Title

Author

Abstract

The Qualitative Simulation Based on Deep Knowledge Model of Signed Directed Graph

Qualitative simulation is an important research and application area of the simulation technology, as well as the artificial intelligence. Recently, the highly advanced computational technology has led to the development of the signed directed graph (SDG) deep knowledge model based qualitative approach. SDG model can be used to express the complex cause and effect relations, and has very large capacity of containing process potential information. Therefore, it is especially suitable for process hazard assessment (PHA), hazard identification and fault diagnosis. The structure of SDG model is much simple and easy of establishment and modification. However, the high quality SDG model must be verified and improved by plant information, experience or dynamic quantitative simulation. The key technique of using SDG for qualitative simulation is the high efficiency two-direction inference engine, which is performed by software automatically. Up to now, SDG is the most effective method in computer aided process hazard assessment. In fault diagnosis, SDG has better completeness, but lower resolution. The future directions would aim at the combination of qualitative and quantitative methods, SDG automatic modeling, the generalized inference engine software environment and high resolution fault diagnosis.

Petrochemical Safety Simulation Techniques and Applications

The safety simulation technique is an important branch of safety science and technology. Along with the development of information and computational technology, the simulation technique is becoming more and more useful in the petrochemical process hazard assessment, hazard identification, safety control systems design and operating personnel training. At present, the simulation technique has been used successfully in disclosure of fault propagation, fatigue failure of devices, 3D simulation of materials fire, release and explosion, as well as in on-line fault diagnosis and safety related training. Simulation technique is a key technique in safety engineering.

Integrated SDG Modeling, Inference and Post-Processing Software Platform                      

Process safety issues have caused ever-increasing public concern and worry. Enormous labour and material resources have been invested in Process Hazard Analysis (PHA), most of which are carried out manually. Recently, by importing SDG, the automatic computer based PHA becomes practical. During the process, the most important task is to create a practical SDG platform for modeling and inference. In this paper, we discuss the progress in our laboratory on safety analysis research for process industry. Based on research in HAZOP (Hazard & Operability) analysis and according to its steps and requirements, we have designed and developed a graph based, layered structured and components reusable software integrated platform. This platform, which is based on a SDG Engine and a graphic modeling environment developed by ourselves, fully considers progress of SDG research, and uses a variety of iterative development technique and object orient programming (OOP). Within this platform, various algorithms can be used together and their components can be easily reused. It is a good start for SDG graph modeling, inference engine structure and dual direction inference.

SDG Modeling Approach for Chemical Engineering Process                             

SDG is an important technology of building qualitative model, especially for hazard evaluation and fault diagnosis. Two modeling approaches based on mathematics model and flow chart are introduced. Besides, a modeling approach based on experience is put forward. Hazard evaluation emphasizes on completeness of revealing latent fault. Fault diagnosis requires high diagnosis speed and fault resolution. Therefore, experience based modeling approach is divided into modeling approach of hazard evaluation model and modeling approach of fault diagnosis model. Hazard evaluation model built by this approach can not only reflect faults completely, but also reveal the rules of fault propagation. Fault diagnosis model built by this approach is simple and has high fault resolution.

Test Graph for SDG Inference Engine and Its Demonstration           

During the research and application of SDG (signed directed graph) in solving safety problems of large-scale complex system, it’s very important to find an efficient and credible way to assess inference Engine’s power. We propose a series of standard and unified test methods based on structured graphs, by which inference engine’s capacities, such as veracity, completeness, speed, and resolution, can be tested systematically. Through adjusting scale, status generation, relationship, position and shape of node and route, we create a flexible method to generate a set of structured graphs which are suitable for various testing application. Besides, software and hardware environmental requirements of engine test have been discussed, and some suggestion of its reasonable scale and test guideline are given. During application, we recommend that the test should be carried on with two scales, e.g. peak and average test, to evaluate its performances respectively under extreme status and normal status. At the end of this article, we give a demonstration of testing the SDG inference engine we developed, and the result proves that it’s practical. As for some problems of our engine which are found in test, some improvement measurements have been proposed.

