System, method and computer program for controlling a production plant consisting of a plurality of plant parts, in particular a metallurgical production plant for producing industrial goods such as metal semi-finished products and/or metal end products
20230251639 · 2023-08-10
Assignee
Inventors
- Thomas Daube (Duisburg, DE)
- Holger Behrens (Erkrath, DE)
- Lutz KÜMMEL (Jüchen, DE)
- Ulrich SOMMERS (Düsseldorf, DE)
Cpc classification
Y02P90/02
GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
G05B2219/32015
PHYSICS
International classification
Abstract
The invention relates to a system (1) for controlling a production plant (3) consisting of a plurality of plant parts (2), in particular a metallurgical production plant for producing industrial goods such as metal semi-finished products and/or metal end products, wherein each plant part (2) has an input quality window (4), an output quality window (5) and a process window (6), wherein the input quality window (4) of a plant part (2) defines the quality characteristics of the input product that are required by the plant part (2) and the output quality window (5) of a plant part (2) defines the quality characteristics of the output product that are allowed by the plant part (2) after processing the input product, wherein, in a production plant (3) consisting of the plurality of plant parts (2), the output quality window (5) of an upstream plant part (2) corresponds to the input quality window (4) of the downstream plant part (2), wherein the process window (6) defines the setting values (7) that can be implemented by the respective plant part (2) for a plant automation unit of the plant part (2), wherein each plant part (2) detects the current state by means of sensors (8) and adapts the process window (6) of the plant part (2) to the detected current state, and wherein the system (1) for controlling the production plant (3) consisting of the plurality of plant parts (2) determines setting values (7) for the respective plant automation unit for each plant part (2), the setting values being within the process windows (6) and that the product produced in the production plant (3) meets the quality characteristics required by the input quality windows (4) and output quality windows (5) of the plurality of plant parts (2).
The invention further relates to a corresponding method and computer program.
Claims
1. System (1) for controlling a production plant (3) consisting of a plurality of plant parts (2), in particular a metallurgical production plant for producing industrial goods such as metal semi-finished products and/or metal end products, wherein each plant part (2) has an input quality window (4), an output quality window (5) and a process window (6), wherein the input quality window (4) of a plant part (2) defines the quality characteristics of the input product that are required by the plant part (2) and the output quality window (5) of a plant part (2) defines the quality characteristics of the output product that are allowed by the plant part (2) after processing the input product, wherein, in a production plant (3) consisting of the plurality of plant parts (2), the output quality window (5) of an upstream plant part (2) corresponds to the input quality window (4) of the downstream plant part (2), wherein the process window (6) defines the setting values (7) that can be implemented by the respective plant part (2) for a plant automation unit of the plant part (2), wherein each plant part (2) detects the current state by means of sensors (8) and adapts the process window (6) of the plant part (2) to the detected current state, and wherein the system (1) for controlling the production plant (3) consisting of the plurality of plant parts (2) determines setting values (7) for the respective plant automation unit for each plant part (2), the setting values being within the process windows (6) and that the product produced in the production plant (3) meets the quality characteristics required by the input quality windows (4) and output quality windows (5) of the plurality of plant parts (2).
2. System (1) according to claim 1, Wherein, after the product to be produced has been processed by a plant part (2), the system (1) updates the setting values (7) for the plant automation units of the subsequent plant parts (2) on the basis of the achieved output quality of the processed plant part (2) and the current process windows (6) of the subsequent plant parts (2).
3. System (1) according to claim 1 or claim 2, wherein the system (1) determines a sequence of products to be produced in the production plant (3), in particular taking into account the current process windows (6) and the achievable quality characteristics of the output products of the plant parts (2).
4. System (1) according to any one of the claims 1 to 3, wherein the system (1) takes into account the possible setting value change rates that can be implemented by the respective plant automation units when determining the setting values (7) for the plant automation units of the respective plant parts (2).
