Ejector Deep Hole Drilling

• Ejector deep hole drilling is a double-tube process that supplies metalworking fluid internally and evacuates chips via the ejector effect, ensuring high bore quality.
• The process benefits from optimized drill-head geometries that reduce vortex formation and chip stagnation, resulting in lower coolant demand and improved feed force control.
• Experimental and SPH–DEM modeling reveal that internal flow dynamics and chip morphology critically influence stability and energy efficiency during deep-hole drilling.

Searching arXiv for the specified papers to ground the article and cite them accurately. arxiv_search.query({"search_query":"id:(Rupasinghe et al., 20 Mar 2025) OR id:(Rupasinghe et al., 4 Aug 2025)","max_results":5,"sort_by":"submittedDate","sort_order":"descending"}) arXiv search completed. Processing the returned records for the article. Ejector deep hole drilling is a double-tube deep-hole drilling process for bores with large length-to-diameter ratios, typically $L/D > 10$, positioned between single-lip gun drilling and BTA/STS systems by its fluid-supply and chip-removal concept. Metalworking fluid (MWF) is supplied inside the tool through the annular gap between outer and inner tube, passes through outlet bores in the drill head into the cutting zone, and the chip–fluid mixture is then entrained into the inner tube by the ejector effect. In contrast to BTA/STS, the process does not require a pressure head or sealing bush at the workpiece face, which permits operation on conventional machining centers while retaining internal chip evacuation, high material removal rate, and high bore quality (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

1. Process position within deep-hole drilling

Deep-hole drilling comprises single-lip gun drilling, ejector drilling, and BTA/STS drilling. All address bores with large $L/D$ ratios, but they differ principally in how coolant is supplied and how chips are removed. Ejector drilling is characterized by internal chip evacuation without the workpiece-face sealing infrastructure required by BTA/STS, and by lower inlet pressures than BTA systems; its process stability therefore depends strongly on internal tool geometry and MWF volume flow rather than pressure alone (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

Method	MWF supply	Chip evacuation / interface
Single-lip gun drilling	Internal through tool	External along grooves at bore wall
Ejector drilling	Through annular gap between outer and inner tube	Internal through inner tube; no pressure head at workpiece face
BTA/STS	Through annulus outside tube	Internal through tool; requires pressure head / seal

Typical applications include shafts, hydraulic cylinders, oil channels in large components, gear shafts, and valve guides, particularly where tight diameter tolerance, good roundness, low straightness deviation, and high surface quality are required. The cited studies consider a Botek type 62 tool with $D = 30\,\mathrm{mm}$ and workpiece materials including 42CrMo4+QT and, for blockage visualization, 18CrNiMo7–6 (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

A recurring practical distinction is that ejector drilling combines the internal chip transport associated with BTA-like surface-integrity advantages and the machine-side flexibility of conventional machining centers. This suggests why the method is often selected when dedicated deep-hole drilling infrastructure is unavailable but bore quality requirements remain stringent.

2. Tool architecture and transport mechanism

An ejector tool consists of an outer tube, an inner tube, and a drill head carrying inner and outer cutting inserts, several guide pads, MWF outlet bores, and chip mouths opening into the inner tube. In the 30 mm Botek type 62 configuration studied numerically and experimentally, the outer cutting edge lies at the periphery, the inner cutting edge is shifted by $180^\circ$ with minimal overlap at the center, and three guide pads are distributed circumferentially (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

Chip transport inside the head is driven by four mechanisms explicitly identified in the studies: bulk MWF flow from the outlet bores across the cutting zone into the chip mouths, the pressure gradient between the cutting zone and the low-pressure inner tube generated by the ejector effect, shear drag on chip surfaces, and rotational motion of the fluid induced by the rotating tool and outlet bores. Guide pads provide the self-guiding effect responsible for improved bore straightness and roundness, but they also introduce strong thermal and mechanical contact loads and require reliable lubrication (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

The hydrodynamic field near the cutting zone is strongly nonuniform. In the simulations, inlet velocity is imposed as $v_{\text{in}} = 5\,\mathrm{m\,s^{-1}}$, but the local flow near the cutting edges decays to about $1$ to $1.5\,\mathrm{m\,s^{-1}}$. This local reduction is significant because chip pickup occurs precisely in that region; inadequate local velocity or adverse flow direction reduces drag, delays chip entry into the chip mouth, and narrows the safety margin against clogging (Rupasinghe et al., 20 Mar 2025).