Event Tree Modeling and Its Application in Petrochemical Hazard Assessing Software

In the petrochemical industry, it is a very important task to conduct hazard assessing. Event tree modeling is one of the useful methods for system hazard assessing. Assessing Industrial Hazards Techniques Standard, published by the World Bank, is widely applied in petrochemical plants in Europe and USA. Based on this standard, we use the method of event tree modeling to develop a petrochemical hazard assessing software. This software package consists of six main modules which cover the whole assessing process such as selecting the failure components, event tree analyzing, model algorithm bases, etc. In the core of this software, the module of event tree analyzing, four different event tree models are built for the four main hazard resources in the petrochemical industry. These four include models of Flammable Gas Event Tree, Toxic Gas Event Tree, Flammable Liquid Event Tree and Toxic Liquid Event Tree. Six different model algorithm sub-bases in the model algorithm base module are presented. They are Outflow model-base, After Release model-base, Dispersion model-base, Fires model-base, Explosion model-base and Effects of Toxic Release model-base. Furthermore, we also discuss the critical technologies related to modeling such as modular modeling, assessing algorithm, and uncertainty analysis etc.

SDG Fault Diagnosis Simulation System for Petrochemical Process

The inference based on SDG is a self-contained method to effectively diagnosis system failures, and it has being a research focus in safety field. Since systemic simulation technology has played a significant role in fault diagnosis, creating a trial system based on simulation will be a good foundation to carry out related research. This article shows hardware and software structure of such system that we have developed for the aim, and deeply discusses key technology involved. The system can be used to assist engineers to create SDG diagnosis model, modify threshold value and so on. Besides, we put it through systematic integrated test according to global standards, including software and hardware reliability.

The Interrelationships between SDG and Process Hazard Analysis

The process hazard analysis (PHA) shall address: the hazards of the process, the identification of any previous incident, the engineering and administrative controls applicable to the hazards, the effect caused by failure of engineering and administrative controls, the status of facility, and the human factors. The methodologies of WhatIf, Checklist, Hazard and Operability Study (HAZOP), Failure Mode and Effects Analysis (FMEA) or Fault Tree Analysis (FTA) are recommended. Most of these methods are based on the same fundamentality as the SDG (Signed Directed Graphs). The SDG is one of the most critical computational technologies in process hazard analysis.

High Efficiency Simulation Software Platform for Process System with Misoperation Autodiagnosis

By using of signed directed graph (SDG) deep knowledge model and inverse direction inference technique, the problem of on-line misoperation autodiagnosis during computer simulation training can be solved effectively. The simulation case studies show that the SDG diagnosis method has good completeness, fine resolution and detailed explanation facility of abnormity caused by misoperation and it’s propagation in process. Based on the object oriented programming technique, the message driven software structure, the “three elements” classification of process operation and full running time data compression technology, the new developed simulation software platform has the features of high efficiency, low memory space occupation, high running speed and easy application software development. Up to now, a lot of students in 45 universities have conducted training on this simulation software. The training results are very successful.

Architectural Design of Hazard, Safety and Control Simulation Platform for Petrochemical Process

A Hazard, Safety and Control Simulation Platform (HSC-SP) is described for Petrochemical Process. HSC-SP is an important tool in hazard and operability study (HAZOP) for petrochemical process. The platform is based on HIL (Hardware In the Loop) simulation technology. After analyzing the function and target of HSC-SP, the architecture of this platform and the functions of each sub-system are presented. A new approach for a hybrid simulator using HIL simulator in combination with a complex fully digital simulator is developed. By using this hybrid simulator, a petrochemical process can be studied. Some key technologies are also given, such as communication system, software integration, and real-time database, etc.

Research and Development of Fault-Tree Synthesis Software Based on SDG Qualitative Simulation

FTA (Fault Tree Analysis) is one of the useful methods to analyze the reliability of the system. It’s very important to conduct fault-tree auto-synthesis for the complex systems. From the method of qualitative simulation based on the signed directed graph (SDG), we have developed a software package ViSFATS (Visual SDG-based Ault-Tree Synthesis). This software package consists of three main modules which cover the whole process of fault-tree auto-synthesis such as Graphical User Interface (GUI), information-processing center and inference engine. In the design of the core of this software inference engine, two improvements on the conventional Lapp&Powers algorithm are made. Firstly, a complete and quick qualitative inference mechanism based on inverse-direction search index and modular operator configuration is presented. Secondly, a novel and highly-efficient topological-matrix algorithm which is used to search all NFBLs (negative feedback loops) and FFLs (feedforward loops) is also presented. In the end of this paper, the applicability and efficiency of this ViSFATS has been demonstrated with a case study of fault-tree auto-synthesis for a nitric-acid cooler.

