LEAN Basic Tools Concepts - Part 3

Lean Tools: Generic Pull System
Purpose - To place a limit or cap on the maximum number of things or work in process (TIP or
WIP), so that lead times are known and predictable (see Little's Law, slide 14). Then
apply improvement tools to reduce WIP/TIP by eliminating the effects of variation
(traditional Six Sigma) and batch size.

When to Use It - Whenever lead times are critical to satisfy customers and when NVA cost is significant
compared to VA cost.

How to do it -
Part 1: Determine WIP Cap (Maximum amount of work or things that should be in process at any one time)
Step 1: Determine Current Total Lead Time (TLT)
•Option 1: Track individual items through the process, measure lead time
•Option 2: Use Little’s Law to get an average lead time
Step 2: Determine current Process Cycle Efficiency (PCE, see slide 13)
Step 3: Identify Target PCE
* Level at which the process should be operating based on “world class”
(high end efficiency
* If current PCE is:
<< low end, multiply current PCE by 10 to get target PCE
< low end, use low end figure as target PCE
= or > low end, use high end as target PCE
>> high end, move toward one step flow
Step 4: Calculate target lead time for the process.
* Lowest process cycle time with the current process
Target Lead Time = VA time for critical path/Target PCE
Step 5: Calculate WIP Cap
* Maximum WIP allowed within the process at any given time
WIP CAP = Target Lead Time x Exits

Part 2: Releasing Work into the System (Usually current WIP level is significantly greater than WIP CAP. Therefore, reduce current WIP and release work into the process to match the exit rate.)
Step 1: Count the WIP in the process
Step 2: Determine if you can release work or not:
* If the WIP > WIP CAP, do not release any more work
- If this will harm customers, options are to temporarily increase capacity
to lower WIP or perform a triage (next slide) of current WIP to see if
some work can be set aside to make for new work
* If WIP < WIP Cap, release enough work to get to the WIP CAP
Step 3: Identify how you will know when more work can be released into the system. CAUTION: As the PCE of a process reaches world class levels, the effects of variation are magnified. Do not reduce TIP or WIP too much without addressing issues of variability or a process could be starved for work creating a constraint.
Step 4: Create a triage system for determining the order that future work is released into the system
•Option 1: First in-First Out (FIFO)
•Option 2: Triaging or working on highest potential items first. Set up a criteria for rating or ranking new work requests as to their potential for your company (i.e. what’s important to you in light of VOC).
•Option 3: If there is capacity for parallel processing, shift work from overloaded process steps or adding/shifting resources
Step 5: Set up processes for maintaining the Generic Pull System
•Identify the person who will release work into the system
•Develop alerts, signals or procedures that will let this person know when WIP has fallen below CAP, e.g. flags, emails, alert cards, etc.
•Train people in the new procedures
•Develop a transition plan to cover the current high WIP state to future WIP Cap state
•Implement and monitor results; adjust as necessary

Replenishment Pull Systems

Purpose - To eliminate shortages or over stocking of supplies by creating a system where items are automatically replaced as they are used up

When to Use It -
When in-process or end-item products, supplies or consumables (or any item for which shortages or stock outs are not acceptable), meet the following criteria:
•Usage of the item is repetitive
•Demand for the item is inconsistent
•Stocking levels have a significant impact on service level for internal or external customers

A Replenishment Pull system should never be installed without a Generic Pull system in place

How to do it -
Step 1: Determine Work Demand Rate (DMD)
•Average weekly or average daily usage
- Based of history or forecasting (backlog) or a combination
- Caution: Historical data may not reflect changes in future business needs
•Recalculate frequently
•Handling seasonality
-Resize monthly if demand changes past a hurdle rate, e.g. if demand changes > 20%
-Use historical data or forecasts to determine signals that mean a resizing is necessary
-Forecast window should be at least the average Lag Time Weighted by Volume to account for lag times between demand and order receipt
-Use historical/future demand weighting tools to smooth ramp ups/downs
* Larger up swings >> higher weighting in forecast
* Smaller upswings >> lower weighting on forecast

Step 2: Determine Replenishment Lead Time (LT) and Order Interval (OI)
•Replenishment Lead Time (LT): The time from when a part or supply has been consumed until new parts/supplies are received
- For purchased items such as supplies etc. add together: Time to generate a purchase order + Supplier lead time + Transportation time + Receiving inspection/stocking time
- For manufactured items, add together: Time to generate the work order +
Total process lead time + Receiving/inspection time
•Order Interval (OI): Can be expressed as either the interval between orders or the order quantity to be purchased
- Changing OI allows for trade offs between transactions, capacity and inventory

Step 3: Determine Optimal Safety Stock (SS) Level
•Safety Stock (SS) can be calculated in a number of ways. This method is generally preferred. Key Assumption: The demand profile is normally distributed
Safety Stock = s service level X (LT) to the power b
Where
• s service level = desired service level (stock out coverage) = number of standard deviations, relative to the mean, carried as safety stock. For example:
- Service Level = 1 means that one standard deviation of safety stock is carried, and on average, there will be no stock outs 84% of the time
- Service Level = 2 means that two standard deviations of safety stock is carried, and on average, there will be no stock outs 98% of the time
•Lead Time (LT) = Replenishment Lead Time
• b = a standard lead time reduction factor (generally set at 0.7)

No comments: