On-line monitoring of boring tools for control of boring operations

In order to carry out precision and quality control of boring operations, on-line monitoring of boring tools is essential. Fourteen features were extracted by processing cutting force signals using virtual instrumentation. A Sequential Forward Search (SFS) algorithm was employed to select the best combination of features. Backpropagation neural networks (BPNs) and adaptive neuro-fuzzy inference systems (ANFIS) were used for on-line classification and measurement of tool wear. The input vectors consist of selected features. For the on-line classification, the outputs are boring tool conditions, which are either usable or worn out. For the on-line measurement, the outputs are estimated value of the tool wear. Using BPN, five features were needed for the on-line classification of boring tools. They are the average longitudinal force, average value of the ratio between the tangential and radial forces, skewness of the longitudinal force, skewness of the tangential force, and kurtosis of the longitudinal force. Three features, the average longitudinal force, average of the ratio between the tangential and radial forces, and kurtosis of the longitudinal force, were needed for on-line measurement of tool wear. Using ANFIS, three features were needed for the on-line classification of boring tools. They are the average longitudinal force, average of the ratio between the tangential and radial forces, and kurtosis of the longitudinal force. Only one feature, kurtosis of the longitudinal force, was needed for the on-line measurement of tool wear using ANFIS. Both 5×20×1 BPN and 3×5 ANFIS can achieve a 100% success rate for the on-line classification of boring tool conditions. Using a 3×20×1 BPN for neural computing, the minimum flank wear estimation error is 0.29% while the minimum flank wear estimation error is 2.04% using a 1×5 ANFIS.     

Investigation and FEA-based simulation of tool wear geometry and metal oxide effect on cutting process variables

In machining, it is clearly noticed that the cutting tool wear influences the cutting process. However, it is difficult with experimental methods to study the effects of tool wear on several machining variables. Thus, in the literature, some earlier studies are performed separately on the effect of tool flank wear and crater wear on cutting process variables (such as cutting forces and temperature). Furthermore when the workpiece material adheres in cutting tool, it affects considerably the heat transfer phenomena. Accordingly, in this work the finite element analysis (FEA) is performed to investigate the influence of combination of tool flank and crater wear on the local or global variables such as cutting forces, tool temperature, chip formation on the one hand and the effects of the oxidized adhesion layer considered as oxide (Fe₂O₃/Fe₃O₄/FeO) on the heat transfer in cutting insert on the other hand. In this investigation, an uncoated cutting insert WC–6Co and medium carbon steel grade AISI 1045 are used. The factorial experimental design technique with three parameters (cutting speed Vc, flank wear land VB, crater wear depth KT) is used for the first investigation without adhesion layer. Then, only linear investigation is performed. The analysis has shown the influence of the different configurations of the tool wear geometry on the local or global cutting process variables, mainly on temperature and cutting. The simulation’s results show also, the highly influence of the oxidized adhesion layer (oxide Fe₂O₃/Fe₃O₄/FeO) on the heat transfer.     

Aspects of burr formation in bandsaw teeth manufactured by milling operation

Steel manufacturers and stockholders prefer bandsawing for cutting off raw materials compared to other techniques as it enjoys competitive advantages of higher accuracy of cut, better surface finish, lower kerf loss, better straightness of cut, long tool life and high metal removal rate. Along with the geometries of the bandsaw tooth, bandsaw cutting edge condition (e.g., edge sharpness and burr) significantly affects the cutting performance of a bandsaw. Currently the production of bandsaw is largely done by milling operation due to the scale of manufacturing and the economics of milling compared to other processes (e.g., grinding). Ideally, the bandsaw teeth should possess sharp cutting edges with no burr. In general, two types of burr are commonly seen in the bandsaw teeth manufactured by milling operation namely tooth tip burr and side burr. Current research undertaken at Northumbria University in collaboration with a major bandsaw producer is focused on the mechanism of burr formation in the bandsaw teeth. This paper briefly outlines the factors affecting the burr formation in bimetal (High Speed Steel edge wire and soft steel backing material) bandsaw teeth manufactured by milling process and suggests the necessary steps to be considered for manufacturing burr free bandsaw with sharp cutting edges. The investigation showed that flank wear in the milling cutter has a major influence on the side burr formation in the bandsaw teeth, whereas tooth tip burr was influenced by both flank wear and “V” type notch wear found at the crossover point on the flank face. It was also concluded that TiN coating on the milling cutter could control the burr formation in bandsaw teeth to some degree.     