Computer Aided HAZOP Analysis Based on SDG

Hazard and Operability (HAZOP) analysis is widely used and recognized as the preferred safety analysis approach in the process systems. However, the manual HAZOP analysis is time consuming, strenuous and costly. A computer aided HAZOP analysis system based on SDG is introduced. SDG qualitative model is used in this system. As process model is built by unit models in the model base, the system is very adaptive. Besides, it can handle the huge amount of specific process information. Compared with the manual HAZOP analysis, the system has many advantages, such as good maturity, time saving, effort saving, money saving, systematic result, and so on.

PetriNet-PaTNet Method for Maintenance Task Simulation

For effectively conducting maintenance task simulation, a research has been made on both the description of maintenance tasks and the simulation by means of animation technique. Based on the human modeling software package-Jack, a method titled PetriNet—PaTNet is presented to help simulate the maintenance process. The description of input information of task procedure, the object model of scenes and the framework for implementing the simulation are studied. Finally, this paper concludes with a case study, which validates the effectiveness of the PetriNet—PaTNet method.

Structure-Oriented Linear Graph for Rigid Multibody Dynamical System           

From the discrete viewpoint of mass, dynamical energy and its transmission information, the linear graphic methodology for rigid multibody system dynamical model was studied. The rigid-bodies and their connections are abstracted to the objects modularized in implementations and interfaces. When they are assembled together according to the topological isomorphism of system graph, the corresponding interface’s parameters are complied with Network Law. When disturbance happens to system, the energy flows are changed in some connection’s edges. System design, therefore, can be oriented towards Minimum Resistance tendency, so control strategy can be made pertinently.

Research on Methods of Extrinsic Evolvable Hardware           

Extrinsic EHW is a kind of evolvable hardware method that binds hardware description language to evolutionary computation. This paper gives a brief introduction to the basic concept and principle of EHW, points out some issues experimentally against the methods of extrinsic evolvable hardware, and addresses EHW to the core that must be solved for implementing extrinsic EHW on platform of EDA.

Discussion and Formal Abstract Description of Conceptual Model of Mission Space

Conceptual Model of the Mission Space (CMMS) is very important to M&S interoperation and reuse. But there is not a consensus description. This paper researches on the application of CMMS in simulation; and discusses the relationships between CMMS, military doctrine and scenario. Then we present formal abstraction of the conceptual model and at last a model structure is provided as well.

The Research of the Method to Model Validation on Complicated Simulation System..

The research on the theory and methods of assessing confidence for complicated simulation systems is a difficult problem in simulation systems. According to the characteristics of complicated simulation system, an efficient approach to model validation for complicated simulation systems by Bayes theory has been put forward in this paper. The test rules are derived and algorithm is given. Some actual examples have proved that this method is accurate and efficient. The method of complicated hypothesis test with Bayes used in this paper is very valuable to simulation systems analysis under the circumstances of small, even extra-small samples.

Application of FNN in the Recognition of Force of Foot about Athlete

In this paper, a recognition method of force of foot about athlete is introduced. This method is based on the theory of FMMNN. This method achieves a good performance when it is applied to identify actions of weight lifting. This work lays foundation of our digital sportsman human model and simulation.

DNA Computing in Analysis of Cases           

DNA computing is a new method of simulating molecular biology structure of DNA by means of molecular biology technological. This method suggests a new way of solving an NP-complete problem. Logic calculus of Discrete Mathematics has been applied to analysis of cases, in particular, on information handling. In the paper, we have solved a problem of analysis of cases with fluorescence labeling technique based on surface, which is an attempt to apply DNA computing to analysis of cases. Our method has some significant advantages such as simple encoding, low cost, and short operating time.