5. System (1) according to any one of the claims 1 to 4, wherein the system (1) creates a prediction model (9) for the future states of the plant parts (2) and takes into account the states of the plant parts (2) predicted by the prediction model (9) and the resulting process windows (6) of the plant parts (2) when determining the setting values (7) for the plant automation units of the plant parts (2).
6. System (1) according to claim 5, wherein the prediction model (9) is based on the states of the plant parts (2) detected by means of the sensors (8), the achieved product qualities of the plant parts (2), other measured values associated with the production plant (3) or the like.
7. System (1) according to any one of the claims 1 to 6, wherein the system (1) takes into account a plurality of plant parts (2) of the same type, such that the processing of a production step can alternatively take place on different plant parts (2).
8. System (1) according to any one of the claims 1 to 7, wherein the plant parts (2) comprise at least in part optical sensors (8) for detecting geometric information of the products produced in the respective plant part (2).
9. System (1) according to any one of the claims 1 to 8, wherein the setting values (7) for the plant automation units of the plurality of plant parts (2) are determined by means of a model (12).
10. Method for controlling a production plant (3) consisting of a plurality of plant parts (2), in particular a metallurgical production plant for producing industrial goods such as metal semi-finished products and/or metal end products, wherein each plant part (2) has an input quality window (4), an output quality window (5) and a process window (5), wherein the input quality window (4) of a plant part (2) defines the quality characteristics of the input product that are required by the plant part (2) and the output quality window (5) of a plant part (2) defines the quality characteristics of the output product that are allowed by the plant part (2) after processing the input product, wherein, in a production plant (3) consisting of a plurality of plant parts (2), the output quality window (5) of an upstream plant part (2) corresponds to the input quality window (4) of the downstream plant part (2), wherein the process window (6) defines the setting values (6) that can be implemented by the respective plant part (2) for a plant automation unit of the plant part (2), wherein the method comprises the steps of: detecting the current states in the plurality of plant parts (2), in particular by means of sensors (8), adapting the respective process windows (6) of the plurality of plant parts (2) based on the detected current states of the plurality of plant parts (2), and determining respective setting values (7) for the plant automation units of the plurality of plant parts (2) of the production plant (3), wherein the determined respective setting values (7) are within the adapted respective process windows (6), and wherein the product produced in the production plant (3) meets the quality characteristics required by the input quality windows (4) and output quality windows (5) of the plurality of plant parts (2).
11. Method according to claim 10, further comprising the step of updating the setting values (7) for the plant automation units of the subsequent plant parts (2) after the product to be produced has been processed by a plant part (2), wherein the updating takes place on the basis of the achieved output quality of the processed plant part (2) and the current process windows (6) of the subsequent plant parts (2).
12. Method according to claim 10 or claim 11, further comprising the step of determining a sequence of products to be produced in the production plant (3), in particular taking into account the current process windows (6) and the achievable quality characteristics of the output products of the plant parts (2).
13. Method according to any one of the claims 10 to 12, comprising the step of taking into account possible setting value change rates that can be implemented by the respective plant automation units when determining the setting values (7) for the plant automation units of the respective plant parts (2).
14. Method according to any one of the claims 10 to 13, comprising the step of building a prediction model (9) for the future states of the plant parts (2), wherein the states of the plant parts (2) predicted by the prediction model (9) and the resulting process windows (6) of the plant parts (2) are taken in account when determining the setting values (7) for the plant automation units of the plant parts (2).
15. Method according to claim 14, wherein the prediction model (9) is based on the states of the plant parts (2) detected by means of the sensors (8), the achieved product qualities of the plant parts (2), other measured values associated with the production plant or the like.
16. Method according to any one of the claims 10 to 15, comprising the step of taking into account the plurality of plant parts (2) of the same type, such that the processing of a production step can alternatively take place on different plant parts (2).
17. Method according to any one of the claims 10 to 16, comprising the step of detecting geometric information of the products produced in the respective plant part (2), in particular by means of optical sensors (8) in the respective plant parts (2).
