DE10261124A1 - Planning complex production sequences in commercial operation of, e.g., steel plant employs stepwise process to achieve optimized groupings and product or production sequence - Google Patents

Abstract

The method is based on a mixed linear programming (MILP), mixed integer optimization. In a stepwise process, taking into consideration predetermined rules and product dimensions, optimized association and sequencing of manufactured products into families and groups is carried out. An optimized product sequence or production sequence is determined. In the first step (2) based on product parameters and rules, ordered sequential pre-sorting places the products to be made into individual families. In the second step (4) within a each product family, a rule-based grouping of the manufactured products is made in terms of at least one product parameter, into individual product groups. In a third step (6) for each product group, a compatibility matrix is determined and a mathematical model is constructed. In a fourth step (8) common MILP solution methods are applied to the compatibility matrix or mathematical model. This yields an optimized sequencing of products within a each product group, and of product groups within a product family. Values are arranged in rising or falling order. Automated and/or manual weighting of the various product parameters are carried out. Based on the weightings of the individual product parameters, a range list is constructed. Predetermined control and/or weightings of the various product parameters are co-determined or impressed through respective plant operation and/or the respective manufacturing process. Weightings of the various product parameters are taken into consideration in the predetermined rules. Sequential pre-sorting is carried out, based on weighted parameters. Both rising and falling product parameters are allowed or enabled within the product sequence of each product group. An Independent claim is included for the corresponding system

Description

The invention relates to a method for optimized planning of complex production sequences in large-scale Plant operations, especially in the steel industry and / or the metal processing industry using program technology implemented mixed-integer optimization process.] In the large industrial Manufacturing is due to the immense quantity or number of pieces to be manufactured Products and the incredible amount or mass of raw materials and complexity of process flows to avoid economic damage, for example due to unwanted Plant downtimes, particularly due to Changeover and / or cleaning work when changing products, special planning efforts to guarantee a smooth and uninterrupted manufacturing process or plant operation required.

The above applies in particular in the steel industry, where through appropriate selection of the sequence of each Slabs or steel blanks, which are essentially in terms of their Geometric dimensions and steel quality differ, one as possible continuous casting process to ensure the steel and / or iron blanks to be processed is. The same also applies to the further processing of the generated ones Blanks using rolling mills but also in the chemical / pharmaceutical industry and / or the pulp or paper industry application. Here too is one to preserve preferably efficient production process an optimized depending on the request or specification Product sequence or production sequence, for example sheet metal, observed. The specification data can include physical, chemical and / or geometric properties, such as Thickness, length, Width, density, structure, elasticity, hardness, chemical composition, regarding include.

A well-known approach to determination the product sequence in steel making is in Grossman's article and Harjunkoski “A decomposition approach for the scheduling of a steel plant production "(Computers and Chemical Engineering, 25, 2001, pages 1647-1660). In the aforementioned Article is geared towards increasing the efficiency of the production operation by extending or expanding the continuous casting sequences to achieve what is a reduction in steel production or production Downtimes or downtimes means. The increase in efficiency of a steelworks was made to order by presorting Steel blanks according to product parameters, such as for example delivery date, width, quality and thickness achieved by Blanks of the same quality and thickness have been combined into a product family. Within Corresponding casting sequences then became a product family formed, but disadvantageous within a casting sequence a constant or decreasing width of the steel blanks from previous to subsequent blank was a requirement.

The invention is based on the object improved planning of complex production sequences in large-scale Plant operations to be specified.

The above task is accomplished by a Method and a system of the type mentioned with the features the independent Expectations solved. Advantageous embodiments of the invention are in the following Description as well as in the dependent claims specified.

According to the procedure, planning becomes optimized complex production sequences in large-scale plant operations, in particular the steel and / or pulp and / or paper and / or Chemical / pharmaceutical industry, through program application one on methods and algorithms of mixed integer linear programming (MILP) constructive, mixed-integer optimization method and taking into account Predefined rules and product parameters are gradually optimized Grouping and optimized sequencing according to the order backlog Products carried out and as a result an optimized product sequence or production sequence certainly.

