Condition monitoring - RWTH Publications - RWTH Aachen University

A review on real time measurement techniques for tool condition monitoring and analysis of metalworking fluids

191

Condition monitoring: A review on real-time measurement techniques for tool condition monitoring and analysis of metalworking fluids Anna Lena Demmerling, M.Sc., Research & Development, Rhenus Lub GmbH & Co KG

Univ.-Prof. Dr.-Ing. Dirk Söffker, Chair of Dynamics and Control, University of Duisburg-Essen

1 Introduction In times of rising interest in Industry 4.0 applications for process monitoring are focused by industrial and scientific researchers. Process control has become more and more important to decrease the number of process failures and to ensure products with constant quality. In engineering there is an increasing demand for higher productivity, reproducible products and lower costs. Machining processes like turning, drilling, or milling are monitored in real-time to predict tool breakage and to prevent machine failure or work piece malfunctions. The application of monitoring systems reduces costs for machining tools and improves the process reliability. Tool condition monitoring (TCM) is implemented into tooling machines as standard. The main techniques used are cutting force (thrust, support force, torque), vibration, or acoustic emission measurements. Advanced technologies are not only appointed with a measurement system but also with implemented neural networks for the diversity of tools. The regular process of each tool is trained so that the abnormal process can be detected by changes in measurement results and a prediction of the tool life time becomes possible. The tool change is indicated by the system if corresponding thresholds had been implemented. The detection rate, the amount of false alarms as well as the necessity of human intervention are seldemly discussed. During many machining processes the tool is splashed with a metal working fluid (MWF). The choice of the lubrication concept depends on the application, i.e. material of work piece, machining process, etc. Exterior or interior cooling with full jet, minimum quantity, air or cryogenic cooling, or dry machining is possible. A MWF improves the transportation of chips and heat and minimizes

Institut für Maschinentechnik der Rohstoffindustrie, RWTH Aachen, 2016

192


the friction between tool and work piece through lubricating ingredients. The MWF influences the development of wear and can increase the tool life time [1]. Two main groups of MWF can be distinguished: Non water mixable and water mixable MWF. Water mixable MWF are divided into water soluble and emulsifiable concentrates. For application the concentrates are mixed with water typically in a ratio from 1:20 to 1:10. On the basis of standardized laboratory fluid tests and empirical earned experiences special fluids are proposed for each application. The condition of the MWF is very important for performance and reliability of the machining process because a degraded fluid does not save the tool from wear or the work piece from corrosion as a wellconditioned fluid. Therefore, the right and suitable choice of MWF in good condition is important. On-line monitoring is state-of-the-art for engine oils. Operators realized the importance of a well-conditioned oil to keep the engine running reliably. An ignorance of suitable indicators would raise wear and would damage the engine irreversibly. Therefore, the condition monitoring of oils for example in engines or bearings is well studied and much applied. Concerning machining processes often only flow and pressure of the MWF is controlled on-line but not its condition or its suitability for the current process. Tool wear is nowadays monitored in real-time but the condition of the MWF that could inhibit this wear is not monitored. On the one hand this paper reviews on interesting and forward-looking tool condition monitoring systems with a focus on techniques using acoustic emission (AE). On the other hand, this paper deals with analyzing methods of water mixable fluids that are capable for realtime applications. The idea behind this review is a combination of TCM methods and MWF analysis so that a multisensory monitoring of the MWF and its suitability for the current machining process could be possible. The main idea is: Is it possible to detect changes in the fluid in real-time before the tool wear increases or the surface quality gets worse?

2 Tool condition monitoring techniques Monitoring of machining processes has become an important part for increasing process stability and reliability. The condition of the tool plays a significant role in reaching demanded surface qualities or demanded machining accuracy. The tool’s lifetime depends on the work piece material, the tool’s material, the cutting parameters, and the suitability of the used MWF. Machining of aluminum or titanium and tools made of high-speed steel (HSS) or solid carbide steel with or without a coating differs. According to [2], the wear of tool increases with increasing machining time. Institute for Mining and Metallurgy Machinery, RWTH Aachen University, 2016


193

High friction between cutting tool and work piece exists. The tool material has to deform the work piece material plastically, so that the work piece material is removed by forming chips. If the cutting edges are coated, firstly the coating will be rubbed off, then the full material of the tool. A loss of coating always leads to a change in material combination. Usually wear develops faster when the coating does not protect the weaker full material anymore. Tool wear appears in several forms: Flank wear, crater wear, plastic deformation, chipping, breakage, and built-up edge [3]. A highly worn tool cannot perform as well as a new tool and is not able to fulfill the continuously rising demands for constant qualities of the machining results. Broken tools can damage the work piece irreparable. A TCM system can warn against highly worn tools so that those can be replaced in time. Process stability can be saved and production failures caused by tool breaks can be reduced. As a result, a TCM system can reduce production costs [4].

