• Manufacturing environments vary greatly with respect to their process structure — the manner in which material moves through the plant The decision of what process structure to use depends on expected demand and volume
    • Job Shops - Small lots are produced with a high variety of routings through the plant. Flow through the plant is jumbled, setups are common. This setup allows for more customizability and is more suited for products with high customizability, low volumes, and slower rates.
    • Disconnected Flow Lines - Product batches are produced on a limited number of identifiable routings (i.e., paths through the plant). Although routings are distinct, individual stations within lines are not connected by a paced material handling system, so that inventories can build up between stations. Used to make similar items on a repeat basis, where the volumes are not high enough to allow for a connected line flow
    • Connected Flow Lines - The classic Assembly line. Product is fabricated and assembled along a rigid routing connected by a paced material handling system. Suited for a narrow product range where production can be automated to produce high volumes of standardized at a faster rate
    • Continuous Flow Processes flows automatically down a fixed routing, no longer divided into discrete units. Suited for very high volume, commodity products.

Push and Pull

  • In a push production control system, work releases are scheduled in advance based on customer due dates (demand), but must respond to changes in the plant

    • Production is triggered by a schedule.
    • Push systems control throughput and observe WIP
    • As soon as work on a part is complete at a workstation, it is “pushed” to the next workstation.
    • All machines work as long as they have the parts.
    • It is typically made to order
    • A push system requires higher WIP levels to attain a given throughput level.
    • Push systems are not flexible because of being prone to high WIP and low flexibility (because design changes cannot be incorporated in time.)
    • Setting release dates in a push system involves looking at capacity, which is not as simple. Push systems are not easily observable and hence harder to optimize
    • In a push system, WIP levels are independent of one another which makes cycle times more variabile.
    • Push systems are easier to plan for.
  • In a pull production control system, work releases are authorized based on changes in the plant, but must be adjusted for customer due date.

    • Production is triggered by demand.
    • Pull systems control WIP and observe throughput
    • An operator requires both parts and an authorization signal to work
    • *It is typically made to stock *
    • The main benefit of a pull system is that there is a limit on the maximum amount of inventory in the system. Any disruptions do not cause WIP to grow beyond a predetermined level and gives more predictable flows (cycle time)
    • Pull systems are flexible because they can easily adjust to changes by reducing costs due to these changes.
    • Pull systems have better timing of work releases
    • Low WIP levels (independent of how this is achieved) also promote more effective quality control since we only need to inspect fewer jobs.
    • We can also perform work aheads when plant conditions are favorable (up to a certain limit of reasonability. For example, we do not want to work on tasks with speculative requirements).
    • Pull systems are easily observable all we have to look for is the WIP.
    • Pull systems introduce negative correlation between WIP levels at different stations. This dampens variability.
  • In a hybrid approach, the factory can be divided into push and pull segments connected via push-pull interfaces.

    • The choice of where to put push-pull interface is important
    • The closer the push-pull interface is to the customer, the less lead time we have at the cost of flexibility.
    • The options for positioning the push-pull interface are strongly affected by the process itself.
    • The economics of push-pull interface placement are affected by how the product is customized as it progresses through the system
      • Having fewer end items may make it feasible to put the interface at the finished goods.
      • Having many items makes it more feasible to put the interface at the start 1
    • In a system in which the product becomes increasingly customized as it progresses down the line, moving the push-pull interface upstream can reduce the amount of safety stock that needs to be carried as protection against demand variability
  • CONWIP (Constant Work In Process) is a protocol where a new job is introduced to the line each time a job departs. CONWIP results in a WIP level that is very nearly constant. This assumes two things (subject to extension)
    • The production line consists of a single routing, along which all parts flow.
    • Jobs are identical so WIP can be measured in units.
  • Compared to Kanban, CONWIP is easier to control since we do not need cards for each station.
    • In Kanban, cards are part number specific. In CONWIP, cards are line specific (cards are instead matched against a backlog). CONWIP scales much better when there are many parts.
    • CONWIP requires steady volume but is more robust than kanban compared to product mix.
  • A CONWIP system resembles a closed queueing network where customers never leave.
  • In CONWIP, bottlenecks are seen more easily as places where WIP accumulates
  • CONWIP subjects operators to less pacing stress since authorization for intermediate materials is implicitly provided. 2
  • We can make use of a Mean Value Analysis Model where we examine the average behavior assuming all other jobs are evenly distributed according to the average behavior.

    • Each station is treated as if they behaved like a 1 queues. The average time a job spends in station in a system with jobs remaining is given by the following (using MVA)

  • Law of CONWIP Efficiency. For a given level of throughput, a push system will have more WIP on average than an equivalent CONWIP.

    • For a given level of throughput, a push system will have longer average cycle times than an equivalent CONWIP system.
  • Law of CONWIP Robustness. A CONWIP system is more robust to errors in WIP level than a pure push system is to errors in release rate.

    • If we tried to plot profit as a function of WIP level, we find that the curve is relatively flat

Shop Floor Control

  • Shop Floor Control encompasses a variety of functions
    • Material Flow Control - we decide which jobs to release into the factory, which parts to work on at individual workstations, and what material to move between workstations.
    • WIP Tracking - involves identifying the current location of parts in the line.
    • Status Monitoring - surveillance of other parameters in the line besides WIP position (i.e., machine and staffing status)
    • Throughput Tracking - measures the output against an established production quota and/or customer due dates which dictates if we need overtime.
    • Statistical Throughput Control - allows us to track progress towards meeting the production quota.
    • Capacity Feedback - ensures that high level planning modules are consistent with low-level execution. It also lets us estimate capacity.
    • Quality Control - allows us to address reworks which require parts and may add delays.
    • Work Forecasting - anticipates the expected amount of work needed. It should consider the effects of delays from rework or yield loss.

Links

Footnotes

  1. Otherwise we have to deal with combinatorial explosion since we would need appropriate inventory for many end items.

  2. Still, the first operator in the line has to fill voids to meet WIP. This is unavoidable just as in a regular kanban line.