18CSE483T - INTELLIGENT MACHINING UNIT 2 & 3 - 12M
12M:
Open architectures
Open architecture for machining process:
Based on open standards
Easy to modify and to exchange data between different types of software application
Supported by a large community of developers
Eg: combination of TCP/IP and HTTP
Machine tool open system architecture for intelligent control (MOSAIC):
Developed at New York University
To support information flow required task
Based on memory mapped backplane architecture (VME)
Is an agent-based machining system architecture
Is a hardware running on POSIX OS
Principles:
Open architecture
Common OS at all levels of machines
Standardized object oriented knowledge exchange
Next Generation Machine/ Workstation Controller (NGC):
Was initiated by the US government for restructuring machine tool industry by enabling Hardware, software, process control technologies
NGC provides interface for scalability and third party product development
Software interface built based on [NML] Neutral Manufacturing Knowledge process control
Requirements addressed by NGC:
Hardware, software architecture
Network capabilities
Task planning
Process control
Fault management
Sensor interface
Manufacturing Automation Protocol (MAP):
Was developed by general Motors to meet its manufacturing integration needs
Supports application layer protocols
Uses the ISO reference model
FTAM - File Transfer and Access Method
CLNS - Connector Less Network
ESIS - End System Intermediate System
ACSE - Association Control Service Element
MMS - Manufacturing Message Specification
Open Systems Architecture for Control within Automation Systems (OSACA):
European initiative to define a vendor-neutral, open controller architecture
OSACA’s approach is to develop an API for control applications and an appropriate infrastructure to achieve interoperability, scalability and reusability of applications
Three main areas:
Communication system: to define a hardware and system software independent interface to exchange information between different application modules of a controller
Reference architecture: to determine the functional units of a controller such as Motion Control and Logic Control and specifies the external interfaces of them
Configuration system: to enable dynamic configuration of a controller by combining different application modules at boot-up time
Explain in detail the different methodologies for transforming data to information
Data acquisition is the process of sampling signals and converting the resulting samples into digital numeric values that can be manipulated by a computer
The analog-to-digital conversion (ADC) process converst the input continuous physical quantity from a sensor to a digital number that represents the quantity’s amplitude
Techniques:
Time domain analysis
Frequency domain analysis
Time-frequency domain analysis
Time domain analysis:
Format with the vertical axis as amplitude or voltage and the horizontal axis as time
This method analyzes a sequence of data points, measured typically at successive time instants with uniform time intervals, in order to extract meaningful statistics and other characteristics of the data
3 main time series analysis techniques:
Autoregressive (AR)
Moving average (MA)
Autoregressive-moving average (ARMA)
Frequency domain analysis:
It can provide information with regard to signal components in certain frequency range and is regarded as the most useful method for signal processing
In many machining operations with a rotating spindle, the frequency spectrum of measured signals, such as vibration or cutting forces, carries a great deal of information that can be used to monitor and diagnose the process and tool condition
An inverse DFT can transform the resultant frequency spectrum back into the time domain
Time-frequency domain analysis:
Although FFT based methods are widely used, one main limitation is for the methods is that they are not suitable for nonstationary signals
A nonstationary signal means that its frequency changes over time
FFT is unable to detect the frequency changes
To address this issue, a number of time-frequency domain analysis techniques have been developed, including
Short-time fourier transform (STFT)
Wigner-Ville distribution
Wavelet transform
Explain in detail about techniques of knowledge representation in AI
Knowledge representation:
Describes the representation of knowledge
It is a study of how the beliefs, intentions and judgements of an intelligent agent can be expressed suitably for automated reasoning
Techniques of knowledge representation in AI:
Logical representation:
Represents conclusion based on various conditions
Each sentence translated to logic using syntax and semantics
Syntax - determine which symbol use in KR
Semantics - rules to interpret sentence in logic; assign meaning to sentence
Advantage: helps to perform logic reasoning
Disadvantage: may not be very natural, have some restrictions and are challenging to work with
Semantic network representation:
Representation in the form of graphical network
Objects as nodes and arcs to give relation between objects
Consists of 2 types of relations:
IS-A relation
Kind-of relation
Advantage: simple and easy to understand
Disadvantage: take more computational time at runtime
Frame representation:
A frame is a record like structure that consists of a collection of attributes and values to describe an entity in the world
Basically it consists of a collection of slots and slot values of any type and size
Slots have names and values which are called facets
Advantage: easy to understand, visualize and add slots
Disadvantage: inference mechanism cannot be smoothly proceeded
Production rules:
In production rules, agent checks for the condition and if the condition exists the production rules fires and corresponding action is carried out
The system consists of 3 main parts:
Set of production rules
Working memory
Recognize act cycle
Advantage: highly modular and can be easily removed or modified
Disadvantage: does not exhibit any learning capabilities and is inefficient
Discuss in detail about tools and techniques for conceptual design
Conceptual design is the part of the design process where
By identifying the essential problems, searching for appropriate working principles and combining these into a working structure
The basic solution path is laid down through the elaboration of a solution principle
conceptual design specifies the principle solution
AI and design:
AI is concerned with the application of knowledge
Within AI, 3 main directions of reasoning
Reasoning by logic
Reasoning by learning
Reasoning by analogy
Reasoning by logic:
Expert systems apple rule-based reasoning (RBR) technique, a form of reasoning by logic
An expert system is useful when the domain knowledge can be formalized into simple rules and when common sense does not play an important role
Reasoning by learning:
Reasoning by learning can be implemented with artificial neural networks
An ANN consists of nodes connected vis adjustable weights
Reasoning by analogy:
The third form of reasoning, reasoning by analogy, is best exemplified by Case-Based Reasoning (CBR)
Cases are stored in a case-base to create a reservoir of problem-solution combinations
AIDA system:
AI- supported design of Aircraft
Built of three modules and a central interface
Case-based reasoning module
Functional module
Geometrical module
Central user interface
Comments
Post a Comment