POSSIBILITIES OF 3D MACHINING OF MATERIALS BY ABRASIVE WATER JETS

Machining of materials through classical way, i.e. using conventional tools for turning, drilling, milling, grinding and polishing, has some limits that can be overcome applying an abrasive water jet (AWJ). Therefore, some possibilities of 3D machining by AWJ placed on 6 axes robot have been tested. Programming of traverse speeds and tilting angles of cutting head was based on Hlaváč’s theoretical model. Low pressure pump has been used for tests. Because of very low pumping pressure, a selfdesigned and manufactured special mixing chamber was used in the experiments. The article deals with preliminary results and points at the direction of further research.


INTRODUCTION
Abrasive water jet (AWJ) machining has been known for over 40 years. It was introduced, described and presented by Hashish [1]. It is often used to cut either semi-finished products or even final products, namely from plan-parallel plates of material. Nevertheless, applications of abrasive water jets for milling [2], turning [3], grinding [4] or polishing [5] are tested more and more often, because they bring some benefits regarding classical machining processes. Utilization of abrasive water jet as a machining tool for composite materials and rocks is getting broader [6][7][8]. One of the important benefits of AWJ utilization is low probability of damage of this tool due to sudden material strength changes. This fact can be a big advantage in cases, when various materials are to be machined, e.g. for decorative purposes or small-scale production. Therefore, job-shops have arisen besides big firms applying AWJ for their large-series production of rarely variable semi-finished or final products. The job-shops machine semi-final or final products from materials demanded by customer, therefore, need to cover a whole scale of material strengths' changes. The tested machining system is based on a low pressure pump with approximately three times higher flow rate regarding commonly used high pressure pumps and six axes robot. The first experience, experimental results and further research plans are presented in this article.

THEORETICAL BACKGROUND
The theoretical base for AWJ machining control has been published few years ago by Hlaváč [9] and Hlaváč et al. [10,11]. It is focused on the two important parameters closely related to the jet penetration through materiallimit penetration depth and limit traverse speed. The limit penetration depth is the maximum average one (for selected traverse speed, material type and jet parameters) that can be reached in material by AWJ (Eq. 1).
Similarly, the limit traverse speed is the maximum average one (for selected material type, thickness and jet parameters) that enables to provide dividing cut (Eq. 2). Both these factors are closely connected with the two main problems limiting AWJ machining accuracy: the trailback and the taper. The typical simplistic description of jet penetration through material is replacement of the real trajectory by simple curves, namely of a parabolic shape. The respective equations describing the trailback and the taper are presented in articles published by Hlaváč et al. [10,11]. Equation (3) (5) It is evident that compensation of influence of the diameter of an abrasive focussing tube, the trailback and the taper shift can be suppressed by jet tilting and correction of the trajectory radius. Therefore, these corrections were tested in the experimental part of the research work.

EXPERIMENTAL SET-UP
Experiments were performed with a special injection abrasive water jet head for low pressure and high flow rate. The deformation of column samples and reduction of difference between the top and the bottom diameters was tested. The photo of the robot used for sample preparation with various tilting of cutting head is presented in Fig. 1. The experimental conditions used in all presented tests are summarized in Table 1. Results of column cutting with cutting head without and with tilting are presented in Fig. 2. This figure also shows the typical striations on the samples' walls. It is evident that non-tilted jet makes more noticeable striations and the sample is rather truncated cone shaped then a "column" shaped. By contrast to it, the tilted jet produces rather barrel shaped samples with striations better visible even in the bottom part. It can also be noticed by the naked eye that diameter of the top base of the sample produced by a non-tilted jet is smaller than that of the tilted jet and some slight increase of the diameter of the bottom base can be also noticeable.   The "mesh" specification is commercial indication provided by suppliers.
Several columns were cut: one half of them with jet axis perpendicular to the surface of the plan-parallel sheet of composite plate, the second half with tilting of the cutting head compensating deformation caused by trailback. Both sets of samples were measured on the top and bottom to compare their diameters with each other.

DISCUSSION
Preliminary results aimed at column sample distortion proved that tilting of the cutting head is a proper way for reduction of trailback and the taper. The difference between diameters of column bases on the inlet side and the outlet one has been reduced by 204 % eliminating the trailback. Elimination of the taper causes additional 50 % of reduction. The resulting average diameters after tilting in both directions (compensation of trailback and taper) are 18.48 mm on the top and 18.58 mm on the bottom, i.e. the diameter difference is only 0.1 which means 0.52 % of real top diameter (0.48 % for the set-up diameter). Difference between the set-up and the real diameter is caused by leaving out the jet radius being about 1.5 mm. For real object diameter 20 mm the setup diameter should be approximately 21.5 mm. The experiments have also proved that even a low pressure AWJ can efficiently cut composite materials. Therefore, the costs of cutting can be reduced, because pump pressure can be lowered and it means much lower capital costs and also operational costs (pump maintenance). The benefit of the AWJ composite cutting is negligible production of air pollution, namely composite material dust and toxic fumes.
Provided that a robot is used for manipulation with cutting head, the possibilities of 3D machining will increase substantially. Unfortunately, programming of cutting of the 3D objects by abrasive water jet is quite difficult, because it is necessary to take into account that residual energy of the AWJ is still efficient in material damage. Therefore, the programming process needs to calculate with anticipated directions of residual jet deflection. For well-prepared 3D AWJ machining some operations can be less time consuming and more precise. However, the proper programming is not possible without deep and exact knowledge of deflected jet behaviour. To obtain all necessary information the further research of AWJ, both theoretical and experimental, is inevitable.

CONCLUSIONS
The preliminary experiments aimed at AWJ machining of composite materials proved that such machining is possible with a relatively high precision. The accuracy of the machining is limited by precision of used machines and respective operation software. Nevertheless, the first tests show that product distortion and/or difference from entered contour can be substantially decreased, by more than 200 %. The resulting distortion comparing with ideal shape was below 1 %, even without any optimization. This result indicates that proper optimization process can improve the production of final products by AWJ to be competitive with classical machining tool production, and simultaneously, much lower amount of health hazardous and risky by-products like dust and fumes. Therefore, further research and development aimed at improving AWJ machining for composite materials is strongly recommended.