Tool wear and surface roughness prediction in turning steel under minimum quantity lubrication

Abstract

Minimum quantity lubrication (MQL) refers to the use ofcuuing fluid, of only a minute amount typically of a flow rate of 50 to 500 ml/hour which is about three to four. orders or magnitude l()\'er than the <IlTIount commonly u~ed in nood cooling condition. The concept of minimum quantity lubrication (MQL) has been suggested since a decade ago as a means of addressing the issues of environmental intrusivene,s and occupational hazards associated with the airborne culling fluid particles on factory shop floors. This research work deals with experimental investigation on the role of MQL by VG-68 cutting oil on chIp thickness ratio, cutting tcmpcraturc, cUlting forces, tool wcar and surf<lce roughness in turning medium carbon steel at industrial speed-feed combinations by uncoatcd carbide insert and also to develop an Al1ificialNeural Network (ANN) modcl to predict tool wear and surface roughness in a MQL environment. The encouraging reSLIlts from expcrimcntal investigations include significant reduction in 1001 wcar rate, dimensional inuceuraey und surface roughncss by MQL over dry lTI<lchiningmainly through rcduction in the cutting zone temperature and favorablc chunge in the chip-tool and work-tool interaction. Tool wcar and tool1ife influence the productivity, quality, surfacc intcgrity, cost and profit in any machining process. So, prediclion of tool wear and surface roughness plays a significant rolc in indu~try for highcr produ~livity and surfocc integrity and for JnunufaclUringproec~" planning. Artificial Neural Network (AN1':) is a very promising tool in the field of modeling and monitoring of machining operations and for process optimization. It can recognize pattern in the past data and base on that patlern il can act as a pattcrn matching englnc to forecast the future data. The advantage of ANNs over mathcmatical model is that thcy do not reqUire a precise formulation of phy~ical relationship; they only need experimental results, In lhis study, an Artificial Neural Network (ANN) model has been developed for prediction of tool wear and sllrfacc roughness. The input parumeters of the ANN model are the four machining parameterscutting speed, feed rate, dcplh of cut, machining time and thc oulput parumeters arc four process parameters which arc principal flank wear, maximum principal flank wear, auxiliary flank wear and surface roughness_ The proposcd model can predict tool wear and surface rollghness which is very clo>e to experimental values. The rc>ults of the ANN model show that thc model can be used for the optimization of the cutling process Le. culling parameters can be set in advance prior to pcrform machining operations for cfficient and economic production and for the purpose of manufacturing process planning in a properly designed MQL environment.