2.1

Manufacturing solutions

Manufacturers of milling or turning machines recognized the need for monitoring the tool’s condition and offer several TCM systems optionally for their products. In the following some companies are mentioned that have invested a lot of research and development in TCM systems: KOMET BRINKHAUS, MCU, PROMETEC, Schwer + Kopka, and Nordmann. The following overview gives a glance at the used sensors and focusses on acoustic emission techniques. KOMET BRINKHAUS offers monitoring systems for example for tapping processes [5]. The system is able to learn and to differentiate between a normal and an abnormal process. The products of MCU GmbH contain sensors measuring the active power, force, structure-borne noise (acoustic emission), vibration/oscillation, and collision. The limit of TCM measurements by power current is given for very small tools. When the current consumed power of the spindle is as low as the no-load current power, a detection of process failures is not possible. Here acoustic emission is used to monitoring the condition of very small tools [6]. PROMETEC uses acoustic emission sensors to detect collisions and gaps and additionally vibration sensors to detect tool break, tool wear, and imbalances of the spindle, the table, or the work piece [7]. Schwer + Kopka developed process monitoring systems for cold and hot forming, stamping, and assembling [8]. Different piezoelectric sensors are used to measure forces, strain, acoustic emission signals, electrical power consumption, and tool breakage. Nordmann divided its sensors in seven categories [9]: Effective power and torque, force, acoustic and vibration, distance and gab, tool length and work Institut für Maschinentechnik der Rohstoffindustrie, RWTH Aachen, 2016

194


piece position control, work piece dimension control, and tool position control. Acoustic emission is used to detect tool breakage through the cooling jet or to control the gap for grinding machines. All these TCM systems are available on the market. The reliability of failure detection is not specified for each system or each sensor.

2.2

Scientific research

Intensive scientific research has been carried out concerning the relation between TCM and cutting forces ( [10], [11], [12], [13], [14], [15]), vibrations ( [2], [16], [17], [18], [19], [20], [21]), wear images ( [22], [16], [23], [24], [25]), sound energy, power consumption, ultrasonic emission, or temperature [26]. Soft computing techniques as neural network, fuzzy logic, or artificial intelligence are used to predict process data and to optimize process stability [27]. The main part of the investigated techniques is based on the monitoring of forces, vibration, and acoustic emission. The techniques were tested for several machining operations like milling, drilling, or turning. Furthermore, acoustic emission techniques were used intensively for forming and grinding operations. This paper focusses on the research concerning acoustic emission. There have been several reviews on TCM including acoustic emission ( [26], [27], [28], [29]). These reviews give a quite detailed view over the monitoring techniques in several machining operations. In [26] the application of acoustic emission is described for turning, drilling, face milling, and end milling. The authors in [27] focus the monitoring systems using digital images. A review on artificial neural networks during turning is focused in [28]. The paper of [29] was published in 1995 and includes a detailed review on the acoustic emission techniques. The present review paper focusses on approaches and techniques of the past 15 years. In the following parts proven monitoring techniques based on acoustic emission are presented. 2.2.1 Turning In turning the wear of the tool was monitored by acoustic emission [30]. A piezoelectric sensor was placed on the tool holder. Acoustic emission signals were analyzed by power spectral density technique. The value was high for a new tool, decreased at the middle of the tool life and finally increased until the end of tool life. The average of the power spectral density correlated with the maximum flank wear of the turning tool [30]. By implementing suitable thresholds for each tool, the tool’s usability could be limited.

Institute for Mining and Metallurgy Machinery, RWTH Aachen University, 2016


195

The root mean square value of the acoustic emission signal in turning was investigated in [31]. Turning tests were performed at different cutting speeds, feed rates, and depths of cut. Flank wear and acoustic emission was measured during the tests. It was found out that the important signals mainly occur within a frequency range of 30 – 60 kHz. The root mean square value correlates with the flank wear [31]. The relation of acoustic emission and vibration signals during turning was investigated in [32]. With increasing flank wear the surface roughness as well as the signals based on acoustic emission first decreased to a minimum value and then increased steadily until the end of tool life. Acoustic emission and vibration signals were evaluated to be both useful to monitor the turning process. The progression of tool wear could be monitored well by acoustic emission whereas surface roughness could be indicated better by vibration signals [32]. The quality of the machined surfaces was also focused in [33]. The acoustic emission was measured during turning and an indirect correlation between the root mean square value of the acoustic emission and the measured surface roughness could be drawn [33]. 2.2.2 Drilling In [34] acoustic emission in correlation to the thrust force was investigated. Tools with five different wear states were used. Acoustic emission and thrust force were recorded simultaneously. The acoustic emission sensor was placed on the work piece. The root mean square value, the measured area under the rectified signal envelope energy, the amplitude, and the mean power were selected as acoustic emission based features. The author concluded that the mean power could be used for tool wear characterization as shown in the following diagram.


