also be effective prove.

In the following it is checked whether that Termination criterion for the optimization met has been. Possible Examples of termination criteria are:

• - exceed one a maximum number of iteration steps,
• - exceed one a maximum number of iteration steps with none or only a certain marginal improvement or
• - on Falling below a minimal change in current vectors between two iteration steps.

If the termination criterion is not met, be the from the solution the improved current vectors resulting from the reformer optimization problem as new current vectors for predefined the multi-glare optimization problem.

Then the multi-glare optimization problem is solved again and then the refor problem of optimization up to the fulfillment of the optimization criterion.

Claims (21)

Process for controlling the operation of the plants provided in a refinery or a petrochemical industrial complex, whereby - rough specifications for required starting materials and product quantities to be produced are determined in a planning level, - the rough specifications determined in the planning level are converted into data in a scheduling level , which correspond to the temporal courses of reference variables, - the data determined in the scheduling level are supplied as reference variables to the control and regulation units provided in the plants and - the control and regulation units of the installations generate control variables for the respective installation taking into account the reference variables , characterized in that - to generate the data generated in the scheduling level, a decomposition of an equipment system into at least two individual systems, each with at least one variable suitable for numerical optimization is carried out, - the variables of one of the individual systems are optimized and - factors determined in this optimization are used for the purpose of coordination when optimizing the variables in the other individual system. A method according to claim 1, characterized in that regarding each individual system is given a target function and the target function as part of the optimization process by each other Determined individual system factors is changed. A method according to claim 2, characterized in that the objective function is dependent of quantities and properties of flowing material flows is changed. Method according to one of the preceding claims, characterized characterized that using a variable optimization an iteration process takes place. A method according to claim 4, characterized in that a check is carried out as part of the iteration process to determine whether the determined variable met an abort criterion. Method according to one of the preceding claims, characterized characterized that start values at the beginning of an optimization process for the variables to be optimized are specified. Method according to one of the preceding claims, characterized characterized that the decomposition is a horizontal decomposition and the optimization of the variables of the individual systems a horizontal Coordination is. Method according to one of the preceding claims, characterized characterized that the scheduling level is divided into two sub-levels one of which is a long-term scheduling level and the other is a short-term scheduling level. Method according to one of the preceding claims, characterized characterized that there is vertical coordination, at which a level is linked to at least one other level. A method according to claim 9, characterized in that vertical coordination by specifying results from the optimization of a higher level as ancillary conditions for the Optimization problem at the lower level or through use the objective function of a respective subordinate level, to which the the objective pursued at the upper level in each case at the lower level to take into account is made. Method according to one of the preceding claims, characterized characterized that high quality during the optimization process Process models are used that are used to model product properties correspond. A method according to claim 11, characterized in that a non-linear modeling is carried out men will. Tool to control the operation of the in a Refinery or a petrochemical industrial complex Attachments, being used to generate those generated at the scheduling level Data a decomposition of an equipment system in at least two individual systems, each with at least one for a numerical one Optimization of suitable variables is carried out which tool Means for optimizing the variables of one of the individual systems and on the use of factors determined in this optimization for the purpose of coordinating the optimization of the variables in the other individual system. Tool according to claim 13, characterized in that it has means related to each individual system for setting a target function and for changing the objective function in the course of the optimization process by the of factors determined for each other individual system are provided. Tool according to claim 14, characterized in that there are means of change depending on the objective function of quantities and properties of flowing material flows. Tool according to one of claims 13 - 15, characterized in that that it's using means to optimize a variable an iteration process. Tool according to one of claims 13 - 16, characterized in that that there are means to carry out horizontal decomposition and horizontal coordination having. Tool according to one of claims 13 - 17, characterized in that that there are means to carry out vertical coordination. Tool according to claim 18, characterized in that there are means of vertical coordination by default Results from the optimization of a higher level as a secondary condition for the Optimization problem at the lower level or through use the objective function of a respective subordinate level, to which the The objective pursued at the upper level in each case at the lower level to take into account having. Tool according to one of claims 13 - 19, characterized in that that it has data indicating a modeling of product properties correspond. Tool according to claim 20, characterized in that the models are non-linear.