41 results about "SimulTrain" patented technology

A flexible and extensible automated design partitioning mechanism that facilitates simulation sessions employing two or more simulators is provided. A simulation backplane includes partitioning logic that identifies the design blocks of an overall design pertaining to each of a plurality of simulators. Once the partitions have been identified, nets that cross simulator boundaries (e.g., mixed nets) are determined and inter-simulator connectivity information is generated for the simulators. According to one aspect of the present invention, the partitioning logic is able to accomodate arbitrary (e.g., instance-based) partitioning. A design source expressed in a design representation upon which a first simulator may operate is received. Design blocks to be partitioned to each of a plurality of solvers are identified based upon one or more partitioning directives and the design source. A first instance of a cell is assigned to a first solver and a second instance of the cell is assigned to a second solver. Netlist like information is generated for those of the design blocks that are partitioned to a non-design source solver. To accomodate a folded representation of a design block containing the first or second instance of the cell, one or more additional cells are created. According to another aspect of the present invention, name space mapping is retained across all simulators. A design source upon which a first simulator may operate is read. Based upon a set of rules, a primary partition that is to be simulated by a first solver is identified. The primary partition includes a top cell of the design representation. Additionally, a secondary partition that is to be simulated by a second solver is identified. Subsequently, netlist information is generated for the second solver while retaining name space mapping in the secondary partition by adding one or more levels of hierarchy so as to include information about the top cell in the secondary partition.

A method of creating a secure intellectual property (IP) representation of a circuit design for use with a software-based simulator can include translating a hardware description language representation of the circuit design into an encrypted intermediate form and compiling the intermediate form of the circuit design to produce encrypted object code. The method further can include linking the encrypted object code with a simulation kernel library thereby creating the secure IP representation of the circuit design. The secure IP can include an encrypted simulation model of the circuit design and a simulation kernel configured to execute the encrypted simulation model.

The invention relates to a distributed simulation system and method supporting a large-scale complicated delay tolerant network (DTN). The conventional DTN simulation tool can only run on a single machine, is limited in operation capacity and low in simulation performance and can support a small scale of simulation nodes. The system comprises a main controller and a plurality of simulators which are positioned in the same local area network; the DTN is partitioned into grids, and an administration region is distributed to each simulator; and the main controller processes a simulation event by promoting a simulation clock and manages simulation of the simulators to be synchronous with the states. By the distributed simulation system and the distributed simulation method, a great number of calculation demands in simulation of the DTN are distributed to different simulators, so that the simulation speed is obviously increased; by adopting a basic network topological structure with one main controller and a plurality of simulators, the expandability is high; the requirement on a DTN network simulation scheme is low; distribution simulation can be realized by the system and the method only when the simulation scheme supports linear partitioning of a simulation network; and the generality is high.

A simulator includes a memory for storing a first netlist, a timing library, and a standard parasitic exchange format (SPEF) file; and a processor configured to compensate for delay to synchronize digital and analog signals. The processor includes a delay calculator module for generating one of a rising time and a falling time and a standard delay format (SDF) file using the first netlist, the timing library, and the SPEF file; an SDF file converter module for adjusting an interconnect delay description included in the SDF file to compensate for delay using the one of the rising time and the falling time; and a digital simulator module for generating an event using a first driving cell according to a compensated interconnect delay description.

The invention discloses a distributed combined simulation system based on an FMI and a construction method. The distributed combined simulation system based on the FMI comprises a server, a master control terminal software, a sub-calculation terminal, a terminal software and an emulator. The emulator utilizes FMI combined simulation ports which are imported by various kinds of heterogeneous platforms to perform model calculation according to the FMI port logic; the master control terminal software finishes variable connection relation configuration between a system dispatching model and a sub-system calculation model; and the terminal software finishes subsystem network configuration. In the system operation process, the emulator operation system in the server operates the system dispatching model and is synchronized with the system data, and the emulator in the sub-calculation terminal operates the sub-system model. The distributed combined simulation system based on the FMI divides the system simulation into the system dispatching and system calculation, which reduces the coupling between the sub-systems, can fast realize the cooperation and combined simulation between the heterogeneous platforms, can fully utilize the calculation resource, and can solve the problems that the mass simulation system construction process is complex and the calculation efficiency is low.