196


Figure 1: Differentiation of tool wear states with mean power of AE, acc. to [34] The correlation between acoustic emission, flank wear, and drill hole shrinkage was evaluated in [35] concerning drilling of glass fiber reinforced epoxy resins. Flank wear, thrust force, and hole shrinkage are influenced by cutting parameters. Acoustic emission signals were evaluated in frequency domain. The power of the measured signal increases with the number of drilled holes. Furthermore, the root mean square correlates to the thrust force, the flank wear and the hole shrinkage in regard to the number of holes [35]. The values increase with the number of drilled holes. The root mean square of acoustic emission correlates best with the flank wear so that a determination of the flank wear via acoustic emission was possible in this case. The effect of tool coatings was investigated by using acoustic emission based on the quantification of wear and the identification of five different wear states [36]. An increasing tool wear results in an increase of the measured rectified signal envelope of the acoustic emission. Due to the wear of the tool the peaks in acoustic emission change [36]. 2.2.3 Milling A decision making model was developed in [37] based on acoustic emission data collected during milling tests. With increasing tool wear amplitude and distribution of the acoustic emission signal changed. The signals were analyzed by wavelet packet transform and the wavelet packet energy was calculated. The normalized energy increased with machining time and increasing wear [37]. Institute for Mining and Metallurgy Machinery, RWTH Aachen University, 2016


197

Acoustic emission was used to detect tool and work piece malfunctions during milling in [38]. The aims were to identify acoustic emission signals independent from the tool path to calibrate acoustic emission sensory against proceeding tool wear and to detect surface anomalies resulting from machining failures. Milling tests were conducted at an Inconel 718 work piece and the acoustic emission sensor was placed onto the work piece. During the tests acoustic emission, forces in three dimensions, and torque were recorded. From the test evaluation resulted that monitoring of the cutting insert’s wear was possible by applying the Short-Time-Fourier-transformation to the acoustic emission signals. With the help of the area underneath the envelope of the resultant cutting force, the measured area of the rectified acoustic emission signal envelope, and metallographic investigations a relation between the acoustic emission, cutting forces, and the surface quality could be established [38]. In further works [39] a cutting tool with an intentionally damaged insert was investigated. The acoustic emission signal was processed by time-frequency domain analysis and advanced processing techniques. A differentiation between normal cutting and cutting with resulting surface anomalies could be detected. The application of acoustic emission and cutting forces for TCM in micro-milling was researched in [40]. A strong relation between acoustic emission and tool wear was found. Although the cutting force signals were disturbed by vibrations the monitoring of tool wear was possible. The author suggested that the usage of more signals is preferable to minimize the diagnosis uncertainty and to make TCM more reliable. Better results are achieved by the measurement of cutting forces and acoustic emission signals than a technique only based on acoustic emission measurement [40]. 2.2.4 Combinations Acoustic emission measurement was combined with other measurement techniques as vibration or cutting forces. The correlation of these methods was investigated for example in [41]. A multiple sensor system based on a cutting force sensor, a vibration sensor, and an acoustic emission sensor was tested during turning. The measurement results were evaluated by neural network and a tool life time prediction model was implemented. With the help of an online cognitive decision making system it was possible to realize a machine learning algorithm. In [42] artificial neural network based on features of cutting forces, vibrations, and acoustic emissions was developed and tested during milling tests. The model provided a higher precision than a model based on single features. Sensing combination systems based on cutting force, vibration, and/or


198


strain was developed for example in [43] or [44]. The fusion of vibration features and digital images was investigated in [45]. 2.2.5 Others An acoustic emission technique for precision machining was investigated in [46]. It was concluded that acoustic emission was very sensitive to control parameters in high frequency range. Normally force or vibration sensors lose accuracy in the high frequency range because of limited band width and due to noise. High frequency signals of acoustic emission were measured for a small length scale of material removal whereas the frequency declines for increasing material removal. That means acoustic emission is more sensitive for small material removals and high levels of precision whereas the use of forces or vibration signals is better for larger chip thicknesses. Concerning the ultraprecision machining and micro cutting mechanisms [46], signal measurements based on acoustic emission are suitable techniques for monitoring because noise and disturbances caused by machine elements could be minimized. Acoustic emission was also tested to control the wear status of gears ( [47], [48]) or for forming and grinding operations ( [49], [50], [51]). Friction and wear states of sliding plates were investigated in [52] and [53]. The lubrication effect at a head/disk interface was examined with acoustic emission in [54]. Three different lubricants were investigated and a difference between the acoustic emission responses has been observed due to higher wear of the disk. Concerning the head/disk interface the lubricant was identified “as one of the most important factors”. The occurring debris in the interface could be measured by changes in the acoustic emission signal. The levels of the acoustic emission signal were higher for increasing wear and increasing lubricant degradation. 2.2.6 Summary With acoustic emission an indirect measurement of the wear of turning, drilling, and milling tools is possible. The roughness of turned surfaces is related to the acoustic emission signals. Special features of the acoustic emission signal can be used to detect different wear states of a drilling tool. In drilling the correlation between acoustic emission and drill hole shrinkage was stated. Changes in cutting speeds, feed rates, and tool coatings could also be detected during drilling. Concerning milling processes the differentiation between normal and abnormal machining could be made by measuring the acoustic emission signals. A relation between acoustic emission, tool wear, and surface quality was found. For head/disk interfaces the detection of higher wear caused by Institute for Mining and Metallurgy Machinery, RWTH Aachen University, 2016


199

lubricant degradation was possible by measuring the acoustic emission signals. The differentiation of very similar lubricants differing only in one ingredient was not investigated.

