WO1990000290A1

WO1990000290A1 - Simulator

Info

Publication number: WO1990000290A1
Application number: PCT/US1989/002885
Authority: WO
Inventors: Mitchell M. Tseng; Edmund P. Oriuch; Bruce Hayden; John Schaefer; Erik Sand; Nelson Velez; Beth Willstead
Original assignee: Digital Equipment Corporation
Priority date: 1988-06-30
Filing date: 1989-06-29
Publication date: 1990-01-11

Abstract

A simulator including a process model processor which processes a mathematical model of a physical system in connection with input data and a heuristic knowledge base controlled by an inference engine. In addition to the input data, from time to time during processing of the mathematical model, the process model requires heuristic knowledge from the heuristic knowledge base. At such times, the process model processor suspends processing of the mathematical model and enables the inference engine to obtain the required heuristic knowledge and transfer it to the process model processor. Upon receiving the heuristic knowledge, the process model processor resumes processing of the mathematical model.

Description

SIMULATOR

Background of the Invention

1 , Field of the Invention

The invention relates generally to the field of digital computers, and more specifically to simulation systems for performing simulations of physical systems and phenomena.

2. Description of the Prior Art Digital computers have been used for a number of years to assist in the development and design of new products and processes (along with the apparatus for implementing the processes) for manufacturing the products. Collectively, products, processes and apparatus may be generically identified as "systems". In one facet, computers have been used to record aspects of a design, that is, computers, executing computer-assisted design and engineering programs, have been used to maintain and provide specifications, drawings, parts lists, interconnection lists and so forth, of systems being developed. In addition to merely recording during the design phase, computers also assisted in determining when designs could not be implemented. Thus, for example, computers verified that a connection could be made, or that multiple connections were not made, in an electrical circuit, or that surfaces could be machined as designed in a physical device.

More recently, computers have also been used, not only to verify that a system can be made, but also to verify that it would operate in a desirable manner by simulating its operation. Computers have been used to simulate a number of aspects the operations of products and the processes and equipment for manufacturing. To enable a simulation to be performed, a programmer creates a mathematical model of the physical system, that is, the product, process or manufacturing system, to be simulated. The mathematical models, which take the form of mathematical equations, represent the actions of the various components of the system. An operator supplies data representing the initial conditions and the various parameters of operations, and as the simulation proceeds it shows the conditions of the system, either the conditions of the product or the manufacturing equipment or process, at various points in time after the initiation of the simulation.

Summary of the Invention The invention provides a new and improved simulator for performing simulations in connection with physical systems, inlcuding products, and processes and apparatus for manufacturing such products.

In brief summary, the simulator includes a process model processor which processes a mathematical model of a physical system in connection with input data, and a heuristic knowledge base controlled by an inference engine. In addition to the input data, from time to time during processing of the mathematical model, the process model requires heuristic knowledge from the heuristic knowledge base. At such times, the process model processor suspends processing of the mathematical model and enables the inference engine to obtain the required heuristic knowledge and transfer it to the process model processor . upon receiving the heuristic knowledge, the process model processor resumes processing of the mathematical model using the heuristic knowledge it received from the inference engine.

The simulator includes a process framework generator that generates the mathematical model of the physical system to be simulated in response to identification of various items or components of the physical system. The process framework generator uses information in a products knowledge base and an operations library, the products knowledge base containing entries each associated with an item which may be included in the physical system and identifying one or more mathematical model primitives in the operations library. The process framework generator, upon receipt of the identification of an item in the physical system to be simulated, links the mathematical model primitives associated with the item together to create a mathematical model, or framework, for the item. In addition, the process framework generator links the frameworks for all items comprising the physical system together to create a mathematical model for the entire physical system.

The data used by the process model processor is provided by a pre-processor . The pre-proσessor receives, from an operator, data for a particular simulation run. The data includes a number of items, including data that the process model processor actually uses in processing the mathematical model and identification of external-sources of data so used. In response to the data identifying an external source, the pre-processor enables data to be obtained from the external source. The pre-processor creates a formatted input data file containing data from the operator and from the external source, formatted as required by the process model processor.