The invention discloses an intelligent operation and maintenance method based on deep reinforcement learning, and relates to the technical field of automatic operation and maintenance, and the method comprises the following steps: collecting data information in advance, and importing the data information into an architecture simulator through a data collector; constructing an operation and maintenance digital twin environment for data information acquired by the data acquisition unit; loading the model from the rule storage database by the reasoning service; and after an operation and maintenance event occurs, performing reasoning through reasoning service, obtaining a matched PlayBook file suitable for being used by the Ansible, and triggering an operation and maintenance action. According to the method, various operation and maintenance data are used, the training environment of deep reinforcement learning is kept consistent with the production environment, the reliability and the real-time performance are enhanced, the operation and maintenance rules are trained by adopting the deep reinforcement learning method, manpower and material resources of rule maintenance are reduced, the accuracy and the effectiveness of the operation and maintenance rules are improved. Automatic operation and maintenance rules are found intelligently, the operation and maintenance rules are updated fully automatically, and the condition of rule aging in automatic operation and maintenance is greatly reduced.

Disclosed is a unified modeling method supporting design and simulation of a complex system. Two forms of models of the complex system are established according to a preset graph and text modeling framework, mutual conversion of the two forms of the complex system model is achieved through an image-text conversion modeling tool, and finally the code is explained into a C++ code through an interpreter, and simulation runs in a DEVS simulator environment. System design and simulation verification of a complex system model are integrated through two modeling forms of graphs and texts, the efficiency of the research and development process of the complex system is improved, the learning cost of modeling personnel is reduced, the difficulty in changing the design of the complex system model is avoided, and the consistency is difficult to guarantee.

The invention belongs to the technical field of digital physical hybrid simulation, and particularly provides a digital-analog hybrid simulation interface test platform based on double real-time simulators. The platform comprises two real-time simulators respectively simulating a digital side and a physical side, the simulators run in parallel, and the actual working condition of the hybrid simulation system is simulated through real physical interface connection. The test platform is closer to the actual working condition of the hybrid simulation system than off-line simulation, the verification of the validity of the simulation interface under different working conditions can be quickly realized through the platform, the parameters can be set on line, and the research and development process is greatly accelerated. The platform solves the problems of lack of actual working condition verification and difficulty in parameter setting in interface algorithm research.

The invention relates to a model continuity checking method, which comprises the following steps of: respectively checking the continuity of a model current and a model charge by using Verilog-A Compeller; performing model current and charge continuity check according to the circuit matrix numerical solution; and reporting inspection result information. According to the method for checking the continuity of the model, an analytical formula of the model is analyzed by utilizing Verilog-A Compiler, so that the continuity of a segmented bias part can be quickly checked; dynamically checking the current and charge values in the simulation process; quickly checking discontinuous bias points through dynamic checking, reporting discontinuous point information, providing insight for problem analysis for simulator developers, and improving software development efficiency. Static inspection is performed first and then dynamic inspection is performed so that excessive inspection steps can be avoided.

The invention discloses a panoramic information physical fusion system simulation architecture and a working method thereof. The system comprises a software simulator and an external device system which are connected with each other. The method comprises the steps of obtaining state information data of a physical system through the software simulator; and sending the state information data to theexternal device system according to a data acquisition protocol. The software simulator is connected with an external device through a network socket interface, interaction between a CPS simulation device and the external device / system is achieved through UDP protocol interaction, information-physical simulation interaction is achieved, system maintenance is convenient, and a simulation scale canbe increased or decreased at any time.

The invention relates to a TMS32C6748 program burning method based on serial port high reliability. The TMS32C6748 program burning method is technically characterized by comprising the steps: planning18 sectors of Flash planned; generating a user bootstrap program; and pre-burning a user bootstrap program UBL at the planned Flash storage address; when the user bootstrap program UBL is started, judging a burning mark of the SA18 sector pair; if the burning mark is the burning mark, carrying out burning, and otherwise, loading a working program. The functions of achieving communication and completing FLASH writing by rapidly using a debugging interface of the system in real time without disassembling or assembling a simulator are achieved. Compared with direct programming by using an emulator, the method has the advantages that the step of frequent disassembly is omitted; and meanwhile, programming inquiry is actively sent to each serial port address when a program is started, and the programming inquiry can be directly used in each project without debugging different modification codes of the serial ports for each project, so that the universality and the code security are improved.

The invention provides a model-based TC1797 diesel engine control algorithm development tool. The invention comprises the following steps of: customizing a TC1797 target platform Simulink model library, seamlessly installing the TC1797 target platform Simulink model library in a Simulink environment, and configuring relevant parameters through a graphical interface; customizing an embedded frame code and an embedded code generation template of the TC1797 target platform; automatically integrating an integrated development environment of TASKING and a one-button model to C code generation, automatically compiling an executable program and downloading to the target platform through an emulator, achieving rapid design, test and verification of an algorithm. According to the method, the Simulink module library is driven through the customized graphical target platform interface to control the algorithm model to quickly generate the executable code of the target platform, and convenient configuration of the development environment of the target platform and unified management of engineering projects can be realized.