3 Analyzing methods of metal working fluids The condition of a MWF, oil, or lubricant within a tribology system is significant for its functionality. The performance of the used processing fluid is important to provide machine elements of failure and to enhance the reliability of processes. With a processing fluid optimized for the individual application and each metalto-metal contact a reduction of wear of machine elements or tools is possible. Concerning the machining of metals, a suitable MWF is able to enhance surface qualities and to increase process stability. Regular investigations are necessary to fix the best condition of the MWF. The technical rules for hazardous substances 611 define the correct use of MWF and prescribe weekly checks of pH value and nitrite content [55]. Other characteristics influencing the MWF’s performance can be analyzed by several measurement techniques for example the pH level or the concentration. For industrial applications fluid control systems were developed and are offered for example by Rhenus Lub and Tiefenbach ( [56], [57]). Depending on the system flow, concentration, conductivity, pH value, temperature, germ count, and the filling level of storage tanks – if the system provides automatic refill or dosage – can be monitored. An on-line measurement of MWF can be realized and a fast intervention in form of an adjustment of concentration or a treatment of microbiological activity is possible. Here measurement techniques for analyzing processing fluids are described. Such techniques with applicability for on-line measurements of water-miscible MWF are focused. The aim is to figure out the diversity of techniques and characteristics to be measured and analyzed.

3.1

Oil concentration in MWF

In practical applications the oil concentration of water mixed MWF is measured with the help of a refractometer. After a short calibration of the refractometer the refraction index of the measured fluid can be measured. The MWF’s concentration can be determined approximately with the known refractive coefficient and the density of the measured fluid. That is a practical and stateof-the-art measuring method in lubricant industry. Institut für Maschinentechnik der Rohstoffindustrie, RWTH Aachen, 2016

200


An interesting measurement strategy that could be applied real-time for MWF is the monitoring of oil viscosity by ultrasonic-based testing. A method indicating changes in viscosity was described in [58]. The authors state that the system measures the acoustic absorption by leading an ultrasonic wave through the fluid. The ultrasonic wave sent out is weakened by frictional effects caused by the fluid’s viscosity so that a detection of viscosity changes is possible: The higher the viscosity, the longer the relative transit time of the existing emitted ultrasonic waves. The application for concentration measurements in MWF could be considered. The technique of dynamic light scattering was applied and tested for industrial degreasing baths [59]. To monitor the condition of the degreasing bath the micelles are measured optically. In degreasing baths, the micelles get larger and multiple when the condition of the bath gets worse. The capacity of degreasing declines by rising micelles. With the help of this technique the determination of the micelles’ sizes is possible. The contamination rate and the loss of detergent could be concluded from this measurement. Possibly the dynamic light scattering technique could also be applied to determine the concentration or the dispersity in emulsified MWF.

3.2

Fluidic contaminants

A processing fluid can be contaminated with fluidic contaminants such as tramp oils, water, gasoline, or coolant. An optical sensing method was developed in [60]. The object shape-based methodology was able to detect different concentrations of water, gasoline, or coolant by capturing and processing images of the fluid. The on-line applicability of this system was verified for engine lubricant oil with artificial generated contaminant concentrations. The optical sensing method could be feasible for water-miscible MWF concerning the contamination with tramp oil or coolant. Because oil droplets within the fluid could cause incorrect measurements, the method could only be practical for water-soluble MWF.

3.3

Additive/Water content

The detection of water content, the oxidation number, or the total acid number plays a significant role for the condition of lubricant oils. Water can lead to corrosion or to a tearing-off of the lubricating film and later in use to an increase of wear. The total acid number or the oxidation number can be indicators for degraded lubricant oils. The total acid number increases due to a continuously degrading oil [61]. The oxidation number changes between a fresh and a used



201

oil [62]. Important anti-wear additives like phosphorus contribute to less wear by depositing on the metallic surfaces [63]. An optical measurement system was developed to determine significant relations between optical measurement results and the total acid number, the content of water, and phosphorus [61]. The experimental system consisted of three optical fiber setups: Absorption, fluorescence, and scattering measurement. Single setups are not able to give evidence about the condition of the fluid. From the combination of the measurements derived a statistical dependence of absorption, fluorescence, and scattering with the characteristics of the fluids. The phosphorus content could be estimated with a coefficient of determination of 96.9 %. The coefficients of determination for the water content and the total acid number were 84.3 and 73.9 %. The experimental setup is generally suitable for on-line applications. The determination of the total acid number or the water content shows great deviations between estimated and true values. The coefficient of determination showed that only great changes in water contents and acid numbers can be reliably determined. The authors state that a determination of phosphorus content in oils was possible and quite precise. The suitability for MWF has to be tested. The oxidation number changes between fresh and used oil and can be measured with the help of the absorption in an infrared spectrum. In [62] an oil condition monitoring device was developed using infrared spectrometry. Because of the high dilution grade in water mixed MWF a measurement of the infrared spectrum gains worse results. Despite a good applicability for oils this technique is not considerable for MWF.