18. Method according to any one of the claims 10 to 17, comprising the step of creating a model (12) for determining the setting values (7) for the plant automation units of the plurality of plant parts (2).
19. Computer program comprising instructions that, when the program is executed by a computer, cause the computer to execute the method according to any one of the claims 10 to 18, in particular that the system (1) according to any one of the claims 1 to 9 executes the method according to any one of the claims 9 to 18.
Description
[0073] The invention is explained in more detail below with reference to exemplary embodiments shown in the figures. The following are shown:
[0074]
[0075]
[0076]
[0077] Thereby, the system in accordance with the invention can be arranged inside or outside the production plant 3, wherein the system 1 is designed to communicate with the production plant 3.
[0078] Each plant part 2 of the production plant 3 has an input quality window 4, an output quality window 5 and a process window 6. The input quality window 4 of a plant part 2 defines the quality characteristics of the input product that are required by the plant part 2, and the output quality window 5 of a plant part 2 defines the quality characteristics of the output product that are allowed by the plant part 2 after processing the input product. With a production plant 3 consisting of the plurality of plant parts 2, the output quality window 5 of an upstream plant part 2 corresponds to the input quality window 4 of the downstream plant part 2, which is shown in
[0079] The process window 6 defines the setting values 7 that can be implemented by the respective plant part 2 for a plant automation unit of the plant part 2. The process window 6 is symbolized in
[0080] Each plant part 2 detects its current state by means of suitable sensors 8. On the basis of the detected current states of the plant parts 2, the process windows 2 of the plant parts 2 are adapted. For example, the process windows 6 may move and/or change their size in a specified parameter space. This is shown in
[0081] The system 1 in accordance with the invention for controlling the production plant 3 consisting of the plurality of plant parts 2 determines setting values 7 for each plant part 2 for the respective plant automation unit. The setting values 7 determined by the system 1 in accordance with the invention are within the respective process windows 6 and are selected such that the product produced in the production plant 3 meets the quality characteristics required by the input quality windows 4 and the output quality windows 5 of the plurality of plant parts 2. Since the product produced in the production plant 3 is in the output quality window 5 of the last plant part 2, the quality requirements imposed on the produced product are met.
[0082] Preferably, after the product to be produced has been processed by a plant part 2, the system 1 in accordance with the invention updates the setting values 7 for the plant automation units of the subsequent plant parts 2 on the basis of the achieved output quality of the processed plant part 2 and the current process windows 6 of zo the subsequent plant parts 2. The setting values 7 for the plant parts 2 still to come are thus continuously adapted, taking into account the processing already carried out by previous plant parts 2.
[0083] In an advantageous variant, the system 1 determines a sequence of products to be produced in the production plant 3, in particular taking into account the current process windows 6 and the achievable quality characteristics of the output products of the plant parts 2. The system 1 preferably takes into account the differences and/or similarities of products to be produced when determining the sequence of products to be produced in the production plant 3.
[0084] When determining the setting values 7 for the plant automation units of the respective plant parts 2, the system 1 takes into account the possible setting value change rates that can be implemented by the respective plant automation units.
[0085] According to a preferred variant, after one or more plant parts 2, the system 1 comprises a quality control, for checking the achieved product qualities, in particular for adapting the setting values 7 for the plant automation units of the subsequent plant parts 2. Thus, at least partially after the plant parts 2, there is a check of whether the produced product is within the output quality window 5. Depending on the determined quality, i.e. the position of the produced product within the output quality window 5, the setting values 7 for the plant automation units of the subsequent plant parts 2 are adapted.
[0086] The sensors 8 of the plant parts 2 detect, for example, the wear, the maintenance state or the like of the plant part 2. In particular, the plant parts 2 comprise at least in part optical sensors 8 for detecting geometric information of the products produced in the respective plant part 2. From the geometric information, the system 1 can derive the product quality of the product produced in the plant part 2 and/or the plant state of the plant part 2, preferably by means of statistical methods and/or machine learning.