In a first step, dependent on of product parameters, such as for example delivery date, length, Width, thickness, height, Weight, density, goodness, Quality, Purity, strength, elasticity and chemical composition of the respective product, as well as by Predefinable rules, in particular through the manufacturing process co-determined, a sequential pre-sorting of those to be produced Products carried out in individual product families.

Consider the above rules here essentially geometric, chemical and / or physical Requirements and or restrictions of the respective manufacturing process and / or of the product manufacturing plant operation used.

Consider geometric rules decisive here restrictions and / or general conditions regarding the used for the production Equipment or the available standing infrastructure.

Chemical rules take into account, for example, different chemical compositions of the products to be manufactured and the raw materials, auxiliaries and operating materials required for their manufacture, whereby it must be ensured that a second product is a previous first product without major financial and / or technical expenditure, such as cleaning work of the equipment used to manufacture the first product. Accordingly, for example Steel production Always cast or manufacture steels with the required very low carbon content before steels with a higher carbon content, i.e. a sequence of products or steels with increasing carbon content must be taken into account. Furthermore, it must be ensured that successive product groups of the specification required by the order can be manufactured at all. For example, the manufacturing process of some high-purity steels, the so-called “wash grades”, requires a thorough preparation of the equipment to be used in the preparation in order to remove impurities and / or residues from previous melts or steels.

Take physical rules into account here physical quantities to be observed in particular the manufacturing process regarding, such as temperature, pressure and general Specifications and settings of the various to be observed used equipment.

The second step is inside of each product family based on at least one pre-selected Product characteristic, for example of quality and / or width, rule-based grouping of those to be produced Products carried out in individual product groups.

In a third step, a compatibility matrix is determined for each product group based on parameters and based on rules, and a mathematical model based on this is created. The compatibility matrix represents or specifies the complex and / or non-linear rules on which the mixed-integer optimization method is based. The matrix elements can have the values 0 or 1. The matrix element P _{ii '} has the value 1 if product i' can be produced after product i. If this is not the case, then the matrix element P _{ii has} the value 0. This matrix together with the target functions to be optimized, for example minimum downtimes of the plant operation or a continuous production process, embedded in the MILP approach, provide all the information required to carry out a Successful sequencing ready or meet all the requirements for successful sequencing.

According to the procedure, either Staying the same and / or increasing or increasing or staying the same and / or the decrease or decrease of product parameters within allows or enables the product sequence of each product group.

The mathematical model has here the aforementioned flexibility to suffice.

Finally, in a fourth step by using more common MILP-solving methods on the compatibility matrix or the mathematical model an optimized sequencing of the Products within each product group, the product groups within each product family and therefore also the products implemented and achieved within a product family.

The execution of the method according to the invention is preferably carried out using a system for optimized planning complex production sequences in large-scale plant operations, in particular the steel and / or pulp and / or paper and / or Chemical / pharmaceutical industry, which means for program engineering Application of one, to methods and algorithms of the mixed integer Linear programming (MILP) building, mixed-integer optimization method has, whereby means are available, which are set up under consideration Predefined rules and product parameters are gradually optimized Grouping and optimized sequencing of products to be manufactured in product families and product groups and an optimized product sequence or to determine the production sequence.

A computer program that has the characteristics of the method according to the invention has and is executed by suitable hardware, in particular a Data processing device with input and display device as well as possibly with network and / or internet connection and which possibly interacts with a data store leads to a preferred embodiment of the system according to the invention. A computer program, in particular one stored on a data carrier Computer program containing the features of the method according to the invention is therefore expressly included in the disclosure content of the present application.

These and other advantageous embodiments and embodiments of the invention are the subject of the dependent claims.

Using one in the attached figures illustrated embodiment the invention, the invention, advantageous embodiments and particular advantages of the invention are explained and described in more detail.

Show it:

1 Exemplary steel manufacturing process

2 exemplary process flow

3 Exemplary pre-sorting in sequences or product groups with the product parameters quality, width, thickness

4 according to the pre-sorting 3 based compatibility matrix

5 Optimized grouping and sequencing of the products to be manufactured, provided only that the widths within a sequence decrease

6 Grouping and sequencing of the products to be produced, optimized according to the invention. The further explanation of the invention is based on the example in FIG 1 shown steel manufacturing process in a steel mill.