APPLICATION OF ABDUCTIVE POLYNOMIAL NETWORK AND GREY THEORY TO DRILL FLANK WEAR PREDICTION

An abductive polynomial network for drill flank wear prediction was established, in which grey relational analysis was incorporated to explore the effect of various drilling parameters on flank wear. An abductive polynomial network usually includes multiple layers, each of which contains different polynomial functional nodes. It can automatically synthesize the optimal network structure, including the optimal number of layers and the optimal form of functional nodes. The correlation between the drilling input parameters, including the average thrust force, torque, cutting speed, feed and drill diameter, and drill flank wear can be achieved through this network model.

Based on experimental data, the developed network of this paper attained better accuracy in predicting drill flank wear, given the CPM of 0.1. The findings prove that the network is feasible and accurate in predicting flank wear.

In addition, grey relational analysis was used in this paper to investigate the effect of the aforementioned five drilling parameters on flank wear. According to the analytical results, the most influential factor on flank wear is drill diameter, followed by the average thrust force.     

Tool condition classification using Hidden Markov Model based on fractal analysis of machined surface textures

The texture of a machined surface generated by a cutting tool, with geometrically well-defined cutting edges, carries essential information regarding the extent of tool wear. There is a strong relationship between the degree of wear of the cutting tool and the geometry imparted by the tool on to the workpiece surface. The monitoring of a tool’s condition in production environments can easily be accomplished by analyzing the surface texture and how it is altered by a cutting edge experiencing progressive wear and micro-fractures. This paper discusses our work which involves fractal analysis of the texture of surfaces that have been subjected to machining operations. Two characteristics of the texture, high directionality and self-affinity, are dealt with by extracting the fractal features from images of surfaces machined with tools with different levels of tool wear. The Hidden Markov Model is used to classify the various states of tool wear. In this paper, we show that fractal features are closely related to tool condition and HMM-based analysis provides reliable means of tool condition prediction.     

Flank wear detection of cutting tool inserts in turning operation: application of nonlinear time series analysis

It has been established that turning process on a lathe exhibits low dimensional chaos. This study reports the results of nonlinear time series analysis applied to sensor signals captured real time. The purpose of this chaos analysis is to differentiate three levels of flank wears on cutting tool inserts—fresh, partially worn and fully worn—utilizing the single value index extracted from the reconstructed chaotic attractor; the correlation dimension. The analysis reveals distinguishable dynamics of cutting characterized by different values for the dimension of the attractor when different quality tool inserts are used. This dependence can be effectively utilized as one of the indicators in tool condition monitoring in a lathe. This paper presents the experimental results and shows that tool vibration signals can transmit tool wear conditions reliably.     

Knowledge-based diagnosis of drill conditions

One major bottleneck in the automation of the drilling process by robots in the aerospace industry is drill condition monitoring. This paper describes a system approach to solve this problem through the advancement of new machine design, sensor instrumentation, metal-cutting research, and intelligent software development. All drill failures can be detected and distinguished: chisel edge wear, flank wear, crater wear, margin wear, corner wear, breakage, asymmetry, lip height difference, and chipping at lips. However, in the real manufacturing environment, different workpiece materials, drill size, drill geometry, drill material, cutting speed, feed rate, etc. will change the criteria for judging the drill condition. The knowledge base used for diagnosing the drill failures requires a huge data bank and prior exhaustive testing. A self-learning scheme is therefore introduced to the machine in order to acquire the threshold history needed for automatic diagnosis by using the same new tool under the same drilling conditions.     