After the process model processor has finished processing the mathematical model, it creates an ouput file that contains the result of the simulation. A post-processor receives the output file, in response to requests from the operator identifying particular format and post-processing information, creates a formatted output file in which the data is organized as required by the operator.

The new simulator provides a number of benefits. In separating the heuristic knowledge and the mathematical model, the simulator may separately include capabilities for separately processing the heuristic knowledge, which is in the form of rules, and the mathematical model, which is in the form of a procedure. In addition, the heuristic knowledge may be modified or updated independently of the mathematical model or primitives therefor.

Furthermore, by providing the process framework generator, the simulator frees the operator from having to know how to generate a mathematical model. The operator need only identify the items or elements in the physical system to be simulated, and the process framework generator generates the mathematical model. In addition, the pre- and post-processors simplify including a commercially-available process model processor in the simulator. The pre- and post-processors format the data going to, and output from, the process model processor.

Brief Description of the Drawings This invention is pointed out with particularity in the appended claims. The above and further advantages of this invention may be better understood by referring to the following description taken in conjunction with the accompanying drawings, in which:

Fig. 1 is a functional block diagram of a simulator constructed in accordance with the invention;

Figs. 2A through 2E comprise flow diagrams illustrating various operations performed by the modules depicted in the simulator shown in Fig. 1. Detailed Description of an Illustrative Embodiment Fig. 1 is a functional block diagram of a simulator constructed in accordance with the invention. With reference to Fig. 1, the simulator includes a process model processor 10 which processes a mathematical model, in a mathematical model module 18, representing a portion of the system being simulated, and a heuristic knowledge base 11 which contains heuristic knowledge representing other aspects of the operation or characteristics of the system being simulated. The mathematical model module 18 receives the mathematical model, in the form of a PROCESS.DESC process description, from a process framework generator 17. The process model processor 10 processes the mathematical model, in conjunction with the heuristic knowledge base 11, using parameters and data which it receives from a FORM.IN.DAT formatted input data file from a pre-processor 12. After the process model processor has completed processing, it generates an OUT.DAT output data file. Upon receiving an output request from the operator interface 13 identifying output data formats and post-processing information, a post processor module 16 generates a FORM.OUT.DAT formatted output data file which is provided to the operator interface 13.

The pre-processor 12, which receives an INP.DAT input data file from an operator interface 13 and an EXT.DAT external data file from one or more external data sources 14. Some of the data in the INP.DAT input data file comprises data that is to go directly into the FORM.IN.DAT formatted input data file to be used during processing by the process model processor 10. Other data in the IN.DAT input data file identifies external data to be used during processing by the process model processor 10. The data identifying the external data enables the pre-processor 12 to fetch an EXT.DAT external data file from one or more external data sources (not shown) through external data source module 14. For example, an external data source may comprise a component in a factory, from which digital data may be obtained over a network (not shown) . In that case, contents of the INP.DAT input data file idenfity the component from which the data is to be obtained. In response to those contents, the external data source module 14 retrieves the element of digital data from the component and generates the EXT.DAT external data file which it couples to the pre-processor 12. The pre-processor aggregates the contents of the INP.DAT input data file to be used by the process model processor 10 and the EXT.DAT external data files from the external data source module 14 retrieved in response to other contents of the INP.DAT input data file and forms the FORM.IN.DAT formatted input data file, which it couples to the process model processor 10. In* response to the FORM.IN.DAT formatted input data file, the process model processor 10 executes the mathematical model contained in the mathematical model module 18. Now and then during the execution of the mathematical model, the mathematical model requires heuristic knowledge. When that occurs, the mathematical model enables the generation of a HEU.K.REQ heuristic knowledge request, which the process model processor 10 directs to an inference engine 15 that controls the heuristic knowledge base 11. At that point, the process model processor 10 suspends processing of the mathematical model until it receives a response from the inference engine 15.