A common oversimplification is to treat ejector drilling as a process governed primarily by nominal coolant pressure. The reported results indicate instead that internal flow direction, chip-mouth inflow, and the spatial distribution of low-velocity regions are decisive.

3. Chip forms, stagnation, and blockage dynamics

The studies distinguish three representative chip types produced by the split cutting edge: a center chip from the center cutting edge, an inner chip from the inner part of the outer cutting edge, and an outer chip from the outer part of the outer cutting edge. Each type has its own characteristic axial length, chip formation frequency $f_s$, and number of chips per revolution $n_s$; these depend mainly on cutting speed $v_c$, feed $L/D$0, and total MWF volume flow $L/D$1 (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

The reported evacuation behavior is not uniform across chip classes. Outer chips are usually picked up quickly and transported into the chip mouth. Inner chips are more critical: in both the reference and modified geometries they can stagnate near the outer cutting edge, rotate about their own axis, and remain in the cutting-zone vicinity for a comparatively long time before full entrainment. In the SPH–DEM results, this rotation is attributed to velocity differences between chip corners and the resulting hydrodynamic torque. Center spiral chips are less frequent and slower to evacuate, and only part of their transport could be resolved within the reported simulation windows (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

A coherent vortex near the outer cutting edge and chip mouth is repeatedly identified numerically and, through prior visualization work cited by the studies, experimentally. Small chips can recirculate in this vortex or remain in near-wall low-velocity regions. If the evacuation rate falls below the chip generation rate, chips accumulate, MWF flow through the chip mouth is throttled, cooling degrades, mechanical loads rise, and drill breakage becomes possible (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

High-speed optical blockage analysis provides a more detailed failure sequence. In a transparent polycarbonate tube with water as visualization fluid, a longer outer-edge chip was first not removed, then subsequent chips accumulated on it, and finally a complete blockage formed at the outer chip mouth. The total time from first non-evacuated chip to complete blockage was approximately seven tool revolutions. Simultaneous telemetry showed a rapid rise in feed force, leading to the definition of a critical feed force $L/D$2 as the termination criterion for a stable process without chip blockage; the healthy feed-force range was reported as $L/D$3 (Rupasinghe et al., 4 Aug 2025).

These observations directly contradict the assumption that chip clogging is primarily a consequence of insufficient gross flow rate. The documented mechanisms are localized: vortex trapping, boundary-layer-induced stagnation, and chip–geometry interference at the chip-mouth entrance.

4. Experimental chip representation and SPH–DEM modeling

The numerical investigations are built on experimentally obtained chip geometries. Physical drilling tests with a transparent polycarbonate bore shell and high-speed imaging were used to observe chip formation mechanisms at the individual cutting edges, identify representative morphologies, and measure chip formation frequencies and lengths. Representative chips were then collected, scanned to generate 3D surface meshes, smoothed, reduced in triangle count, and exported as STL geometries for simulation. Each chip therefore enters the model as a rigid 3D body with realistic shape rather than as a sphere or ellipsoid (Rupasinghe et al., 20 Mar 2025).

The fluid phase is simulated by weakly compressible smoothed particle hydrodynamics and the chips by the discrete element method. In the WCSPH formulation, the fluid follows the Lagrangian continuity equation

$L/D$4

and pressure is closed by Tait’s equation of state

$L/D$5

with $L/D$6 for water-like fluids and $L/D$7 chosen as about ten times the maximum expected fluid velocity to keep density fluctuations below about $L/D$8 (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

Solid boundaries and chips are represented as triangular surface meshes. Fluid–solid interaction is handled by a modified Lennard–Jones potential, while chip motion follows Newton–Euler rigid-body dynamics with unilateral penalty contacts and friction for chip–chip and chip–wall interaction. The coupling is two-way: fluid pressure, viscous forces, and boundary interaction act on chips, while chips act as moving boundaries that alter local pressure and velocity fields (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