Lightweight Data Indexing and Compression in External Memory

In this paper we describe algorithms for computing the Burrows-Wheeler Transform (bwt) and for building (compressed) indexes in external memory. The innovative feature of our algorithms is that they are lightweight in the sense that, for an input of size n, they use only n bits of working space on disk while all previous approaches use Θ(nlog n) bits. This is achieved by building the bwt directly without passing through the construction of the Suffix Array/Tree data structure. Moreover, our algorithms access disk data only via sequential scans, thus they take full advantage of modern disk features that make sequential disk accesses much faster than random accesses. We also present a scan-based algorithm for inverting the bwt that uses Θ(n) bits of working space, and a lightweight internal-memory algorithm for computing the bwt which is the fastest in the literature when the available working space is o(n) bits. Finally, we prove lower bounds on the complexity of computing and inverting the bwt via sequential scans in terms of the classic product: internal-memory space × number of passes over the disk data, showing that our algorithms are within an O(log n) factor of the optimal.     

Adaptive neuro-fuzzy inference system modeling of cryogenically treated AISI M2 HSS turning tool for estimation of flank wear

This study deals with modeling the flank wear of cryogenically treated AISI M2 high speed steel (HSS) tool by means of adaptive neuro-fuzzy inference system (ANFIS) approach. Cryogenic treatment has recently been found to be an innovative technique to improve wear resistance of AISI M2 HSS tools but precise modelling approach which also incorporates the cryogenic soaking temperature to simulate the tool flank wear is still not reported in any open literature. In order to obtain data for developing the ANFIS model, turning of hot rolled annealed steel stock (C-45) by cryogenically treated tools treated at various cryogenic soaking temperatures was performed in steady state conditions while varying the cutting speed and cutting time. The model combined modeling function of fuzzy inference with the learning ability of artificial neural network; and a set of rules has been generated directly from experimental data. It was determined that the predictions usually agreed well with the experimental data with correlation coefficients of 0.994 and mean errors of 2.47%. The proposed model can also be used for estimating tool flank wear on-line but the accuracy of the model depends upon the proper training and selection of data points.     

Evaluation of the performance of backpropagation and radial basis function neural networks in predicting the drill flank wear

This study compares the performance of backpropagation neural network (BPNN) and radial basis function network (RBFN) in predicting the flank wear of high speed steel drill bits for drilling holes on mild steel and copper work pieces. The validation of the methodology is carried out following a series of experiments performed over a wide range of cutting conditions in which the effect of various process parameters, such as drill diameter, feed-rate, spindle speed, etc. on drill wear has been considered. Subsequently, the data, divided suitably into training and testing samples, have been used to effectively train both the backpropagation and radial basis function neural networks, and the individual performance of the two networks is then analyzed. It is observed that the performance of the RBFN fails to match that of the BPNN when the network complexity and the amount of data available are the constraining factors. However, when a simpler training procedure and reduced computational times are required, then RBFN is the preferred choice.     

Computer observer for in-process measurement of lathe tool wear

In the laboratory, tool wear is measured by direct visual observation of flank and crater wear dimensions using a tool maker's microscope. On the shop floor, the journeyman machinist uses chip appearance, sound, vibration, and surface finish to access tool condition. More precise information can be provided by between pass measurements of work piece dimensions. Although there is a body of research directed toward in-process measurement of tool wear, none has found practical application on the shop floor. This, despite the demands of unattended operation of machine tools in the automated factory. Two of the authors have developed a state space model of metal cutting on a lathe. Operation of that model was experimentally accessed using the Advanced Continuous Simulation Language (ACSL) [1]. The state space model embodied in the simulation is comprised of six state variables. In an actual lathe, four of those variables (spindle speed and torque, and cutting speed and force) would be directly measurable. The other two variables, flank and crater tool wear, would not be measurable. This paper describes a linear observer that reconstructs the two tool wear state variables based on system inputs and the four measurable state variables. In actual use the observer would be implemented using a microcomputer dedicated to the lathe. In this study, the previously developed ACSL model was substituted for the lathe, and the linear observer was incorporated as an extension to the simulation. Operation of the observer, including response to initial errors, is demonstrated.     