In response to the receipt of a HEU.K.REQ heuristic knowledge request from the process model processor 10, the inference engine 15 exercises the heuristic knowledge base 11 to obtain heuristic results therefrom and generates a HEUR.K.RESP heuristic knowledge response, which is returned to the process model processor 10. The knowledge in the heuristic knowledge base is in the form of rules, which the inference engine 15 processes using information in the HEUR.K.REQ heuristic knowledge request from the process model processor. The rules that fire in response to the processing by the inference engine 15 provide the heuristic knowledge required for processing of the mathematical model by the process model processor 10. Upon receipt of the heuristic knowledge, the process model processor 10 may resume processing the mathematical model, using the heuristic results in the HEU.K.RESP heuristic knowledge response from the inference engine 15.

After the process model processor 10 has finished processing the mathematical model, it generates an OUT.DAT output data file, which it transfers to a post processor 16. The post processor 16, upon receiving an output request from the operator interface 13 identifying the data formats in which the operator wishes to view the output data, and post-processing information identifying options in which the output data in the OUT.DAT output data file may be combined or otherwise processed, generates a FORM.OUT.DAT formatted output data file, which it transfers to the operator interface 13.

As noted above, the mathematical model processed by the process model processor 10 is generated by a process framework generator 17. The process framework generator 17 operates in response to PHYS.DESC physical description data from the operator interface 13 which describes the system being simulated. . The process framework generator 17 uses the contents of an operations library 20, a business requirements library

21 and a products knowledge base 22, along with the PHYS.DESC physical description data from the operator, to generate a PROC.DESC process description file which contains the mathematical model to be processed by the process model processor.

The products knowledge base 22 contains descriptions of the various products or elements that may be included in the physical system to be simulated. Thus, if the physical system to be simulated is a system for manufacturing a device, the products knowledge base

22 will contain descriptions of the various machines that may be required in a manufacturing environment. The contents of the products knowledge base 22 essentially comprise templates, each template being related to a type of element that may be included in the system. The template may require additional information which requires information from an operator at the operator interface. The template also identifies o contents of the operations library 20 and business requirements library 21 which are necessary for the product framework, that is, the portion of the mathematical model related to the particular element. The business requirements library 21 contains 5 descriptions of various business requirements to be considered in connection with the simulation. The particular business requirements information depends upon the physical system being simulated. If the physical system being simulated is a system for

3₀ manufacturing a device, the business requirements may include, for example, such information as the amount of time each day or week the system may be active, the amount of space available for the system at a plant, cost and profit information, tax information, and so 35 forth. The operations library 20 contains mathematical model primitives of the various physical system elements in the products knowledge base 22 and the business requirements information in the business requirements library 21. In response to each product framework generated in response to physical description input from the operator interface 13, the process framework generator 17 obtains from the operations library 20 the mathematical model primitives related to the product framework and business requirements and links them together. The process framework generator 17 provides the result in the PROCESS.DESC process description file, and couples it to the process model processor 10 for processing in response to the contents of the FORM.IN.DAT formatted input data file from the pre-processor 12.

The simulation system depicted in Fig. 1 provides a number of benefits. First, by dividing the simulation processing between the mathematical model, processed by the process model processor 10, and the heuristic knowledge base 11, exercised by the inference engine 15, the mathematical model can be simplified and used only for the portion of the processing requiring procedural knowledge, with the heuristic knowledge base 11 having heuristic knowledge that may be used during the simulation. The contents of the heuristic knowledge base 11 are in the form of rules which are processed by the inference engine 15 using information in the HEU.K.REQ heuristic knowledge request from the process model processor 10. The process model processor 10, when the mathematical model requires access to the heuristic knowledge in the heuristic knowledge base 11, can, by means of the HEU.K.REQ heuristic knowledge request, pass control to the inference engine 15 essentially as a procedure or subroutine call operation. The provision of the pre-processor 12 and post-processor 16 also simplifies construction of the simulation system. In one specific embodiment, the process model processor 10 is a commercially-available simulation tool "SLAM II". The pre-processor 12 receives the INP.DAT input data file from the operator interface 13, which may be organized in a format logical for the operator in entering the data at the operator interface, and develops the FORM.IN.DAT formatted input data file in a format required by the commercially available simulation tool. Similarly, the organization, of the OUT.DAT output data file provided by the commercially available simulation tool may not be optimum for conveying the particular results of the simulation, and the post-processor 16 provides, in the FORM.OUT.DAT formatted output data file, a desirable output format for presentation to the operator by the operator interface 13.