[0087] In accordance with a particularly preferred variant of the invention, the system 1 creates a prediction model 9 for the future states of the plant parts 2 and takes into account the states of the plant parts 2 predicted by the prediction model 9 and the resulting process windows 6 of the plant parts 2 when determining the setting values 7 for the plant automation units of the plant parts 2. The prediction model 9 is based, for example, on the states of the plant parts 2 detected by means of the sensors 8, the achieved product qualities of the plant parts 2, other measured values associated with the production plant 3 or the like.
[0088] The prediction model 9 is based on methods of statistical data analysis and/or machine learning, in particular linear or quadratic programs, genetic optimization, reinforcement learning with Q-tables, neural networks, simulated annealing, Metropolis, swarm algorithms, hill climbing, the Lagrange multiplier method or the like.
[0089] The prediction model 9 is trained, for example, continuously or cyclically, in particular on the basis of the states of the plant parts 2 detected by means of the sensors 8, the achieved product qualities of the plant parts 2, other measured values associated with the production plant 3 or the like.
[0090] According to an advantageous variant, the system 1 stops the production process in the production plant 3 if the required quality of the produced product cannot be achieved. If possible, the system 1 can change the production process to a different product to be produced, the required quality of which can still be achieved by the current production process. Expediently, the change to a different product to be produced is limited to the pre-planned production of products to be produced within a specified period of time.
[0091] Moreover, the system 1 can take into account a plurality of plant parts 2 of the same type, such that the processing of a production step can alternatively take place on different plant parts 2.
[0092] The system 1 from
[0093] To display information relevant to the production process, the system 1 comprises a user interface 11. The user interface is provided on a portable device, for example, such that the information is also available within the production plant, for example during an inspection.
[0094] The system 1 in accordance with the invention further comprises a model 12 for determining the setting values 7 for the plant automation units of the plurality of plant parts 2. The model 12 is based on laws of physics or on methods of statistical data analysis and/or machine learning, in particular linear or quadratic programs, genetic optimization, reinforcement learning with Q-tables, neural networks, simulated annealing, Metropolis, swarm algorithms, hill climbing, the Lagrange multiplier method or the like. Like the prediction model 9, the model 12 can be trained continuously or cyclically.
[0095]
[0096] The invention differs from the prior art by detecting the current states in the plurality of plant parts 2, in particular by means of sensors 8, adapting respective process windows 6 of the plurality of plant parts 2 based on the detected current states of the plurality of plant parts 2, and determining respective setting values 7 for the plant automation units of the plurality of plant parts 2 of the production plant 3, wherein the determined respective setting values 7 are within the adapted respective process windows 6, and wherein the product produced in the production plant 3 meets the quality characteristics required by the input quality windows 4 and output quality windows 5 of the plurality of plant parts 2.
[0097] According to the prior art, it is possible that, starting from the processing by a previous plant part 2 and the current process windows 6 of the current plant part 2, the setting values 7 for the plant automation unit of the current plant part cannot be set in such a manner that the produced product is within the output quality window 5 of the current plant part 2. This is symbolized in
[0098] In accordance with the present invention, the current process windows 6 of all plant parts 2 are taken into account when determining the setting values 7 for the plant automation units of the plant parts 2. Thus, the process windows of the plant parts 2 at the end of the production process are already taken into account for the plant parts 2 at the beginning of the production process. The setting values 7 are thus determined in that the product produced in the production plant 3 is within all input quality windows 4 and output quality windows 5 of the plant parts 2. This is shown in
LIST OF REFERENCE SIGNS
[0099] 1 System
[0100] 2 Plant part
[0101] 3 Production plant
[0102] 4 Input quality window
[0103] 5 Output quality window
[0104] 6 Process window
[0105] 7 Setting values
[0106] 8 Sensors
[0107] 9 Prediction model
[0108] 10 Data storage unit/cloud storage unit
[0109] 11 User interface
[0110] 12 Model for predicting setting values