3.4

Changes in pH level

In application of MWF a change in pH level can influence the corrosion protection for metallic components significantly or could be an indicator for growing microbiological activity. Thus the monitoring of pH level is important for the MWF’s condition. A capacitance-based measuring system was tested for the detection of changes in pH level [58]. The capacitance between to copper electrodes was measured in a test fluid acting as a dielectric. From the conducted experiments at different pH levels a relation between the pH value and the relative capacitance could be extracted. A decreasing pH value affects an increasing relative capacitance. Experiments were conducted at a neutral until slightly acidic environment. Deviations of several measurement points were not specified so that the precision or suitability of this measurement technique for low pH changes cannot be evaluated. The tested pH levels were


202


not application-relevant for MWF so that a particular evaluation for pH values between 7.8 and 9.8 is necessary to judge the applicability for MWF.

3.5

Microbiological activity

Water-mixed lubricant systems are typically open systems and thus exposed to biological degradation. The MWF’s degradation is accelerated by bacteria and fungi that subsist on the MWF’s components [64]. As an effect of microbiological activity functional MWF additives are depleted and the MWF system fouls. The unpleasant odor results from volatile organic chemicals produced by the microorganisms. According to [65] the life of the MWF and the machining tools will be shortened if the operator did not intervene by adding bactericides or fungicides or – if the attack is further progressed – by cleaning the whole system and refilling fresh MWF. The microorganisms can deplete corrosion inhibition additives so that the risk of work piece corrosion rises. A severe microbiological attack of the MWF often means a machine downtime. Therefore, a condition monitoring of the microbiological activity is important to realize an early intervention and correction of the MWF. In practice several off-line testing methods to determine the microbial load exist: Adenosine triphosphate test methods like ASTM E2694 or ASTM 7463, culture tests like dip-slides or the standard plate count method, and the laser desorption mass spectrometry [64]. An on-line application of culturing tests can be excluded from the outset as well as the ATP methods whose samples have to be pretreated with fiber filters, buffer reagents, and/or centrifuges. A lot of scientific research dealt with the on-line applicability of the matrix assisted laser desorption ionization-time of flight mass spectrometry. A technical device was developed in [66] to deposit the sample continuously onto the repelled plate of the mass spectrometry. On the plate laser desorption took place and the sample was analyzed via matrix assisted laser desorption ionization-time of flight mass spectrometry. This technique enabled on-line measurements but applicability for MWF was actually not tested.

3.6

Corrosion inhibition

The corrosion inhibition is a substantial property of a MWF. Antioxidant additives deposit on the metal surface of the work piece or of the machine components and build a passivation layer [67]. The layer saves the material for oxidation caused by water from the fluid or by air humidity. Because of chemical deposition the additives are discharged by and by with the machined work pieces. Ferrous, non-ferrous, or nonmetals materials as cast iron, copper, or Institute for Mining and Metallurgy Machinery, RWTH Aachen University, 2016


203

aluminum respond differently to the same fluid because electronegativity and redox potential differ between the different combinations. According to its additives a fluid can cause heavy corrosion at ferrous materials but does not show corrosion on nonmetals. The type and content of the right additive is the decisive factor. If the content of additives against non-ferrous corrosion reduces, the risk for copper corrosion will increase. The knowledge of the additive’s loss is important for the avoidance of corroded and irreparable work pieces. A corrosion sensor based on the electrical resistance of a thin copper film was tested under laboratory and field conditions [67]. The corrosion sensor was hold into the fluid and the electrical resistance was measured. During the corrosion process the copper film lost material and the thickness of the layer decreased. As an effect of proceeding copper oxidation the sensor resistance increased and the corrosion rate decreased [67]. Different types of corrosion effects lead to different behaviors in sensor resistance. The reproducibility of this method increased for thinner copper layers. A copper layer of 100 nm thickness shows a very good reproducibility but needs a replacement after every 24 h [67] which is not considerable for industrial applications. The layer thickness of 1000 nm enables an operating time of 15 days requiring that the measured fluid was not optimized during the operating time. The experiments were conducted with consistently aging oils without adding substances to weaken the corrosion. The effect on adding special corrosion inhibited substances during the test was not investigated. This is the aim of a monitoring system in industrial practice and the feasibility for this measurement system has to be tested. The sense of an on-line application of this corrosion sensor in MWF can be considered.

3.7

Summary

The on-line measurement of concentration, pH value, conductivity, and microbiological activity is possible for MWF and already available in industrial products. The content of special additives for example phosphorus, sulfur, or glycol is not measured on-line but these could be very important for machining performance or corrosion effects. Of the introduced methods above particularly interesting is the developed copper corrosion sensor that could also be tested for aluminum or specially corrosion susceptible steel types and for the applicability in MWF. The developed device containing absorption, fluorescence, and scattering measurements could be interesting for additive determination. Institut für Maschinentechnik der Rohstoffindustrie, RWTH Aachen, 2016

204


4 Conclusions Because of higher efforts of technical applications and the necessity of constant qualities, less failures, and lower production costs in machining monitoring of the process are necessary. In this paper tool condition monitoring (TCM) techniques to detect increasing tool wear or to determine the tool’s usability have been presented. Tool wear and the quality of the machined work piece surfaces can be characterized by TCM techniques based on acoustic emission. Furthermore, analyzing methods for metal working fluids (MWF) that could be capable for real-time applications have been introduced. Promising theories were presented concerning measurements of pH level, additive content, and corrosion inhibition. Tool condition monitoring techniques and MWF monitoring systems are already applied in industrial processes. Advantages like waste and cost reduction and quality control have been experienced. A system checking the suitability of the MWF for the current machining process in real-time was not found in literature or in industrial products.