In the example used here Steel production process will be two electrically powered arc furnaces 01 and 02 as blast furnaces used, which are operated alternately and in which melted New steel combined or scraped with scrap iron or scrap is melted. About that In addition, an argon oxygen is used for the metallurgical treatment of the melt Decarburization unit D1 to reduce the carbon content of the Melt, the so-called "fresh", and for monitoring and regulation of the chemical composition and temperature of the Melt a ladle with metallurgical devices G1, for example a so-called “converter” final Shed of molten steel takes place by means of modern continuous casting processes in a continuous casting plant S1, which is a constant solidification and an optimal structure of the manufactured Ensure slabs (steel blanks) and or round bars. A continuous one casting process usually includes between 8-10 furnace fillings. Once these have been processed, so become regular Maintenance work and / or maintenance measures on the continuous caster carried out. The maintenance cycle is around 8-10 completed furnace fillings or melting.

Acting in steel making it is an extremely time critical Process with a comparatively high energy requirement, in which planning decisions As a rule, the sequence of the steel manufacturing process still manual, that is be met by an expert, which is particularly the case with highly complex process flows and a variety of considerations Framework proves to be difficult to almost impossible.

Especially the last processing step, namely continuous casting the melt into slabs to order and / or round steels is to be considered and fulfilled due to the large number Rules and / or framework conditions to achieve the required steel quality as the most difficult process to implement in terms of planning.

The aforementioned rules include essentially geometric, chemical and / or physical rules and / or requirements.

Geometric rules stir here essential of restrictions and / or general conditions regarding the used for the production Equipment or the available existing infrastructure, especially in known Process assumed that the casting process in steel manufacturing only in the order of decreasing width or width of the molds and / or products feasible is.

Take chemical rules into account, for example the starting materials and / or required for the production of different steels Admixtures, both

• - ensure is that the next Melt the previous one without major expense, for example Cleaning of the ladle, can follow, for example, products or steels should required very low carbon content always before products with higher Carbon content are cast, that is, a sequence of products or steels with increasing carbon content has to be considered as well
• - ensure is that successive product groups of the required specification can be manufactured at all are, for example, require some high-purity steels and / or Stainless steels, the so-called "wash degrees ", before the beginning a thorough cleaning of the manufacturing process Operating resources, for example of contaminants and residues from previous ones Melting or steels and / or Casting operations.

Physical rules apply here in particular temperatures to be observed, for example the ovens, the Melt and / or critical temperatures of the equipment used, as well as general specifications and settings to be observed of the various equipment used. For example the change in thickness cast slabs or steel blanks ("slabs") is a time-dependent process that is avoided if possible should be. The equipment used also has limits, so tolerated the continuous caster, the "caster", only 8 oven fillings ("Heats") or melt batches, before extensive maintenance has to be carried out.

The aforementioned rules must be observed or applied, for example, between two consecutive batches of melt and the product groups on which they are based, and for each sequence change. A continuous, uninterrupted casting process is possible within a sequence, which usually comprises 8 to 10 batches of melt and / or furnace fillings. However, a change of sequence requires maintenance and repair work to be carried out, so that the production or continuous casting process must be interrupted , Accordingly, one seeks to achieve the longest possible casting sequences in order to reduce the downtimes of the plant operation or the production failures and accordingly to increase the efficiency of the plant operation or manufacturing process. However, no longer than the required maintenance interval.

Interruptions of continuous casting process occur, for example, when different for the orderly production Products, especially steel blanks, for example different Geometric dimensions of the equipment, here the continuous casting plant become necessary.

At plant downtimes, here the continuous caster, to a minimum and a continuous manufacturing process using one if possible long period possible Efficient design makes optimized planning of the product sequence unavoidable.

As in 2 this is shown in a first step 2 , depending on the product parameters, delivery date, width, quality ("quality"), quality ("grade") and thickness of the respective product, as well as by predetermined rules, which are determined in particular by the manufacturing process, a sequential pre-sorting of the products to be manufactured in individual product families carried out. All products of the same thickness and quality form a product family.

Consider the above rules here essentially geometric, chemical and / or physical Requirements and or restrictions the steelmaking process and / or the product manufacturing process used continuous caster.