Furthermore, the generation by the process framework generator 17 of the mathematical model of the physical system being simulated simplifies generation of the mathematical model for processing by the process model processor 10. The operator, by means of the PHYS.DESC physical description data of the physical system provided to the process framework generator 17, provides a physical description of the physical system, and the process framework generator 17 actually prepares the mathematical model. Thus, the operator need not be trained to prepare the mathematical model . Fig. 2A is a flow diagram depicting general operations performed by the simulator depicted in Fig. 1, and Figs. 2A through 2E contain flow diagrams detailing operations performed by various elements depicted in Fig. 1. With reference to Fig. 2A, an operator, through the operator interface 13, initiates ^■ l i ¬

the operation of the other modules of the simulator depicted in Fig. 1. Initially, the operator interface 13 enables the process framework generator 17 to generate the mathematical model of the physical system to be simulated (step 100). The operator interface 13 supplies information, in the PHYS.DESC physical description, to the process framework generator 17 identifying the various elements of the physical system. In response, the process framework generator 17 proceeds to construct the mathematical model, using the information from the operator interface 13 and the content^'s of the products knowledge base 22, business requirements library 21 and operations library 20 (step 101). The process framework generator 17, after finishing construction of the mathematical model, supplies it as the PROCESS.DESC process description, to the mathematical model module 18 for processing by the process model processor 10.

After the mathematical model has been prepared, the operator, through the operator interface 13, supplies data to be used in connection with a particular simulation of the physical system. The operator interface 13 generates the INP.DAT input data file, which it supplies to the pre-processor 12 (step 102). In response, the pre-processor generates the FORM.IN.DAT formatted input data file, including the data from the external source module 14, for processing by the process model processor 10 (step 103).

After the process framework generator 17 has supplied the mathematical model to the mathematical model module 18-and the pre-processor 12 has generated the FORM.IN.DAT formatted input data file, the process model processor 10 processes the mathematical model using the contents of the FORM.IN.DAT formatted input data file, referencing the heuristic knowledge in the heuristic knowledge base as appropriate (step 104) . After processing the mathematical model, the process framework generator 17 generates the OUT.DAT output data file, which contains data representing the results of the simulation.

The post-processor thereafter, using the contents of the OUT.DAT output data file, generates the FORM.OUT.DAT formatted data output file, which contains data formatted for display-to the operator (step 105). The operator interface then displays the data from the FORM.OUT.DAT formatted data output file to the operator (step 106). —

Fig. 2B depicts the sequence of operations performed by the process framework generator 17 in connection with generation of the mathematical model of the physical system being simulated. With reference to Fig. 2B, the process framework generator 17 receives, from the operator interface 13, as part of the description of the physical system to be simulated, the identification of an item or element of the physical system (step 110). The process framework generator 17 then retrieves from the products knowledge base 22 an entry relating to the item (step 111) . The entry essentially includes a template, which requires completion using information obtained from the operator through the operator interface 13. The entry also includes other information, such as identification of business information in the business requirements library 21 and identification of simulation primitives in the operations library for constructing the portion of the mathematical model related to the element.

Based on the information received from the operator interface in step 111, the process framework generator 17 then retrieves the appropriate primitives from the operations library 20 (step 112) . Using the retrieved primitives and the business information in the business requirements library 21, the process framework generator creates a mathematical model related to the item of the physical system and links it to previously created mathematical model associated with previously-identified items of the physical system to be simulated (step 113). Thereafter, the process framework generator 17 determines, by interrogating the operator through the operator interface 13, whether the physical system to be simulated contains any additional items or elements (step 114) . If so, the process framework generator returns to step 110 to generate a portion of the mathematical model for that portion of the physical system. On the other hand, if, in step 114, the process framework simulator 17 determines that the physical system to be simulated contains no additional items or elements, it transfers the mathematical model, as the PROCESS.DESC process description file, to the mathematical model module 18 for processing by the process model processor 10.