References [1] D. Biermann and M. Kirschner, "Wahl des Kühlschmierstoffkonzeptes als Schlüssel zum Erfolg," Werkstoffe in der Fertigung 5, pp. 26-29, 2015. [2] A. González-Laguna, J. Barreiro, A. Fernández-Abia, E. Alegre and V. González-Castro, "Design of a TCM system based on vibration signal for metal turning processes," Procedia Engineering 132, pp. 405-412, 2015. Coromant, "Wear types," 2016. [Online]. Available: [3] SANDVIK http://www.sandvik.coromant.com/en-gb/knowledge/drilling/wear-andtroubleshooting/wear_types/pages/default.aspx. [Accessed 20. 09. 2016]. [4] Schubert System Elektronik GmbH, "BK Mikro Werkzeugueberwachung," 2016. [Online]. Available: http://www.bkmikro.de/werkzeugueberwachung.html. [Accessed 15. 09. 2016]. [5] KOMET GROUP GmbH, "Youtube-Video: Tool and Process Monitoring with ToolScope Tapping," 2016. [Online]. Available: https://www.youtube.com/ watch?v=aS756V9l3Ds. [Accessed 15. 09. 2016]. [6] MCU GmbH, "Acoustic Emission," 2016. [Online]. Available: http://www.mcu-gmbh.de/?id=22. [Accessed 15. 09. 2016]. [7] PROMETEC mbH, "Products," 2016. [Online]. Available: http://www.prometec.com/ GE/products_machining.html. [Accessed 15. 09. 2016]. [8] Schwer + Kopka GmbH, "Sensors," 2016. [Online]. Available: http://www.schwer-kopka.de/ english/sensors.html. [Accessed 15. 09. 2016]. [9] Nordmann International GmbH, "Sensors," 2016. [Online]. Available: http://www.toolmonitoring.com/ sensoren.html. [Accessed 15. 09. 2016]. [10] F. Cus und U. Zuperl, „Real-Time Cutting Tool Condition Monitoring in Milling,“ Journal of Mechanical Engineering 57, pp. 142-150, 2011. [11] Y. Wu, G. Hong and W. Wong, "Prognosis of the probability of failure in tool condition monitoring applicatio-atime series based approach," International Journal of Advanced Manufacturing Technology 76, pp. 513-521, 2015.



205

[12] M. Nouri, B. K. Fussel, B. L. Ziniti and E. Linder, "Real-time tool wear monitoring in milling using a cutting cindition independent method," International Journal of Machine Tools & Manufacture 89, pp. 1-13, 2015. [13] S. Afazov, D. Tdebski, S. Ratchev, J. Segal and S. Liu, "Effects of micro-milling conditions on the cutting forces and process stability," Journal of Materials Processing Technology 213, pp. 671-684, 2013. [14] H. Ertunc, K. Loparo, E. Ozdemir and H. Ocak, "Real Time Monitoring of Tool Wear Using Multiple Modeling Method," Electric Machines and Drives Conference, 2001. [15] H. Wiklund, "Bayesian and regression approaches to on-line prediction of residual tool life," Quality and Reliability Engineering International 14, pp. 303-309, 1998. [16] C. Madhusudana, S. Budati and N. Gangadhar, "Fault diagnosis studies of face milling cutter using machine learning approach," Journal of Low Frequency Noise, Vibration and Active Control 0 (0), pp. 1-11, 2016. [17] A. Rmili, A. Ouahabi, R. Serra and R. Leroy, "An automatic system based on vibratory analysis for cutting tool wear monitoring," Measurement 77, pp. 117-123, 2016. [18] K. Arun Vikram, C. Ratnam and K. Sankar Narayana, "Vibration diagnosis and prognostics of Turn-milling operations using HSS and carbide end mill cutters," Procedia Technology 23, pp. 217-224, 2016. [19] K. Venkata Rao, B. Murthy and N. Mohan Rao, "Cutting tool condition monitoring by analyzing surface roughness, work piece vibration and volume of metal removed for AISI 1040 steel in boring," Measurement 46, pp. 4075-4085, 2013. [20] M. Saimurugan, K. Ramachandran, V. Sugumaran and N. Sakthivel, "Multi component fault diagnosis of rotational mechanical system based on decision tree and support vector machine," Expert Systems with Applications 38, pp. 3819-3826, 2011. [21] S. Orhan, A. Osman Er, N. Camuscu and E. Aslan, "Tool wear evaluation by vibration analysis during end millinf of AISI D3 work tool steel with 35 HRC hardness," Independent Nondestructive Testing and Evaluation International 40, pp. 121-126, 2007. [22] W. Weis, "Processing of Optical Sensor Data for Tool Monitoring with Neural Networks," in WESCON/94. Idea/Microelectronics., Anaheim, CA, 1994. [23] D. Lipinski, W. Kacalak and R. Tomkowski, "Methodology of Evaluation of Abrasive Tool Wear with the Use of Laser Scanning Microscopy," SCANNING VOL. 36, pp. 53-63, 2014. [24] S. Dutta, A. Datta, N. Das Chakladar, S. Pal, S. Mukhopadhyay and R. Sen, "Detection of tool condition from turned surface images using an accurate grey level co-occurence technique," Precision Engineering 36, pp. 458-466, 2012. [25] Y. Zhang, Y. Zhang, H. Tang and L. Wang, "Images acquisition of a high-speed boring cutter for tool condition monitoring purposes," International Journal of Advanced Manufacturing Technology 48, pp. 455-460, 2010. [26] A. G. Rehorn, J. Jiang and P. E. Orban, "State-of-the-art methods and results in tool condition monitoring: a review," International Journal of Advanced Manufacturing Technology 26, pp. 693-710, 2005. [27] S. Dutta, S. Pal, S. Mukhopadhyay and R. Sen, "Application of digital image processing in tool condition monitoring: A review," CIRP Journal of Manufacturing Science and Technology 6, pp. 212-232, 2013. [28] B. Sick, "On-line and indirect tool wear monitoring in turning with artificial neural networks: A review of more than a decade of research," Mechanical Systems and Signal Processing 16 (4), pp. 487-546, 2002. [29] G. Byrne, D. Dornfeld, I. Inasaki, G. Ketteler, W. König and R. Teti, "Tool Condition Monitoring The Status of Research and Industrial Application," CIRP Annals - Manufacturing Technology 44, Issue 2 , pp. 541-567, 1995. [30] L. Maia, A. Abrao, W. Vasconcelos, W. Sales and A. Machado, "A new approach for detection of wear mechanisms and determination of tool life in turning using acoustic emission," Tribology International 92, pp. 519-532, 2015. [31] S. Kosaraju, V. Gopal Anne and B. Popuri, "Online tool condition monitoring in turning titanium


