Issue 70, December 2006

Using BlockSim 6 for Planning Just-In-Time Ordering of Spare Parts

Just-in-time (also known as lean production or stockless production) is an inventory strategy that aims to reduce operation costs by reducing carried inventory and its associated cost. Inventory of spare parts is necessary to maintain operations and avoid long downtime, but can also be a major expense. Modern business thinking views inventory as incurring costs instead of adding value. Lean business tries to eliminate inventory that does not add value to the product. The just-in-time (JIT) philosophy is about having the right part at the right time, at the right place and in the sufficient amount. Good predictive and statistical analysis can transform inventory from a burden to a lifesaver.

Just-in-time ordering of spare parts provides many benefits, such as:

• Savings for maintaining the parts.
• Avoiding quiescent types of failure modes (degradation while on the shelf).
• Avoiding acquiring many parts and then discovering that the manufacturer is no longer in business or is retiring that type of part.
• Smaller inventory floor and warehouse space.
• Less overhead cost.
• Reducing the amount and complexity of tracking and control otherwise needed for large inventory.
• Freeing up capital that can be used for growth of the company or value-added activities.

This article presents methods of using BlockSim 6 for comparing different inventory policies to achieve a JIT inventory strategy based on the understanding of the repair and spare parts needs for supporting an operation.

1- Example
In this article, we will use a simple example to illustrate the proposed methodology. It can be expanded to analyze more complex systems from real applications in manufacturing, process or communication systems. The following reliability block diagram (RBD) describes a simple operation.

Figure 1 - RBD for a simple operation

The next table lists the failure and the repair distributions of the different blocks in the above RBD. (Blocks can represent parts, equipment, subsystems, etc.)

Table 1 - Failure and Repair distributions for the different blocks in Figure 1

The following figures show how the failure and the repair distributions can be set for a block in BlockSim 6.

Figure 2 - Entering the failure distribution and repair distribution for a block

In order to evaluate an inventory strategy, we need to consider the costs of maintaining spare parts in the inventory. The costs associated with each type of part include the direct cost per item of the part; the indirect cost, or cost per item per unit time of maintaining each part in the inventory; and the cost of obtaining an emergency spare part if none is present in the inventory. In this example, we use the following costs.

Table 2 - Breakdown of costs of maintaining the inventory of spare parts

Planning the optimum inventory strategy is a very difficult mathematical and managerial problem. It is a very rich area of Operations Research and inventory problems could range in difficulty from easy to intractable or unsolvable. Simulation can be a very adequate and simple approach to evaluate different inventory strategies and to assist in the planning of a "close to optimum" strategy without spending a great deal of effort and time using advanced analytical stochastic calculations (which might demand a high level of expertise) to determine the "optimum" inventory strategy.

2- Estimating the Required Number of Parts
The first step in this analysis is to estimate the number of failures (for each block) that would be expected to occur during a certain operation duration (mission time) of interest. For this example we use a period of 3 years (26280 hrs). The number of failures during the period of interest depends on the failure and repair characteristics and can be estimated using simulation. The next figure shows the simulation settings for the system.

Figure 3 - Simulation settings to simulate 3 years of system operation

After running the simulation, we can obtain many results, reports and plots in BlockSim 6. For example, the System Overview report, shown in Figure 4, can be obtained by clicking the Details button in the Maintainability/Availability Simulation console shown in Figure 3 (which becomes available after you have run the simulation). This is just one of the reports available in the Simulation Results Explorer.

Figure 4 - System overview report example

We note that the availability of the system is expected to be 0.999637 for the duration of 3 years.

Another report of interest for this study is the Block Summary report, located in the Blocks folder in the Simulation Results Explorer..

Figure 5 - Block summary report (partial)
[Click to Enlarge]

From this report, we can obtain the expected number of failures (Expected NOF column), which we can translate to number of spares needed. The following table shows a summary of the total (rounded up) number of spares needed for each type of block.

Table 3 - Summary of the Total Expected Number of Spares Needed

In this initial simulation, the assumption was that the system has access to infinite spares at all times (default setting in BlockSim 6 for spare parts). We used this initial simulation just to get an estimate for the spare parts needs of the system.

3- Setting Up Different Inventory Strategies
Now that we know the required number of parts, we can study different spare part inventory strategies.

Inventory Strategy 1
A very conservative strategy would be to have all the estimated required parts in stock from the beginning of operation. We set up the spare pool (inventory) policies for blocks A, B and C by creating a Spare Pool Policy (select Add New Resource then Add Spare Pool Policy from the Project menu.) The strategy is set up as follows.

Table 4 - Summary of settings for Inventory Strategy 1

Figure 6 demonstrates how to enter these settings using block A as an example.

