About the project
AdvisorAI is the software part of a PhD dissertation on how the physical and mechanical properties of materials should be taken into account when CNC control programmes are prepared.
Research topic
Development of principles for taking into account the physical and mechanical properties of materials when developing automatic control programmes (G-code) for CNC machines.
Specialty 05.02.08 – Mechanical engineering technology. Tashkent State Transport University.
Why it matters
Cutting data chosen without a quantitative account of the material either wastes machine time or overloads the tool, the spindle and the surface of the part. For aerospace alloys such as Ti-6Al-4V and Inconel 718 the cost of such errors is especially high. Handbook tables and CAM libraries give recommended ranges. They do not give the physical state of the process or the reason a value was chosen.
Research question
How can the transition from measurable material properties to the numerical parameters of a control programme be made explicit, reproducible and verifiable?
Scope
- Milling on three-axis machining centres
- ISO 513 material groups P, M, K, N and S
- Carbide tools with TiAlN coating, flood coolant
- Planning stage, before the programme reaches the machine
The four principles
Each principle is stated verbally and as an explicit equation, and each is one stage of the selection algorithm.
P1 · Speed and feed
Cutting speed and feed are placed within the band recommended for the material and tool, then reduced until spindle power stays below 85 % of rating and the homologous temperature of the cutting zone stays below 0.60 of the solidus.
P2 · Depth of cut
The axial depth is the least of the geometric requirement, the depth at which tool deflection stays within tolerance, and 80 % of the regenerative-chatter limit.
P3 · Tool life and tool changes
Tool life comes from the extended Taylor equation, with feed and depth in the denominator. Tool changes are placed so that no edge exceeds its predicted life.
P4 · Feature coupling
Regimes are corrected for the machined feature: the roughness requirement limits the feed, thin walls limit the depth, and deep cavities in low-conductivity materials limit the speed.
Refinement P1-S. Analysis of the full space of admissible regimes showed that placing speed and feed symmetrically in their bands gives inefficient regimes. The tool therefore applies the sequential rule: depth first, then feed, then the speed for a target tool life. For Ti-6Al-4V this predicts 31 % shorter cutting time with 27 % longer tool life than the symmetric rule. See the results.
Scientific contribution
- An explicit, traceable link between material properties and the S, F and depth words of the programme.
- A constrained selection that reports the physical limit governing each regime, which differs between materials on the same machine.
- A dimensionally consistent two-zone temperature model with a homologous-temperature criterion referred to the solidus.
- A multi-tooth force model and a corrected extended Taylor equation.
- A catalogue of implementation errors in physics-based cutting-data software, with the checks that detect them.
Honest status
All quantitative results are calculations with provisional model coefficients. An experimental programme according to ISO 8688-1 has been designed to determine these coefficients and to measure the prediction error. It has not yet been carried out. Until then, the numbers indicate trends and relative differences, not validated values.
Team
H.Khakimov
PhD researcher. Research design, models, software and analysis.
Scientific supervisor
To be added by the project owner.
Institution
Tashkent State Transport University