Ant Colony Algorithm with Characteristics of Sensation and Consciousness

A new ant colony optimization algorithm which simulates the ants’ behavior according to the laws of sensation and consciousness is presented in order to make balance between accelerating convergence and averting precocity as well as stagnation. The algorithm selects path by the mixed influence of consciousness and subconsciousness, and updates the trail information of each path adaptively in each step. The results of simulation on several symmetric and asymmetric TSPs with different sizes indicate that the algorithm has much higher convergence speed and stability than that of classical ant colony algorithm, and is more suitable for solving large scale TSP.

A Spiral Navigation Algorithm for Robotic Exploration of Unknown Environment         

In many cases, robots have no prior knowledge of the environment, it is necessary that the robot need to explore the environment. Navigation algorithms for robots to map the environment completely in short time play a very important role in the robotic task. A spiral navigation algorithm based on virtual force is proposed in this paper to explore the unknown environment using rang sensors. Collisions between robot and obstacle or between robots can be avoided with the application of the proposed navigation rules. The kinematics and dynamics equations of the robot adopting the algorithm are also given. The simulation experiments demonstrate the operation of the algorithm. Several simulation experiments of various representative robotic tasks are carried out, based on the explorative navigation algorithm, which successfully validate the virtual force based spiral navigation algorithm.

The Evolving Artificial Neural Network Based on Genetic Algorithm

This paper presents a genetic multiplayer feedforward neural network in order to evolve neural network architectures and connection weights simultaneously. In the method, the learning efficiency and the convergence are improved greatly by bi-population weights learning and self-adaptive mutate ratio of structure learning. The simulation results show that the premature convergence in genetic algorithm is restrained effectively and the learning efficiency and the convergent precision for the weights of the multi-layer forward neural networks are improved greatly. These results also show that the proposed method in this paper can produce very compact artificial neural networks in comparison with other algorithms.

Detection 3-D Structure Objects from Aerial Image with Bayesian Networks

Perceptual organization is a powerful tool for automatic detection objects. In this paper, a perceptual organization approach based on Bayesian Networks is proposed, and used as a method of detection 3-D man made structure objects, such as buildings et al., from urban aerial image. Objects are detected by means of hypotheses verification, and the approach is divided into four steps: edges and lines segments extraction, parallelograms creation, hypotheses creation, and hypotheses verification. Two Bayesian Networks are designed for information integration in the process of hypotheses creation, and hypotheses verification respectively. With the learning function from Bayesian Networks, the approach is provided with self-adaptation to some extent. The validity of this approach is indicated by experiment.

A Novel Approach For Word Clustering Based On Large Tagged Corpus                 

This paper proposes a novel approach for word clustering based on large tagged corpus. According to the need of decision-making support for a specific problem, this paper review several algorithms developed by previous works, after that, our algorithm is rendered. Firstly, we manually extract several words from a specified class, and then search the corpus for the modifiers of those words to construct modifier vector, for each of other words, count the frequency of its co-occurrence with each modifier in the modifier vector to construct its characteristic vector, finally, apply clustering algorithm to those characteristic vectors to get the result. Proved by experiment carried out on Decision-making Support for Macro Economics, this algorithm is effective for word clustering.

Design and Application of TD Algorithm Based on BP Net

Whereas the standard BP algorithm can not change weight dynamically, Temporal Difference (TD) method is introduced. This paper applies a learning algorithm based on neural network to predict time sequence. We deduct the algorithm which combines TD method with BP net and give a method to perform the algorithm by C++. Through a simulation in which we build a model for signal of time sequence, we illuminate the feasibility of this algorithm.

The Virtual Simulation of Sensitivity Measurement about Initiating Explosive Device in Environment of Web           

This paper produces the Virtual Simulation System of Sensitivity Measurement about Initiating Explosive Device with the implement of Virtual Reality Modeling Language (VRML) being used in WWW, which can do interactive operation by using IE browser and Netscape browser in the Internet, and the key technology to realize this system is also discussed. The system is the implement of virtual reality in the World Wide Web with high real effect and it makes operators feel as if personally on the scene.

Technique and Method of the Scene’s Modeling of Virtual City

The scene’s modeling is an important step to design Virtual City, which includes 3D models of terrain and surface features building, textures mapping and special effect designing. In order to construct terrain model, it is necessary to divide the surface into small parts, calculate the normal line and define color for each part; In order to make surface features’ model, it is required to divide the surface into 2D triangular pieces, construct transparent and opaque face. The texture mapping is used to improve the scene of Virtual City by joining the photos with terrain and surface features; special effect designing is used to make the special sight effect for Virtual City by the techniques such as elimination, illumination and perspective. The technique of LOD (level of detail) can be used to improve the running speed of the complicated scene’s model of Virtual City, but certain compromise between detail information and speed need to be made by adopting the technique of multilevel LOD and by using different LOD at different time while drawing the scene.

Simulation of Decoying to the ARM in the Simulative Fire Control Radar Training System.           

Compute-aid training is an important form of military simulation, with which the operation training of many weapons can be fulfilled safely and efficiently at a low cost. The anti-radiation missile is the archenemy of radar in the current battlefield. Consequently, knowing its feature well and finding a special countermeasure are in urgent need. Based on the Simulative Two-in-One Fire Control Radar Training System, this paper has investigated the simulation of decoying to the anti-radiation missile in ARM subsystem. The mathematic model has been brought forward, besides, the simulation result of representative dual sources decoy has also been presented. The effect of the dual sources’ station has been analysed at last.

Research on Modeling and Simulation of Aircraft Electric Braking System..

This paper investigates the composition and the principle of electric brake of aircraft, sets up the model structure of it, and design the controller by using fuzzy control. It is simulated in the MATLAB/SIMULINK and the result of wet runway of simulation is also presented. The result shows that the system model established in the paper is basically correct and electric brake has many advantages over hydraulic brake in actuating response, braking efficiency and etc.

A Simulative Model for Collision Verification between Missile and its Target

We propose a simulative model for missile attacks and establish an effective calculating method, by which the collision between a single missile and its target can be verified, and all the judging conditions required by the collision verification are given in this paper. This method is applicable to moving targets with any complicated shapes and can be used in all the arm attack simulations.

Emulational Analysis on the Performance of Underwater Bearing-only Passive Target Tracking Using Two Arrays

The tracking performance of data fusion based on underwater bearing-only measurements of two arrays is studied in this paper. The simulation results show that the target motion analysis method with two-dimension bearing-only information have not only eliminated maneuvers needed by conventional bearing-only TMA based on single array, but also improved the stabilization and global convergence for estimation algorithm. The movement factors of target can be fast and steadily estimated under the usual situation of towarding the target using extended Kalman filter. At the same time, the influence of the geometrical situation of our ship to the target on the tracking performance is shown and references to engineering application are provided.

Analysis and Design on Performance Flight-test Assessment System of Fighter’s Fire Control System

Performance flight-test of fire control system is an important method to assess the performance of fighter’s fire control system. As an example, the unitary design frame, the detailed design and implementation of a certain fighter are given, and the software and hardware about the flight-test assess system are also developed. It has been applied to a certain airplane and has been proved to be highly effective.

Research on Accuracy of Localization Algorithm for Bistatic Sonar

Based on the character of bistatic sonar, the bistatic time-only-localization (TOL) algorithm is proposed in this paper. Through simulation, the localization principle and the error of the algorithm is given. The simulating results show that the location accuracy of TOL algorithm is improved when the length of baseline is long. The time measure error has a little effect on the localization accuracy. The location accuracy of bistatic sonar is higher than that of monostatic. The research presents theory basement for establishing the new multistatic sonar system in the future.

Modeling and Simulation for High-Speed Switching Fabrics   

High-speed switching fabric is the core for high-speed router/switch, which is responsible for forwarding cells after table lookup. Switching fabric is one of the important factors to limit router/switch’s capacity. In this paper, we try to model and simulate switching fabrics. Firstly, we propose a general switching fabric model, and describe its components. Then, we stress to discuss one kind of switching fabrics ---- input-queued crossbar, and evaluate its performance measured by throughput and time delay. In a result, input-queued crossbars are suitable for building high-speed, large capacity routers/switches.