The head-focused simulations reported identical baseline fluid properties and kinematics for the compared designs: $L/D$9, $D = 30\,\mathrm{mm}$0, $D = 30\,\mathrm{mm}$1, $D = 30\,\mathrm{mm}$2, and $D = 30\,\mathrm{mm}$3. Initial particle spacing was $D = 30\,\mathrm{mm}$4 and smoothing length $D = 30\,\mathrm{mm}$5. Chips were inserted as fully formed rigid bodies with axial insertion velocities $D = 30\,\mathrm{mm}$6 for outer chips, $D = 30\,\mathrm{mm}$7 for inner chips, and $D = 30\,\mathrm{mm}$8 for center chips; chip density was set to $D = 30\,\mathrm{mm}$9 (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

The methodological limitations are explicit. Chip formation is not simulated, chips do not deform or break, thermal effects are neglected, turbulence is not modeled by RANS or LES, and in the head-only simulations the ejector effect is not explicitly resolved. The models are therefore intended for mechanism analysis and relative comparison of drill-head designs rather than absolute prediction of every process observable (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

5. Drill-head geometry, vortex control, and comparative performance

The numerical and experimental studies compare several head concepts: a reference industrial head, an earlier flow-optimized head emphasizing inward-directed flow, a narrowed chip-mouth design labeled modification II, and an extended chip-mouth design labeled modification IV. The central design variable across these variants is the geometry of the chip mouth and the way outlet bores direct flow from the cutting zone into the inner tube (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

Design	Geometric emphasis	Reported behavior
Reference head	Standard chip mouth and outlet orientation	Vortex near outer cutting edge; inner-chip stagnation; cluster formation near inner wall
Earlier flow-optimized head	More inward-directed flow near outer cutting edge	Reduced inner-chip residence time inside drill head
Modification II	Narrowed, more closed chip mouth	Reduced vorticity, but chip jamming and outer cutting edge failure in experiments
Modification IV	Extended, more open chip mouth; outer wall removed/shortened	Vortex shifted rearward; higher inner-tube velocity; smooth continuous evacuation

In the earlier reference-versus-optimized comparison, the first inner chip was released at about $180^\circ$0, stagnated near the cutting edge, and was not fully evacuated from the cutting zone until $180^\circ$1. For the flow-optimized head, by $180^\circ$2 that inner chip had reached near the end of the drill head, whereas in the reference head it was still within the head at $180^\circ$3 and the first chip remained inside even at $180^\circ$4. The primary reported effect was a qualitative reduction in inner-chip residence time produced by stronger inward flow toward the chip mouth and inner tube (Rupasinghe et al., 20 Mar 2025).

The later three-head comparison clarifies that not all vortex-reduction strategies are equally viable. In quasi-steady SPH results, velocity magnitude in the chip mouth was $180^\circ$5 for the reference head, $180^\circ$6 for modification II, and $180^\circ$7 for modification IV. Inside the inner tube, the reference head and modification II both showed $180^\circ$8, while modification IV reached $180^\circ$9. Reference-head trajectories displayed persistent stagnation zones and periodic chip clusters near the inner wall; modification II reduced early tube-side stagnation but weakened flow near the center cutting edge; modification IV produced smooth continuous evacuation with virtually no stable stagnation clusters once the flow was established (Rupasinghe et al., 4 Aug 2025).

Experimental drilling confirmed the geometric trade-off. Under standard parameters $v_{\text{in}} = 5\,\mathrm{m\,s^{-1}}$0, $v_{\text{in}} = 5\,\mathrm{m\,s^{-1}}$1, and $v_{\text{in}} = 5\,\mathrm{m\,s^{-1}}$2, the reference head showed minor chipping at the chip-mouth edge at $v_{\text{in}} = 5\,\mathrm{m\,s^{-1}}$3, modification II suffered repeated outer cutting-edge failure caused by chip jamming between the mouth edge and outer insert, and modification IV drilled stably up to $v_{\text{in}} = 5\,\mathrm{m\,s^{-1}}$4 with no visible wear on the head (Rupasinghe et al., 4 Aug 2025).

One misconception addressed directly by these results is that reducing vortex formation is sufficient. Modification II reduced vorticity but proved impractical because the constricted mouth increased chip–geometry interference. The reported evidence therefore favors hydrodynamically open chip mouths that preserve conveying capacity instead of merely suppressing recirculation.