206

[32]

[33]

[34] [35]

[36] [37] [38]

[39]

[40]

[41] [42]

[43]

[44]

[45]

[46] [47]

[48]

[49]

[50]


(grade 5) using acoustic emission: modeling," International Journal of Advanced Manufacturing Technology 67, pp. 1947-1954, 2013. M. Bhuiyan, I. Choudhury and Y. Nukman, "Tool Condition Monitoring using Acoustic Emissionand and Vibration Signature in Turning," in Proceedings of the World Congress on Engineering , London, U.K., 2012. C. Beggan, M. Woulfe, P. Young and G. Byrne, "Using Acoustic Emission to Predict Surface Quality," International Journal of Advanced Manufacturing Technology 15, pp. 737-742, 1999. G. Ferrari and M. Gómez, "Correlation Between Acoustic Emission, Thrust and Tool Wear in Drilling," Procedia Materials Science 8, pp. 693-701, 2015. S. Arul, L. Vijayaraghavan and S. Malhotra, "Online monitoring of acoustic emission for quality control in drilling of polymeric composites," Journal of Materials Processing Technology 185, pp. 184-190, 2007. A. Quadro and J. Branco, "Analysis of the acoustic emission during drilling test," Surface and Coatings Technology 94-95, pp. 691-695, 1997. S. Liu, H. Zhang, C. Li, H. Lu and Y. Hu, "Fuzzy Reliability Estimation for Cutting Tools," Procedia CIRP 15, pp. 62-67, 2014. I. Marinescu and D. Axinte, "A critical analysis of effectiveness of acoustic emission signals to detect tool ans workpiece malfunctions in milling operations," International Journal of Machine Tools & Manufacture 48, pp. 1148-1160, 2008. I. Marinescu and D. Axinte, "A time-frequency acoustic emission-based monitoring technique to identify workpiece surface malfunctions in milling with multiple teeth cutting simultaneously," International Journal of Machine Tools & Manufacture 49, pp. 53-65, 2009. K. Jemielniak and P. Arrazola, "Application of AE and cutting force signals in tool condition monitoring in micro-milling," CIRP Journal of Manufacturing Science and Technology 1 , pp. 97-102, 2008. S. Karam, P. Centobelli, D. D'Addona and R. Teti, "Online prediction of cutting tool life in turning via cognitive decision making," Procedia CIRP 41, pp. 927-932, 2016. H. Gao, D. Li, M. Xu, M. Zhao, X. Shi and H. Huang, "Tool Wear Monitoring Based on Novel Evolutionary Artificial Neiral Networks," Sixth International Conference on Naturak Computation, pp. 1339-1343, 2010. D. Shi, D. Axinte and N. Gindy, "Development of an online machining process monitoring system: a case study of the broaching process," International Journal of Advanced Manufacturing Technology 34, pp. 34-46, 2007. H. Chelladurai, V. Jain and N. Vyas, "Development of a cutting tool condition monitroing system for high speed turning operation by vibration and strain analysis," International Journal of Advanced Manufacturing Technology 37, pp. 471-485, 2008. G. Krishna Pradeep, M. Saimurugan and S. Ravi Kumar, "Tool wear monitoring using fusion of vibration signals and digital image," Journal of Chemical and Pharmaceutical Sciences Vol. 9, pp. 537-541, 2016. D. Dornfeld, Y. Lee and A. Chang, "Monitoring of Ultraprecision Machining Processes," International Journal of Advanced Manufacturing Technology 21, pp. 571-578, 2003. T. Toutountzakis, C. Tan and D. Mba, "Application of acoustic emission to seeded gear fault detection," Independent Nondestructive Testing and Evaluation International 38, pp. 27-36, 2005. G. Feng, M.-Y. Tsai and Y.-R. Jeng, "A micromachined, high singal-to-noise ratio, acoustic emission sensor and its application to monitor dynamic wear," Sensors and Actuators A 188, pp. 56-65, 2012. T. Skare, "Dynamiskt belastade tribologiska system under plastisk formning, del II --analyserade genom akustisk emission," 2001. [Online]. Available: http://lup.lub.lu.se/search/record/20182. [Accessed 08. 10. 2016]. D. Lee, I. Hwang, C. Valente, J. Oliveira and D. Dornfeld, "Precision manufacturing process monitoring with acoustic emission," International Journal of Machine Tools & Manufacture 46, pp. 176-188, 2006.