Simulating Studies on Hydrocarbon Migration

In this paper a skeleton-based implicit surface modeling technique is used to construct a simplified porous space model. A numerical simulation platform has been developed and it is proved that Buoyancy Percolation model applies to simulate the process of migration of oil in the porous media saturated with water.

Simulation on the Performance of Plate Evaporator           

The steady distributed parameter model is established in this paper to simulate the plate heat exchanger used as an evaporator. Based on the model, many parameters can be respectively calculated such as the temperature and pressure of refrigerant and water, the velocities of the two fluids and the temperature of the plate along the plate length. The numerical analysis can be carried out about the exchanger performance, and the theories concerned are expounded. The heat exchanging and the fluids flowing in the exchanger are explained well. From these, it is possible to improve the performance and the design of the exchanger.

Simulation Research on the Process of Missile Attacking and Operating Process of Surface Ship    

Anti-ship missile is the primary weapon used in surface ship combat, and it’s very important to simulate the process. Inspired by Lanchester’s Equations, this article makes simulated research on the missile attacking and operating process of surface ship by adopting discrete module after analyzing combat efficiency, uses differential equation group to analyze operating strength, damage coefficient and combat efficiency. After simulating three practical problems, it gains conclusion as follows: combat method, combat capability of missile and defend capability of ship play important role on the result of surface ship combat.

The Real-time Simulation of The Auditorium Acoustics Based on DSP    

The auditorium acoustics, which is the up-to-date research subject in architectural acoustics, is about the acoustical qualities in such halls as concert, cinema, opera, gymnasium, etc. And it’s especially significative for guiding the building design of the above halls by simulating their audible effects. Although there’re many evaluating parameters or indices of acoustic qualities, the final conclusion is determined by man’s subjective criterion with audible feeling or description. So a method is proposed in some way to simulate the auditorium acoustics based on DSP by transit “dry sound” to digital singal followed by convolution operation on the impulse response from the given model and then return to analog singal with A/D or D/A converter. The whole process can be real-time and audible through the ear-phones finally. And that will be helpful to study the auditorium acoustics and its design of good quality.

Control Method for Eliminating Pressure Pulsating of Aircraft Refueling Pipeline..

Aimed at the pressure transient of the aircraft refueling pipeline, a new real time flow control method based on the traditional passive accumulator is presented in this paper. The dynamic characteristic equations of the inner parameters of the transient suppressor are established, and the controllable vectors are also extracted.Then with the help of feed-forward and feed-back technology,the optimal control scheme is determined. At the same time the effects of system’s uncertainties are considered to evaluate the robustness. At last the method mentioned above is utilized to eliminate the very high transient pressure on line. The results of theoretic analysis and simulation show that it is correct and valid.

A New Fuzzy PID Controller....

In this paper, PID parameters are first normalized on the basis of engineering modification, and then are adjusted by fuzzy control principle and fuzzy inference. Thus a new fuzzy PID controller is proposed. The PID controller has the ability of adapting to such complexity and hard control objects, as non-lined Time-variant uncertainty and so on. In the end this algorithm is applied in the simulation research of using over-hot air temperature in a power plant as the subject. The result shows the algorithm is of reliability, high precision and strong ability of anti-interfere and robustness.

Soft Sensor Modeling Based on Support Vector Machines

Soft sensor has been widely used in industrial process control. It makes an important role to improve the quality of product and assure safety in production. The core problem of soft sensor is to construct appropriate mathematic model. Support vector machine (SVM) is a novel machine learning method, which is powerful for the problem characterized by small sample, nonlinearity, high dimension and local minima, and has high generalization. In this paper, soft sensor modeling method based on Least Square SVM (LS SVM) is proposed, and cross validation method is used to select hyper-parameter of LS SVM model. Soft sensor model based on LS SVM is applied to predication of frozen point of light diesel oil. Effective result indicates that LS SVM is of potential application in soft sensor.

Information Management of Virtual Prototype in General Design Phase  

The paper gives the information flow chart of general design, through studying on general design of complex product virtual prototype engineering system. It also presents the idea that Agent technology can manage the information of general design, so as to realize intelligent general design for designing complex product quickly and effectively.