6. Process window, coolant demand, and energy implications

The most explicit quantitative process benefit reported is the reduction in minimum MWF volume flow required for stable drilling. For 42CrMo4+QT at $v_{\text{in}} = 5\,\mathrm{m\,s^{-1}}$5 and $v_{\text{in}} = 5\,\mathrm{m\,s^{-1}}$6, the reference head required approximately $v_{\text{in}} = 5\,\mathrm{m\,s^{-1}}$7, whereas modification IV required $v_{\text{in}} = 5\,\mathrm{m\,s^{-1}}$8, a reduction of about $v_{\text{in}} = 5\,\mathrm{m\,s^{-1}}$9 or $1$0. At $1$1, the reference head required approximately $1$2 and modification IV $1$3, a reduction of about $1$4 or $1$5. In both cases, stability was defined by maintaining feed force below $1$6 (Rupasinghe et al., 4 Aug 2025).

Because pump power is approximated as

$1$7

a reduction in $1$8 at fixed $1$9 and $1.5\,\mathrm{m\,s^{-1}}$0 implies an approximately proportional reduction in pump power. For the $1.5\,\mathrm{m\,s^{-1}}$1 case, the flow reduction from $1.5\,\mathrm{m\,s^{-1}}$2 to $1.5\,\mathrm{m\,s^{-1}}$3 corresponds to $1.5\,\mathrm{m\,s^{-1}}$4, hence an energy-demand reduction of about $1.5\,\mathrm{m\,s^{-1}}$5 for the ejector system under that parameter set (Rupasinghe et al., 4 Aug 2025).

The design guidance emerging from the cited work is specific. Extended, open chip mouths improve conveying by shifting residual vortices away from the cutting front and raising inner-tube velocities. Outlet bores angled at approximately $1.5\,\mathrm{m\,s^{-1}}$6 toward the cutting edges and in the feed direction produce distinct jets directed toward the cutting edges, guide pads, bore bottom, and chip mouth, consistent with the SPH predictions. By contrast, overly narrow chip mouths can reduce vortices but also reduce local volume flow into the mouth and increase the probability of chip impact and jamming (Rupasinghe et al., 4 Aug 2025).

The broader process implication is that flow distribution and direction in the head can be more decisive than absolute flow rate alone. The earlier SPH–DEM study had already shown that even with an imposed inlet velocity of $1.5\,\mathrm{m\,s^{-1}}$7, local edge-region flow was only about $1.5\,\mathrm{m\,s^{-1}}$8 to $1.5\,\mathrm{m\,s^{-1}}$9, so the head’s internal geometry determines whether that available momentum is translated into chip pickup or lost in dead zones and recirculation (Rupasinghe et al., 20 Mar 2025).

7. Limitations, unresolved questions, and future directions

The current state of analysis remains intentionally simplified. Chips are inserted as pre-formed rigid bodies rather than generated by a coupled cutting simulation; there is no chip plasticity, breakage, or wear evolution. Thermal effects, including MWF temperature, viscosity variation with temperature, and guide-pad/bore-wall thermal loading, are not represented. Fluid treatment is isothermal and single-phase, without aeration or cavitation, and turbulence is captured only in a coarse-grained sense through WCSPH unsteadiness and artificial viscosity rather than explicit turbulence modeling (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

Validation is also bounded. The correspondence between simulation and experiment is strongest for qualitative and comparative observables: vortex location, tendency of inner chips to stagnate, relative improvement of evacuation under optimized flow guidance, and minimum volume-flow requirements for stable drilling. Direct quantitative validation against detailed pressure-loss distributions, chip-residence-time measurements, or full torque histories was not reported in the cited studies (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

Several next steps are explicitly identified. One is extension of the simulations to center-chip evacuation so that all chip classes are resolved equally. Another is inclusion of thermal properties, tool wear, and fuller multiphysics coupling, potentially including deformable chips and fluid–structure interaction. A third is further geometric refinement of the chip-mouth region near the outer cutting edge, which remained the critical stagnation site even after optimization. Beyond the drill head, future work targets ejector nozzle optimization inside the inner tube; cited hydraulic-efficiency estimates place current overall ejector-tool efficiency at $f_s$0, with a theoretical maximum around $f_s$1 without cavitation (Rupasinghe et al., 20 Mar 2025, Rupasinghe et al., 4 Aug 2025).

Taken together, the reported results establish ejector deep hole drilling as a process whose reliability is governed by the interaction of chip morphology, local head geometry, and internal hydrodynamics. The most plausible general implication is that future gains will come less from increasing nominal coolant supply and more from shaping the internal flow path so that chips are intercepted, accelerated, and conveyed before localized stagnation can evolve into full blockage.