In a second step ( 4 ) within each product family, based on at least one preselected product parameter, preferably the quality and / or width, a rule-based grouping of the products to be manufactured into individual product groups is carried out, wherein no more than 8 products should be contained within a product group.

In a third step ( 6 ) a compatibility matrix is determined for each product group depending on the parameters and based on rules, and a mathematical model based on this is created. The compatibility matrix represents or specifies the complex and / or non-linear rules on which the mixed-integer optimization method is based. The matrix elements can have the values 0 or 1. The matrix element P _ii has the value 1 if product i 'can be produced after product i. If this is not the case, then the matrix element P _{ii has} the value 0. This matrix together with the target functions to be optimized, for example minimum downtimes of the plant operation or a continuous production process, embedded in the MILP approach, provide all the information required to carry out a Successful sequencing ready or meet all the requirements for successful sequencing.

According to the invention, either Staying the same and / or increasing or increasing or staying the same and / or descending or decreasing in width within the product sequence allowed for each product group.

The mathematical model has here the aforementioned flexibility to suffice and lets specify as follows:

G gibt hierbei eine Abschätzung der maximal erforderlichen Anzahl von Gruppen an.There are 4 cases C1, C2, C3 and C4 to describe the dependencies:
C1: P _{ii '} = 1, w _i ≤ w _i' , t _i ≤ t _{i '}
C2: P _ii '= 1, w _i' ≤ w _i , t _i ≤ t _{i '}
C3: P _i'i = 1, w _{i '} ≤ w _i , t _i' ≤ t _i
C4: P _i'i = 1, w _i ≤ w _{i '} , t _i' ≤ t _i ,
where w is the width of the respective product and t is a type number describing the quality of the respective product. As a rule, it should be noted that the product with type number 1, i.e. with a higher quality, must always be cast or manufactured before the product with type number 2, i.e. with a lower quality.
z _g be a binary variable that takes the value 1 if a sequence g is used, that is, at least one product is contained, otherwise it has the value 0. The goal is to minimize the total number of sequences or product groups. Expressed by the following relation R1

G gives an estimate of the maximum number of groups required.

α_g ist eine binäre Variable mit dem Wert 1 wenn innerhalb der Produktgruppe bzw. Sequenz g die Produktweite w zunimmt. q_ig ist eine Variable, die dazu dient einige der angegebenen Zielfunktionen für das letzte Produkt einer Sequenz aufzuweichen bzw. zu entschärfen. Bei einer Produktabfolge innerhalb einer Produktgruppe mit ansteigender Weite (α=1) muß es für alle Produkte, ausgenommen das letzte Produkt der Folge, ein geeignetes nachfolgendes Produkt geben, ausgedrückt durch Relation R4. Zielfunktion bzw. Relation R5 gibt den entsprechenden Fall bei abnehmender Weite (a=0) innerhalb einer Produktabfolge an.The assignment of products i to product groups g is done by the binary variable x _ig . Each product i must be clearly assigned to exactly one product group g, as indicated by the objective function or relation R2, with an upper limit M _max for the maximum number of products in a product group being observed, which is illustrated by relation R3. The variables for unused product groups have the value 0. I describes the quantity of products to be manufactured.

α _g is a binary variable with the value 1 if the product width w increases within the product group or sequence g. q _ig is a variable that is used to soften or defuse some of the specified target functions for the last product in a sequence. In the case of a product sequence within a product group with increasing width (α = 1), there must be a suitable subsequent product for all products, except the last product in the sequence, expressed by relation R4. Target function or relation R5 indicates the corresponding case with decreasing width (a = 0) within a product sequence.

Corresponding to the two aforementioned objective functions or relations R4 and R5. describe the two following target functions, R6 for increasing widths and R7 for decreasing widths within a sequence, the case that all products, except the first product, must have a suitable preceding product within a product sequence of a product group or a sequence. The target function is weakened for the first product of a sequence by the variable r _ig .

Because the sequence of products within a product group cannot be controlled directly are two further target functions R8 and R9 required to avoid errors or incompatibilities exclude: The target function R8 prevents the appearance of products with a larger width and with a smaller type number, i.e. high quality, and products with a lower one Wide and with higher Type number, i.e. lower quality, for rising or increasing widths within a sequence or a product sequence a product group. The second objective function or relation R9 describes the corresponding case for decreasing width within a sequence. It should be noted here that when pre-sorting the products depending on the quality in index comparison is sufficient.

The relations R10 and R11 allow exactly two exceptions per group, namely for the first and the last Product of a sequence.

The other target functions or Relations serve to further narrow the search space or to Compacting the model. The "flag" for increasing Widths within a sequence are set to 0 for nonexistent ones Sequences, see relation R12. The product groups are dependent ordered by the number of products they contain, see relation R13. additionally is caused by equation R14 to prepend the active groups become. The number of groups that can be formed | G | is determined by the presorting reached. It is usual, that some of these relations are redundant after optimization appear.

The exception variables for the first and last product in a sequence are real variables in the range between 0 and 1. Z G , x ig , α G ∈ {0.1} 0 ≤ q ig , r ig ≤ 1

Finally, in a fourth step by using more common MILP-solving methods on the compatibility matrix or the mathematical model and solution of the aforementioned Objective functions specified model an optimized sequencing of the products within each product group, the product groups within each product family and therefore also the products achieved within a product family.

In 3 An exemplary pre-sorting in sequences or product groups with the product parameters quality, width, thickness is shown. After pre-sorting, there are 5 sequences separated by thicker horizontal lines. The prerequisites here are that products of different grades must be manufactured in the order 101A → 101B → 101C → 101 and the maximum change in width between two successive products 7 . 0 Units may be.

In 4 is one according to the pre-sorting 3 based compatibility matrix shown in 5 is an optimized grouping and sequencing of the products to be manufactured, based on the assumption that the widths within a sequence only decrease 3 and 4 shown.

In 6 is an optimized grouping and sequencing of the products to be produced based on 3 and 4 shown.

Claims (10)

Process for optimized planning of complex production sequences in large-scale plant operations, especially in the steel industry, by means of a programmatic application of one to methods and algorithms of mixed integer linear programming (MILP) Optimization process, taking into account predetermined Rules and product parameters step by step an optimized grouping and optimized sequencing Products carried out in product families and product groups and determines an optimized product sequence or production sequence becomes. A method according to claim 1, characterized in that - in a first step ( 2 ) a sequential pre-sorting of products to be manufactured according to the order into individual product families is carried out depending on the product parameters and rule-based, - in a second step ( 4 ) within each product family, based on at least one product parameter, a rule-based grouping of the products to be manufactured into individual product groups is carried out, - in a third step ( 6 ) a compatibility matrix is determined for each product group depending on the product parameters and based on rules, a mathematical model based on this is created and - in a fourth step ( 8th ) by using common MILP solution methods on the compatibility matrix or the mathematical model, an optimized sequencing of the products within each product group, the product groups within a product family as well as the products within a product family is carried out and achieved. Method according to one of claims 1 or 2, characterized in that that an optimized grouping and sequencing to produce Products in product groups is achieved by relevant Product characteristics either same and / or increasing or increasing or same and / or descending or decreasing values can be achieved. Method according to one of the preceding claims, characterized characterized that automated and / or manual weightings of the different product parameters. A method according to claim 4, characterized in that based on the weightings of the individual product parameters Product characteristics ranking is created. Method according to one of the preceding claims, characterized characterized that the predetermined rules and / or the weights of the various product parameters the respective plant operation and / or the respective manufacturing process co-determined or shaped become. Method according to one of the preceding claims, characterized characterized in that in the predetermined rules weightings of the different product parameters taken into account become. Method according to one of claims 2 to 7, characterized in that that the sequential pre-sorting is carried out based on weighted product parameters. Method according to one of the preceding claims, characterized characterized that both the ascending and descending of product parameters within allows or enables the product sequence of each product group. System for optimized planning of complex production sequences in large-scale plant operations, especially in the steel industry, with means for the technical application of a method and Mixed Integer Linear Programming (MILP) algorithms mixed-integer optimization methods, with resources available are for that are set up taking into account Predefined rules and product parameters are gradually optimized Grouping and optimized sequencing of products to be manufactured in product families and product groups and an optimized product sequence or to determine the production sequence.