Sunday, July 21, 2019

Can complex high end machining be science than art?



My favorite challenge: High end complex machining. Can it be a science rather than art?

Metaphor: ‘3D print skilled machinist like 3D printed parts’

Summary:  When I travel in US I see American pilots, in China Chinese pilots and in India Indian pilots. So, it is possible to work around a complex system like air plane cockpit and air traffic system without importing talent from west. Can this be done to machine a part which has a tolerance smaller than human hair? Can there be autopilot program running a CNC machine? Can we train an operator in simulation room before he runs an expensive machine to produce expensive parts? Will this solve the problem and create a level playing field between West and East? or can we reduce the element of memorizing and facilitate real time guidance?

Before 





  



Potential solution














After 














Solution-  there are number of ways to feed information. For repeat scenarios there could be frequently asked question bank and solutions. For a unique situations there can be expert guidance by sharing or looking at same interface and making a decision before stopping the machine. This will reduce the time needed to seek, search help and build confidence in the worker.



Prototype












In the existing set up one desk stop, couple of cameras can be used to start with. This will create a context like a real set up and help understand what final product needs to take into account.

Day one experiment- Install a camera at workstation and monitor remotely from a desk top. Use blue tooth headphones and existing machine or desk top simulator. Arrange an expert phone call to worker at a chosen station and start feeding instruction step by step based on input from the camera. When a critical decision making point comes intervene and ask worker to change machine parameters. Do this exercise for couple of parts in existing set up on a old machine and less expensive or already scrapped part. This is simplistic version however the real solution need interface with machine and exact position and status of the program needs to be confirmed before any input can be given and this is where the smart glasses who can talk to desk top simulator can add value.

Hypothesis testing-  Learning and remembering complex process can be replaced by real time feeding of instructions to ensure learning is continuous and this will not slow down the worker.

Measure of success- Worker feedback on whether he finds the real time guidance useful and can he produce a part quicker with good quality.

Cost of experience-  Day salary of a worker and expert. Other costs are fixed whether we do this pilot or not.

No comments:

Post a Comment