Condition monitoring of CNC machining using adaptive control

In this work, an adaptive control constraint system has been developed for computer numerical control (CNC) turning based on the feedback control and adaptive control/self-tuning control. In an adaptive controlled system, the signals from the online measurement have to be processed and fed back to the machine tool controller to adjust the cutting parameters so that the machining can be stopped once a certain threshold is crossed. The main focus of the present work is to develop a reliable adaptive control system, and the objective of the control system is to control the cutting parameters and maintain the displacement and tool flank wear under constraint valves for a particular workpiece and tool combination as per ISO standard. Using Matlab Simulink, the digital adaption of the cutting parameters for experiment has confirmed the efficiency of the adaptively controlled condition monitoring system, which is reflected in different machining processes at varying machining conditions. This work describes the state of the art of the adaptive control constraint (ACC) machining systems for turning. AISI4140 steel of 150 BHN hardness is used as the workpiece material, and carbide inserts are used as cutting tool material throughout the experiment. With the developed approach, it is possible to predict the tool condition pretty accurately, if the feed and surface roughness are measured at identical conditions. As part of the present research work, the relationship between displacement due to vibration, cutting force, flank wear, and surface roughness has been examined.     

In-process tool condition forecasting based on a deep learning method

It is widely acknowledged that machining precision and surface integrity are greatly affected by cutting tool conditions. In order to enable early cutting tool replacement and proactive actions, tool wear conditions should be estimated in advance and updated in real-time. In this work, an approach to in-process tool condition forecasting is proposed based on a deep learning method. A long short-term memory network is designed to forecast multiple flank wear values based on historical data. A residual convolutional neural network is built to enable in-process tool condition monitoring, using raw signals acquired during the machining process. The integration of them enables in-process tool condition forecasting. Median-based correction and mean-based correction are adopted to improve the accuracy. IEEE PHM 2010 challenge data has been used to illustrate and validate this approach. Experimental study and quantitative comparisons showed that future flank wear values could be precisely forecasted during the machining process. The proposed approach contributes to prompt and reliable cutting tool condition forecasting, which will support the decision-making about cutting tool replacement in data-driven smart manufacturing.     

Advanced process monitoring—a major step towards adaptive control

Adaptive Control (AC) of machine tools requires many kinds of measured input data. The more information about the complex metal cutting process that can be obtained, the better the process can be controlled.

The paper describes an Adaptive Control Optimization (ACO) system for turning operations. The system continuously chooses Optimal Cutting Data (OCD), taking into account both economical criteria and technical limitations.

The system operates at three different levels:

• • Advanced Process Monitoring

• • Adaptive Control Constraint (ACC)

• • Adaptive Control Optimization (ACO).

Two commercial monitoring systems perform process monitoring. In addition, five independent measurement systems have been developed.

A dedicated vision system has been installed in the lathe to measure the tool flank wear between cuts. The flank wear data are utilized to predict the tool life. Based upon these predictions economical optimum cutting data can be calculated at the ACO level.

To obtain in-process real-time control of the metal cutting process the cutting forces are measured during machining. The forces are measured with conventional piezoelectric force transducers which are located between the turret housing and the cross-slide. The measured force signals are processed by a dedicated microcontroller at the ACC level and cutting data adjustments are fed back to the machine control.

A vibration measurement system, which either can be connected to an accelerometer or use the dynamic force signal from the piezoelectric force transducer, is part of a vibration control module at the ACC level. An ultra-fast signal processor performs the signal analysis.

The remaining two measurement systems—a high frequency tool signal analysis system and a power spectra analysis system—are mentioned in the paper but not further discussed.

Finally, the paper deals with how the strategies at the three different levels will be combined, in order to form an AC system. The monitoring tasks will always reside in the background and be activated if any failure occurs. The ACO subsystem will act as a path-finder and suggest cutting data. The active control tasks will, however, be carried out at the ACC level.     

Dynamic speed control of a machine tool with stochastic tool life: analysis and simulation

We discuss a tool management model for a flexible machine equipped with a tool magazine, variable cutting speed, and sensors to monitor tool wear, when tool life due to flank wear is stochastic. The objective is to adjust the cutting speed as a function of remaining distance, each time a tool change occurs, in order to minimize the expected processing time (sum of cutting and tool setup time). We address the computational aspects of finding optimal decision rules and we present numerical results suggesting that easily computed decision rules of a simple static model are near-optimal for our dynamic programming model. Dynamic adjustment is assessed with simulation experiments.     

Comparing orthogonal force and unidirectional strain component processing for tool condition monitoring

Signal processing using orthogonal cutting force components for tool condition monitoring has established itself in literature. In the application of single axis strain sensors however a linear combination of cutting force components has to be processed in order to monitor tool wear. This situation may arise when a single axis piezoelectric actuator is simultaneously used as an actuator and a sensor, e.g. its vibration control feedback signal exploited for monitoring purposes. The current paper therefore compares processing of a linear combination of cutting force components to the reference case of processing orthogonal components. Reconstruction of the dynamic force acting at the tool tip from signals obtained during measurements using a strain gauge instrumented tool holder in a turning process is described. An application of this dynamic force signal was simulated on a filter-model of that tool holder that would carry a self-sensing actuator. For comparison of the orthogonal and unidirectional force component tool wear monitoring strategies the same time-delay neural network structure has been applied. Wear-sensitive features are determined by wavelet packet analysis to provide information for tool wear estimation. The probability of a difference less than 5 percentage points between the flank wear estimation errors of above mentioned two processing strategies is at least 95 %. This suggests the viability of simultaneous monitoring and control by using a self-sensing actuator.     

Universal routing schemes

Summary.  In this paper, we deal with the compact routing problem, that is implementing routing schemes that use a minimum memory size on each router. A universal routing scheme is a scheme that applies to all n-node networks. In [31], Peleg and Upfal showed that one cannot implement a universal routing scheme with less than a total of Ω(n ^1+1/(2s+4)) memory bits for any given stretch factor s≧1. We improve this bound for stretch factors s, 1≦s<2, by proving that any near-shortest path universal routing scheme uses a total of Ω(n ²) memory bits in the worst-case. This result is obtained by counting the minimum number of routing functions necessary to route on all n-node networks. Moreover, and more fundamentally, we give a tight bound of Θ(n log n) bits for the local minimum memory requirement of universal routing scheme of stretch factors s, 1≦s<2. More precisely, for any fixed constant ɛ, 0<ɛ<1, there exists a n-node network G on which at least Ω(n ^ɛ) routers require Θ(n log n) bits each to code any routing function on G of stretch factor <2. This means that, whatever you choose the routing scheme, there exists a network on which one cannot compress locally the routing information better than routing tables do. Received: August 1995 / Accepted: August 1996     

A new mutation operator for evolutionary airfoil design

A new mutation operator, ℳ _ijn , capable of operating on a set of adjacent bits in one single step, is introduced. Its features are examined and compared against those of the classical bit–flip mutation. A simple Evolutionary Algorithm, ℳ–EA, based only on selection and ℳ _ijn , is described. This algorithm is used for the solution of an industrial problem, the Inverse Airfoil Design optimization, characterized by high search time to achieve satisfying solutions, and its performance is compared against that offered by a classical binary Genetic Algorithm. The experiments show for our algorithm a noticeable reduction in the time needed to reach a solution of acceptable quality, thus they prove the effectiveness of the proposed operator and its superiority to GAs for the problem at hand.     

A 3-Party Simultaneous Protocol for SUM-INDEX

Abstract. We show that the SUM-INDEX function can be computed by a 3-party simultaneous protocol in which one player sends only O(n ^ɛ ) bits and the other sends O(n ^1-C(ɛ) ) bits (0<C(ɛ)<1 ). This implies that, in the Valiant—Nisan—Wigderson approach for proving circuit lower bounds, the SUM-INDEX function is not suitable as a target function.     

New Techniques for Regular Expression Searching

We present two new techniques for regular expression searching and use them to derive faster practical algorithms. Based on the specific properties of Glushkov’s nondeterministic finite automaton construction algorithm, we show how to encode a deterministic finite automaton (DFA) using O(m2^m) bits, where m is the number of characters, excluding operator symbols, in the regular expression. This compares favorably against the worst case of O(m2^m|Σ|) bits needed by a classical DFA representation (where Σ is the alphabet) and O(m2^2m) bits needed by the Wu and Manber approach implemented in Agrep. We also present a new way to search for regular expressions, which is able to skip text characters. The idea is to determine the minimum length ℓ of a string matching the regular expression, manipulate the original automaton so that it recognizes all the reverse prefixes of length up to ℓ of the strings originally accepted, and use it to skip text characters as done for exact string matching in previous work. We combine these techniques into two algorithms, one able and one unable to skip text characters. The algorithms are simple to implement, and our experiments show that they permit fast searching for regular expressions, normally faster than any existing algorithm.