 Figure 6 - First inventory strategy for A

Note: In the above figure, we also used an Emergency Spare Provisions policy. If we do not have a way to obtain new external spare parts once the inventory runs out of parts, then blocks that fail and run out of spare parts will stay down for the rest of the simulation time and that would bias the availability estimates. Therefore, the emergency spare provisions policy is set up as a backup for the cases when the inventory is completely starved (when the number of actual failures turns out to be larger than the average expected number of failures) and there is no way to obtain more parts.

Once the spare pool policies have been created, we specify the appropriate policy to associate with each block in the diagram by selecting the policy from the list of available policies in the Spare Part Pool section in the Maintenance tab of the Block Properties window. The next figure shows how to associate a block with its spare pool policy.

Figure 7 - Specifying the inventory policy for a block

After specifying the appropriate policy for each of the blocks in the diagram, we run the simulation to obtain availability metrics and cost estimates for maintaining the inventory. One of the simulation reports in BlockSim 6 is the Pool Summary, located in the Spare Pools folder in the Simulation Results Explorer. This report lists, among other things, the Total Cost of maintaining the inventory.

Figure 8 - The pool summary report (partial)
[Click to Enlarge]

For this inventory strategy, the estimated total cost for maintaining all the pools is $1,159,196. The estimated mean availability can be obtained from the System Overview report or by using the Simulation Quick Calculation Pad (QCP). The estimated mean availability is 0.999637.

Inventory Strategy 2
In this strategy, we study a possible JIT inventory strategy. We start with one spare part in the inventory (instead of all the needed parts as in Inventory Strategy 1) and use a condition-based rescheduling policy to acquire spare parts when the existing spare part is used for a repair. Compared to Inventory Strategy 1, this strategy will cut down greatly on the cost of inventory but could conceivably reduce the availability level of the system, since the inventory could possibly not have a spare part available when needed. The strategy is set up as follows.

Table 5 - Summary of settings for Inventory Strategy 2

Figure 9 demonstrates how to enter these settings using block A as an example.

  Figure 9 - Second inventory strategy for A

We use the approach described for Inventory Strategy 1 to specify the appropriate policy for each block, run the simulation and obtain the availability and inventory cost estimates. The estimated mean availability is 0.999637 and the estimated cost for maintaining all the pools is $40,654.

Inventory Strategy 3
In this strategy, we study another possible JIT inventory strategy. In this case, we start with one spare part in the inventory and use a periodical rescheduling policy to acquire spare parts. The period of rescheduling can be determined using the expected number of spare parts needed in the period of 3 years summarized in Table 2. By dividing the period of interest (3 years) by the number of needed spares, we can determine an appropriate part restocking frequency.

Table 6 - Determining the Appropriate Part Restocking Frequency

The strategy is set up as follows.

Table 7 - Summary of settings for Inventory Strategy 3

Figure 10 demonstrates how to enter these settings using block A as an example.

Figure 10 - Third inventory strategy for A

We use the approach described for Inventory Strategy 1 to specify the appropriate policy for each block and run the simulation and obtain the availability and inventory cost estimates. The estimated mean availability is 0.9851195 and the estimated cost for maintaining all the pools is $94,118.

4- Comparing Different Inventory Strategies
Many other inventory strategies could be evaluated. For example, variations of Inventory Strategy 2 could be used (using different initial stock levels, different numbers of ordered parts, etc.). Combinations of different strategies could also be used (e.g. mixing on-condition based restocking and periodical restocking).

The following table presents a summarized comparison of the different inventory strategies discussed in this article. The table shows the inventory costs incurred and the mean availability level that would be expected if the inventory policy is implemented.

Table 5 - Comparison of the different inventory strategies

Based on the above comparison, we decide that Inventory Strategy 2 is the most suitable strategy due to its significant cost savings (about $1,118,542 in savings compared to the cost of Inventory Strategy 1) and the fact that it also preserves an acceptable level of operation availability. (In this example, there is no difference between the availability level obtained by the first and second inventory strategies, whereas the third strategy causes a significant decrease in availability and does not provide the same cost savings offered by the second strategy.) For this example, Inventory Strategy 2 can be considered as a very good JIT inventory strategy that avoids carrying more parts than needed at any given time.

5- Conclusion
In this article, we presented a simple approach for assessing different inventory strategies to plan a just-in-time inventory policy. We used a simplified example to illustrate the methodology and the related concepts; however, we can build on this concept and add more details (e.g. delays for obtaining parts). The reader is invited to explore the other features of BlockSim's Spare Pool Policy feature that can help to construct models that mimic the particulars of their real inventory. Considering additional factors, such as the cost of loss of opportunity, would enable us to create a holistic view of the system and study the effect of the inventory strategy on the global Life Cycle Cost or profitability of the system.

(*): Used in Inventory Strategy 1 only, in Section 3. The cost of obtaining an Emergency Part is not related to the Fixed/Hourly inventory costs

Copyright 2006 ReliaSoft Corporation, ALL RIGHTS RESERVED