Performance based system configuration as preprocessing for system peformance simulation
AssigneeSiemens Aktiengesellschaft
InventorsHeinrich Helmut Degen, Arun Ramamurthy, Yunsheng Zhou
Life of this patent
Estimating the term…
Abstract
System for performance-based design space composition in engineering systems includes design configuration module, explainable recommendation module and user interface module. A list of one or more system configurations is generated, each system configuration having a unique set of components that satisfy system goals and system requirements. For each system configuration, system performance values are determined based on performance enabled by each system component. A design dashboard presents the system configurations in rank order and the system performance value for each of the system configuration components.
Claims10
A system for performance-based system design configuration generation as preprocessing for optimal system performance simulation, the system comprising:
- a processor;
- and
- a non-transitory memory having stored thereon modules executed by the processor, the modules comprising:
- a design configuration module configured to receive seed designs, a specification of requirements for a system design, and user defined goals for the system, and to generate a set of system design configurations, each system design configuration comprising a unique set of components that satisfy system goals and system requirements;
- an explainable recommendation module configured to:
- perform a requirement decomposition that decomposes requirements into function and physical block representations that are mappable to the components;
- determine, for each system design configuration, an estimated system performance value associated with each component of the system design configuration,
- wherein the value represents the component-based contribution to performance of the system configuration based on a simulation model;
- and
- rank the list of system configurations according to system performance values;
- and
- a user interface module configured to communicate information exchanged between a graphical user interface and the explainable recommendation module, wherein the graphical user interface displays a design dashboard that presents the system configurations in rank order, and for each system configuration, the estimated system performance value for each of the system configuration components;
- wherein the top ranked system configurations are recommended to the user for performance simulation setup;
- wherein the explainable recommendation module implements a machine learning based model trained to predict performance of system configurations.
A method for performance-based system design configuration generation as preprocessing for optimal system performance simulation, comprising:
- receiving seed designs, a specification of requirements for a system design, and user defined goals for the system;
- generating a set of system design configurations, each system design configuration comprising a unique set of components that satisfy system goals and system requirements;
- performing a requirement decomposition that decomposes requirements into function and physical block representations that are mappable to the components;
- determining, for each system design configuration, an estimated system performance value associated with each component of the system design configuration,
- wherein the value represents the component-based contribution to performance of the system configuration based on a simulation model;
- ranking the list of system configurations according to system performance values;
- and displaying a design dashboard on a graphical user interface that presents the system configurations in rank order, and for each system configuration, the estimated system performance value for each of the system configuration components;
- wherein the top ranked system configurations are recommended to the user for performance simulation setup;
- further comprising training a machine learning based model to predict performance of system configurations.
Dependencies are inferred from the claim text.