207

[51] T. Jayakumar, C. Mukhopadhyay, S. Venugopal, S. Mannan und B. Raj, „A review of the application of acoustic emission techniques for monitoring forming and grinding processes,“ Journal of Materials Processing Technology 159, pp. 48-61, 2005. [52] D. Baccar and D. Söffker, "Wear detection by means of wavelet-based acosutic emission analysis," Mechanical Systems and Signal Processing 60-61, pp. 198-207, 2015. [53] F. Saiedi, S. Shevchik and K. Wasmer, "Automatic detection of scuffing using acoustic emission," Tribology International 94, pp. 112-117, 2016. [54] Y. Liu, C. Jiaa and A. Eltoukhy, "Acoustic Emission Study of Lubricant Effect on Proximity Contact Recording," IEEE Transactions on Magnetics 33, pp. 3160-3162, 1997. [55] Bundesministerium für Arbeit und Soziales, "TRGS 611," 05. 2007. [Online]. Available: http://www.baua.de/ de/Themen-von-A-Z/ Gefahrstoffe/TRGS/ pdf/TRGS611.pdf?__blob=publicationFile. [Accessed 18. 06. 2016]. [56] Rhenus Lub GmbH & Co. KG, "FluidSafe LubControlSystem," [Online]. Available: http://www.rhenuslub.de/en/fluidsafe.html. [Accessed 07. 10. 2016]. [57] Tiefenbach Control Systems GmbH, "Schmierstoffseminar des Verbandes Schmierstoff-Industrie e.V.," in Prozesssichere Kontrolle und Steuerung wassermischbarer Kühlschmierstoffe, Köln, 2016. [58] M. Appleby, F. K. Choy, L. Du and J. Zhe, "Oil debris and viscosity monitoring using ultrasonic and capacitance/inductance measurements," Lubrication Science 2013; 25, pp. 507-524, 2013. [59] M. Menta, J. Frayet, B. Grassl, C. Gleyzes, A. Castetbon and M. Potin-Gautier, "Development of an alternative analytical methodology to monitor industrial degreasing baths by dynamic light scattering," Journal of Cleaner Production 113, pp. 981-988, 2016. [60] E. Bordatchev, H. Aghayan and J. Yang, "Object shape-based optical sensing methodology and system for condition monitoring of contaminated engine lubricants," Optics and Lasers in Engineering 54, pp. 128-138, 2014. [61] A. G. Mignani, L. Ciaccheri, N. Díaz-Herrera, A. A. Mencaglia, H. Ottevaere, H. Thienpont, S. Francalanci, A. Paccagnini and F. S. Pavone, "Optical fiber spectroscopy for measuring quality indicators of lubricant oils," Measurement Science and Technology 20, pp. 1-7, 2009. [62] A. Agoston, C. Schneidhofer and B. Jakoby, "A concept of an infrared sensor system for oil condition monitoring," Elektrotechnik & Informationstechnik 125/3, pp. 71-75, 2008. [63] J. Schulz and W. Holweger, Wechselwirkung von Additiven mit Metalloberflächen, Renningen: Expert-Verlag, 2010. [64] T. Koch, F. Passman and A. Rabenstein, "Comparative study of microbiological monitoring of water-miscible metalworking fluids," International Biodeterioration & Biodegradation 98, pp. 19-25, 2015. [65] T. Koch, "Auswirkungen des mikrobiellen Befalls auf das Zerspanergebnis (Teil 3)," Journal of Heat Treatment and Materials 63, pp. 115-132, 2008. [66] J. Preisler, F. Foret und B. L. Karger, „On-Line MALDI-TOF MS Using a Continous Vacuum Deposition Interface,“ Analytical Chemistry 70, pp. 5278-5287, 1998. [67] A. Agoston, N. Dörr and B. Jakoby, "Corrosion sensors for engine oils - laboratory evaluation and field tests," Sensors and Actuators B 127, pp. 15-21, 2007.


208