Fig. 2C contains a flow chart depicting the operations of the pre-processor 12 in connection with the INP.DAT input data file from the operator interface 13. As noted above, the pre-processor 12 is provided to facilitate interconnection of a commercial simulation system (specifically, a commercially-available process model processor) in a simulator. With reference to Fig. 2C, the pre-processor first receives the INP.DAT input data file from the operator interface (step 120). The INP.DAT input data file may contain a number of items of data, some of which may be used directly by the process model processor 10 in processing the mathematical model in the mathematical model module 18. Other data in the INP.DAT input data file identifies external sources of data which may be so used. Upon receiving the INP.DAT input data file, the pre-processor 12 proceeds to process it item by item. For each item of data, the pre-processor 12 first examines the item to determine whether it identifies an external source of data (step 121). If the item in the INP.DAT input data file does identify an external data source, it transmits to the external data source module 14 a request for the data (step 122), and waits for a response with the data. Alternatively, instead of waiting for the external data source model 14 to respond, it will be appreciated that the pre-processor may return^-~o^"step 121 for the next item in the INPvDAT input data file from the operator interface.

If, in step 121, the pre-processor 12 determines that the item in the INP.DAT input data file is data, rather than the identification of an external data source, or after receiving an item from an external data source following step 122, the pre-processor sequences to step 123. In step 123, the pre-processor determines the relationship between the format of the data item in the INP.DAT input data file or obtained from the external data source module 14 and the format of the data item in the FORM.IN.DAT formatted input data file (step 123). That is, each item of data, either from the INP.DAT input data file or from the external data source module, has one or more fields in particular arrangements, and the item in the FORM.IN.DAT formatted input data file also has one or more fields in particular arrangements. The pre-processor 12 determines relationships between the fields and arrangements of the item from the INP.DAT input data file or external data source module 14 and the fields and arrangements of the item to be inserted into the FORM.IN.DAT formatted input data file. The pre-processor then creates the item in the FORM.IN.DAT formatted input data file and inserts the contents of the item from the INP.DAT input data file or external data source module 14 (step 124).

Following step 124, pre-processor 12 determines whether it has processed all of the items in the INP.DAT input data file. If not, it returns to step 121 to process the next item in the INP.DAT input data file (step 125). On the other hand, if, in step 125, the pre-processor determines that all items in the INP.DAT input data file have been processed, it transfers the now-completed FORM.IN.DAT formatted input data file to the process model processor 10 for processing with the mathematical model in the mathematical model module 13 (step 126) . Fig. 2D depicts the operations performed by the process model processor, which processes the mathematical model contained in the mathematical model module 18 in connection with the contents of the FORM.IN.DAT formatted input data file and the heuristic knowledge in the heuristic knowledge base 11. With reference to Fig. 2D, the process model processor 10 initiates processing of a portion of the mathematical model in the mathematical model module 18, using data from the FORM.IN.DAT formatted input data file (step 130). The process model processor 10 determines, during processing of the various portions of the mathematical model, each relating to a primitive from the operations library 20, whether heuristic knowledge from the heuristic knowledge base 11 is required to process the portion (step 131).

If, in step 131, the process model processor 10 determines that heuristic knowledge is not required in connection with processing of the portion of the mathematical model currently being processed, it completes processing of the portion and sequences to step 132. In step 132, the process model processor 10 determines whether it has processed the entire mathematical model in connection with all of the data items in the formatted input data file, and, if not, it returns to step 130 to process the next portion of the mathematical model. On the other hand, if the process model processor determines in step 132 that it has processed the entire mathematical model in connection with all of the data items in the formatted input data file, it sequences to step 133 in which it creates an OUT.DAT output data—firϊe containing the results of the proceϋ-s-ino of the mathematical model.

