Bajaj Auto Ltd – Waluj. (2nd Runner-up)

Project: -Bajaj Auto Ltd-Waluj
Objective: Easy & reliable JH/PM and Diagnosis with minimum Manual Intervention
Focus: Strengthening JH/PM Activities using Artificial Intelligence
Purpose & Back Ground:

For any business sustainability- we need to minimize investment, optimize the manpower and same time speed up the outcomes. In line with these goals, BAL CVD Maintenance abandoned traditional thinking of improving skilled man power & come out with innovative thinking:

  1. Why do we need skilled man power to inspect same points repeatedly to maintain Equipment health?
  2. Why specialized skilled man power is needed to diagnose the interruptions?
  3. Why not to eliminate inspection & increase in-house skill of our persons for restoration only.

Answer was - why not to build Maintenance Man’s thinking ability into the equipment itself to take decisions or guide the operator itself to take care.

This thinking can benefit us both financially as well as non-financially in saving PM time, Diagnosis time & man power substantially, with indirectly huge Productivity loss due to limited skill of in-house persons, particularly at the time of peak production demands in odd hours of seasons or times. Ultimately Increasing reliability & Uptime of equipment.

Solution was achieved through Present condition deep study, taking challenges as opportunity to improve & proposed new concept-for JH/PM & diagnosis:paint2

To eliminate errors in manual JH/PM inspection and recording, 1st step was to integrate all check sheets into soft & through intelligent unit achieve auto scheduling of pending points, Abnormality & spares reports.
paint3

Next Step was to use max existing sensors to indicate deterioration/malfuction in cycle, to operator as Digital cock pitpaint4

paint5

paint6

 

paint8

 

New Intelligent Method for Diagnosis, eliminates skill

Total Sensors 143                       No Need of high skill to diagnose & search exact interruption

In this method, Each sensor location & its physical picture is linked with its computerised logical address. As sson as the interuption / malfuction of Sensor is triggered in a cycle, its remedial action & location along with picture is displayed to the operator for immediate corrective action by himself.

paint9

paint10

paint11
Pain:

  1. Any single interruption if not detected & attended within 48 seconds lead to loss of one vehicle
  2. Most of the time interruptions are resettled by Operator/ Line supervisors without informing to Maintenance
  3. Interruptions are recorded manually by visiting each machine panel every shift to take corrective action.
  4. Inspection of overhead sensors are unsafe to check daily/weekly.
  5. Hence difficult to reduce minor stoppages occurrences due to such interruptions, affecting OLE.

Idea:  Make all shop equipment-Diagnosis & Sorting Easy & Quickly
Presently all such interuptions are manually recorded in register & latter on are studied for failure reasons. This leads to skipping of root cause actions with direct loss of prodution, inspite of having AND_ON llike system.
Following trend shows how it is taking long time to eliminate such interptions on various lines.

Picture51 Picture52

Idea & Concept: Interlinking of each equipment & each sensor with centralized SCADA systemPicture53

 

paint12

On clicking the equipment, Failed equipment plan view with failed part sensor flashing is shown with its graphical picture location as shown below. Eliminating need of any skill to Diagnose the fault.
Picture54

paint13

paint14

Leave a Reply