Top Banner
Optimal Intelligent Scheduling When Machines Decide Improving Manufacturing Productivity with Data Analytics Vasu Netrakanti Optessa Inc.
9

When Machines Decide Improving Manufacturing Productivity with Data Analytics Vasu Netrakanti

Jan 29, 2016

Download

Documents

When Machines Decide Improving Manufacturing Productivity with Data Analytics Vasu Netrakanti Optessa Inc. Third Industrial Revolution: Digitisation of Manufacturing ( The Economist ). Perspective of optimization solutions provider Predictive analytics  Decision support - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: When Machines Decide Improving Manufacturing Productivity with Data Analytics Vasu Netrakanti

Optimal Intelligent Scheduling

When Machines DecideImproving Manufacturing Productivity with Data Analytics

Vasu NetrakantiOptessa Inc.

Page 2: When Machines Decide Improving Manufacturing Productivity with Data Analytics Vasu Netrakanti

(2)

Third Industrial Revolution: Digitisation of Manufacturing (The Economist)

Perspective of optimization solutions provider

Predictive analytics Decision support

Integrated with plant infrastructure

Real-time decisions: routing (what next), selection (which order) …

In an automated manufacturing environment, computers can help machines decide.

Page 3: When Machines Decide Improving Manufacturing Productivity with Data Analytics Vasu Netrakanti

(3)

OEM

VEHICLE ASSEMBLY PLANT 1

RED

WHITE

BLACK

BLUE

SILVER

RED

WHITE

BLACK

BLUE

SILVER

RED

WHITE

BLACK

BLUE

SILVER

RED

WHITE

BLACK

BLUE

SILVER

RED

WHITE

BLACK

BLUE

SILVER

RED

WHITE

BLACK

BLUE

SILVER

RED

WHITE

BLACK

BLUE

SILVER

RED

WHITE

BLACK

BLUE

SILVER

NON

SUNROOF

SUNROOFNON

SUNROOF

SUNROOFNON

SUNROOF

SUNROOFNON

SUNROOF

SUNROOF

4 DOOR2 DOOR4 DOOR2 DOOR

500K400K

METEOR

RED

WHITE

BLACK

BLUE

SILVER

RED

WHITE

BLACK

BLUE

SILVER

NON

SUNROOF

SUNROOF

2 DOOR

300K

COMET

VEHICLE ASSEMBLY PLANT 2

Optessa SUPPLIER ROOF LINER DIVISION

PRESS LINE MA1 PRESS LINE MA2

Case: Tier 1 On-demand Scheduling

High complexity – 50 variantsHigh volume for a small manufacturer – 25K parts/month“Broadcast” response

Page 4: When Machines Decide Improving Manufacturing Productivity with Data Analytics Vasu Netrakanti

(4)

Feature Data Examples / Note

Vehicle models COMET (300K): 2 DOOR

METEOR (400K, 500K): 2 DOOR, 4 DOOR

Trim COMET: GR5 – GR6, XR5 – XR7, ZR6

METEOR: GT0 – GT9, XT1 – XT9

Color RED,WHITE,BLACK,BLUE, SILVER

Other features/Attributes

Part #, Priority Date, Priority Time, Part Required Quantity etc.

Demand 22,798 units MONTHLY

Due Dates (Priority Date Time – 14 hours) Logistics lead time

BoM BoM Vehicle configuration to part

1St Set of Inputs: Demand

Page 5: When Machines Decide Improving Manufacturing Productivity with Data Analytics Vasu Netrakanti

(5)

Feature Data Examples / Note

Lines 2

Window 1 Month Flexibility required

Shift Pattern Line 1: Day shift,

Line 2: Day and Afternoon

Defines line capacity

Cycle time COMET ~ 83 seconds (based on 43/hour) and METEOR ~ 72 seconds (based on 50 per hour)

Cycle times can differ based on model, features and lines

Tool Setup / Changeover Tool change: 0.75 hours

Color change: 10 minutes

Demand – Capacity mismatch

Smoothing of overtime / idleness required

Line capacity Hard constraint

On Time Delivery High priority

Minimize Changeovers Medium Priority Door – Roof – Color changeovers

Line Restriction High Priority COMET model parts imited to line 1

2nd Set of Inputs: Production Capacity and Rules

Page 6: When Machines Decide Improving Manufacturing Productivity with Data Analytics Vasu Netrakanti

(6)

• Total Orders: 22,798• On Time: 5,653• Late: 13,324• Unscheduled: 3,821

• Total Orders: 22,798• On Time: 22,798• Late: 0• Unscheduled: 0

Results: Current and Optimized Scenarios

CurrentFollow OEM sequence

OptimizedSmoothed OT: 2 hours/dayMinimize changeovers

Page 7: When Machines Decide Improving Manufacturing Productivity with Data Analytics Vasu Netrakanti

(7)

Case: SC Fab Real-time Yield Management

Page 8: When Machines Decide Improving Manufacturing Productivity with Data Analytics Vasu Netrakanti

(8)

Case: Auto OEM ASRS Real-time DecisionsThe flow of painted bodies from the Paint Shop to the ASRS, and then to TCF can be represented through a flow diagram indicating the key vertices and arcs.R30/31: R30/31 is the set of readers on the two lane buffer at the exit of paint (post polish stage). After painted bodies cross the R30/R31 readers, there is no divergence until they reach the next node BT1100. All orders between R30/31 and BT1100 are the look-ahead available for RTS.BT1100: BT1100 is the skid exchange station where the painted body is transferred from Paint skid to Logistics skid. At this point, the decision of storing/bypassing/wax/set out of a painted body is conveyed to the further PLCs.PTT01: PTT01 is the turntable at the entrance of the ASRS. The orders diverge into Lane 1 or Lane 2 of the ASRS after PTT01.PTT02: PTT02 is the turntable at the exit of the ASRS. Orders retrieved from the ASRS converge at this turntable before being passed to the TCF.TT02: This is the turntable which routes LTVs to the LTV TCF buffer. The U202 and W201 bodies continue past TT02.TT03: This is the turntable which routes U202s to the U202 TCF buffer. The W201 bodies continue past TT03.

TT04: Turntable routing W201 to W201 TCF buffer.

At 60 jph, 3 decisions required every minuteRouting: which ASRS cell to store painted bodySelection: sequence outgoing jobAssignment: which sales order for painted body

Page 9: When Machines Decide Improving Manufacturing Productivity with Data Analytics Vasu Netrakanti

(9)

Challenges

Reliable, accessible, timely data

Complement or supplement existing business processes?