18CSE483T - INTELLIGENT MACHINING UNIT 4 & 5 - 12M

12M:

Explain in detail the Neural Network model in intelligent machining process.

Neural network modeling:

• Neural networks are simple models of the way the nervous system operates

• The basic units are neurons, which are typically organized into layers

Neural network:

• A neural network is a simplified model of the way the human brain processes information

• It works by stimulating a large number of interconnected processing units 

• The processing units are arranged in layers

• There are typically three parts in a neural network:

• An input layer, with units representing the input fields

• One or more hidden layers

• An output layer, with a unit or units representing the target fields

• The units are connected with varying connection strengths or weights

• Input data are presented to the first layer and values are propagated from each neuron to every neuron in the next layer

• Eventually, a result is delivered from the output layer

• The network learns by examining individual records, generating a prediction for each record and making adjustments to the weights whenever it makes an incorrect prediction

• This process is repeated many times, and the network continues to improve its predictions until one or more of the stopping criteria have been met.

• Initially, all weights are random, and the answers that come out of the net are probably nonsensical. 

• The network learns through training. 

• As training progresses, the network becomes increasingly accurate in replicating the known outcomes. 

• Once trained, the network can be applied to future cases where the outcome is unknown.

Neural network in machining process:

• Usage: 

• Prediction of cutting forces

• Surface roughness

• Dimensional deviation

• Tool life

• Input neurons correspond to feed(f), cutting speed(v), depth of cut & vibration, acceleration of tool

• Output neurons correspond to surface roughness (Ra)

Most common neural network architectures:

• Feedforward neural networks

• Feedback neural networks

• Self-organizing neural networks

Explain in detail about objective functions and optimization techniques

IM computation methods:

• Optimization using genetic algorithms

• Control of process using fuzzy logic

Objectives of adaptive control system:

• To adjust machining process

• Maximizing performance criteria

• Automatically improve the performance

Types of adaptive control:

• Adaptive control optimization

• To search optimal values of feed rate and spindle speed

• Adaptive control constraint

• To select feasible solution

• Geometric adaptive control

• To adjust tool against material used and temperature to obtain accurate surface

• Vibration adaptive control

• To control vibration of tools

Machining optimization:

Objective in machining problem:

• Minimization of cost of machining

• Maximization of production rate

• Maximization of profit rate

Objective functions:

Constraints: 

• Constraint on tool life

• Low -> production affected

• High -> tools become under utilized

• Constraints of surface finish

• Low surface roughness not recommended 

• Surface roughness affects heat transfer rate

• Constraint on machining process

• Constraint on vibration

• Constraint on dimensional duration

• Constraint on geometric relations

Optimization technology:

• Golden section search method

• Sequential quadratic programming

• genetic algorithm

Golden search method:

• To find minimum of unimodal function fn: which has only one min in certain interval

General procedure of region elimination method:

Genetic algorithm: (STEPS)

• Select encoding type

• Choose population size

• Randomly choose initial population

• Select parental chromosomes

• Crossover and mutation

• Evaluation of offspring

• If stopping criteria not reached, go to step 4

Discuss in detail

a)Fuzzy Inference system

Fuzzy inference:

• The prediction of surface roughness steps:

• Fuzzification

• Rule evaluation step

• Rule aggregation

• Defuzzification 

Fuzzification:

In this step, the crisp input values of feed and cutting speed are transformed into fuzzy values. These values are mapped to fuzzy sets with membership functions. For example:

• Feed: Low, High

• Cutting Speed: Low, High

Rule Evaluation

In this step, fuzzy rules are applied to the fuzzified inputs. The rules are typically in the form of IF-THEN statements. For example:

• Rule 1: IF feed is Low AND cutting speed is Low, THEN surface roughness is Medium.

• Rule 2: IF feed is Low AND cutting speed is High, THEN surface roughness is Low.

• Rule 3: IF feed is High AND cutting speed is Low, THEN surface roughness is High.

• Rule 4: IF feed is High AND cutting speed is High, THEN surface roughness is Medium.

These rules are used to evaluate the conditions and produce outputs for each rule.

Rule Aggregation

• In this step, the outputs from all the rules are combined to form a single fuzzy set. 

• The aggregation process involves taking the maximum membership value from the outputs of each rule. 

• This results in a single fuzzy set that represents the combined output of all the rules.

For instance, if multiple rules suggest different surface roughness levels, the aggregation step combines these suggestions into a single fuzzy set that encapsulates all possible outputs.

Defuzzification:

This final step converts the aggregated fuzzy output set back into a single crisp value. The most common defuzzification methods include:

• Centroid Method (Center of Gravity): This method calculates the center of the area under the fuzzy set curve and uses it as the crisp output.

• Mean of Maximum (MOM): This method takes the average of the maximum values of the fuzzy set.

• Bisector of Area (BOA): This method finds a vertical line that splits the area under the fuzzy set into two equal parts.

For example, if the aggregated fuzzy set indicates surface roughness levels across a range of values, the centroid method would compute the centroid of this set and provide a single crisp value representing the predicted surface roughness.

b) Role of Adaptive control in intelligent machine

Adaptive control:

• An adaptive control mechanism is an integral part of an intelligent machine.

• An adaptively controlled machine is able to adapt to the dynamic changes of the system caused by the variability of the machining process due to changes in the cutting conditions such as the hardness of the work material, tool wear, deflection of the tool and the workpiece, and so on. 

Objectives of adaptive control system:

• To adjust machining process

• Maximizing performance criteria

• Automatically improve the performance

Types of adaptive control:

• Adaptive control optimization

• To search optimal values of feed rate and spindle speed

• Adaptive control constraint

• To select feasible solution

• Geometric adaptive control

• To adjust tool against material used and temperature to obtain accurate surface

• Vibration adaptive control

• To control vibration of tools

FOR THE CASE STUDIES, PLZ REFER THE PPTS IN THE FOLDER

18CSE483T Intelligent Machining Case Study