As noted above, during execution of step 131, the process model processor 10 determines whether, during processing of a portion of the mathematical model, it requires heuristic knowledge from the heuristic knowledge base to process the portion. If it determines that it needs heuristic knowledge, the process model processor sequences to step 134, in which it generates a HEUR.REQ heuristic request for use by the inference engine 15. The HEUR.REQ heuristic request includes, as arguments, information which the inference engine 15 uses in exercising the rules representing the knowledge in the heuristic knowledge base 11. The inference engine 15 exercises the rules in the heuristic knowledge base 11 and determines results which are provided in a HEU.K.RESP heuristic knowledge response to the process model processor 10 (step 135) . The process model processor then resumes processing of the portion of the mathematical model, using the results in the HEU.K.RESP heuristic knowledge response. After processing the portion, the process model processor sequences to setp 132 to determine whether it has finished processing the entire mathematical model (step 136). The inference engine 15, in step 135, exercises the rules in the heuristic knowledge base in a conventional manner. Upon obtaining the HEUR.REQ heuristic request, the inference engine 15 determines which rules fire in response to the arguments in the HEUR.REQ heuristic request, and other data which may be in the heuristic knowledge base. That is, the inference engine determines which rules conditions are satisfied by the arguments and data. The rules which fire may result in firing of additional rules. As the rules fire, the inference engine collects the information contained therein and forms the HEUR.K.RESP heuristic knowledge response. When no further rules fire, the inference engine 15 supplies the HEUR.K.RESP heuristic knowledge response to the process model processor 10. After the process model processor 10 has completed processing, the OUT.DAT output data file generated by it is supplied to the post-processor 16 for formatting for use by the operator interface. The operations performed by the post-processor 16 are shown in the flow chart depicted in Fig. 2E. With reference to Fig. 2E, the post-processor receives the OUT.DAT output data file from the process model processor 10 (step 140). The OUT.DAT output data file comprises one or more items of data, with fields and items arranged in a format determined by the process model processor. The post-processors 16 generates a FORM.OUT.DAT formatted output data file with items of data arranged in a format requested by the operator interface 13. The post-processor 16 determines the relationship between the format of the data items in the OUT.DAT output data file and of the data items in the FORM.OUT.DAT formatted output data file (step 141). The post-processor creates a data item in the FORM.OUT.DAT formatted output data file and inserts into it data from the OUT.DAT output data file (step 142).

The post-processor then determines whether items have been created in the FORM.OUT.DAT formatted

5 output data file for all of the data items in the output data file, and it not, returns to step 141 to continue processing (step 143). If, in step 143, the post-processor determines that data items have been created in the FORM.OUT.DAT formatted output data file ₀ for all of the data items in the output data file, it sequences to step 144, in which it transfers the

FORM.OUT.DAT formatted output data file to the operator interface 13 for display.

As noted above, the new simulator provides a

*ι_5 number of benefits. In separating the heuristic knowledge and the mathematical model, the simulator may separately include capabilities for separately processing the heuristic knowledge, which is in the form of rules, and the mathematical model, which is in the

20 form of a procedure. In addition, the heuristic knowledge may be modified or updated independently of the mathematical model or primitives therefor.

Furthermore, by providing the process framework generator, the simulator frees the operator from having

25 to know how to generate a mathematical model. The operator need only identify the items or elements in the physical system to be simulated, and the process framework generator generates the mathematical model. It will be appreciated however, that an operator may ₃0 directly provide the mathematical model for use in a simulation, without requiring the operation of the process framework generator 17.

In addition, the pre- and post-processors simplify including a commercially-available process 35 model processor in the simulator. The pre- and post-processors format the data going to, and output from, the process model processor.

The foregoing description has been limited to a specific embodiment of this invention. It will be apparent, however, that variations and modifications may be made to the invention, with the attainment of some or all of the advantages of the invention. Therefore, it is the object of the appended claims to cover all such variations and modifications as come within the true spirit and scope of the invention.

Claims

1. A simulator for simulating a physical system in response to input data comprising:

A. a mathematical model representing portions of said physical system;

B. a heuristic knowledge base containing a plurality of rules containing heuristic knowledge;

C. a mathematical model processing means for processing said mathematical model in response to said input data and heuristic knowledge responses and for generating, in response to said mathematical model, a heuristic knowledge request; and

D. inference engine means responsive to a heuristic knowledge request, for exercising rules in said heuristic knowledge base and for generating, in response thereto, a heuristic knowledge response for use by said mathematical model processing means.