Selecting Spare Parts for Additive Manufacturing

Jun 04, 2019 • Features3D printingmanufacturingSpare PartsAdditive Manufacturing

Additive Manufacturing has long been touted as a major solution to be embraced if we are to resolve many of the challenges of the service supply chain. However, identifying which parts are suitable for such an approach is a key part of adopting such a strategy as Atanu Chaudhuri, Associate Professor, Aalborg University explains... 

An industrial products manufacturer has a large portfolio of spare parts in its portfolio. It has 10 year service contract with its customers implying that the company must be able to deliver spare parts on time for 10 years to its customers even though the product may longer be produced. Thus, the company has to maintain inventory of some of those spare parts. Such inventory costs run into millions of dollars. Many of those spare parts are produced by suppliers. Demand for those spare parts are low and unpredictable.

On Monday morning, the spare parts manager received a call from a customer in another continent “ Mr. Paul, our equipment is out of action as it cannot be started in the morning shift. My maintenance guys confirmed we need to replace a part, supplied by you and we do not have it in inventory. I am sending you the part number. Please deliver it within three days otherwise there will be heavy loss.” Mr. Paul checked the part number in the ERP and found that they also do not have in stock.

He found out who the supplier is and called him up. The supplier replied saying that they cannot make only one piece. They will need an order size of at least 20 and it will take 1 week. What will Mr. Paul do? He remembered that the Vice President (VP) of Supply Chain was talking about 3D Printing of spare parts in the last meeting. Can the part be printed? How will Mr. Paul figure that out?

He refers to the presentation by VP- Supply Chain and finds the name of a company called 3D Hubs. Mr. Paul checks out the website of 3D Hubs and found that he can upload the drawing of the parts and get an analysis for printability along with a quote. He uploaded the drawing, got the quote for delivery within three days and went to his boss.

His boss, Mr. James said “Are you crazy? How will I be able to justify that price? Also, have you noticed that they are not printing our part using our specified material? This will never work.” Mr. Paul tried to argue “This is a very strategic customer for us. If they are not satisfied and cancel the service contract, it will be a big trouble for us. We can say that they can run their equipment with it until an actual replacement spare part is produced and delivered by the supplier. I can talk to R&D about the approval for the deviation in material. It doesn’t seem to be a big change.” Mr. James said “You are taking a big risk, Paul. If you want to try, go ahead.”

In the end, everything turned out well. That part was produced using additive manufacturing (AM) and delivered on time. Paul received appreciation for his efforts. VP- Supply Chain was very happy and said to Paul. “Now, I want to evaluate our entire spare part portfolio and develop a systematic process of evaluating which of those parts can be printed and which cannot. It will be great if you can evaluate our portfolio and present in our next month’s meeting.”

Paul has 10 years of experience in spare parts but he does not know much about 3D Printing. How can we help Paul? I am sure that there are many people like Paul in large industrial manufacturers, who need similar help. Lets try to outline a steps-by-step process for selecting spare parts suitable for additive manufacturing.

Step 1: Identify the objectives
Some potential objectives for producing spare parts using AM can be:
• Supply risk reduction;
• Lead time reduction;
• Inventory cost reduction;
• Ensuring local content;
• Minimizing loss of production;
• Reducing carbon foot print across life cycle.

Companies can decide the most relevant ones from those and provide relative importance of those using a method called analytic hierarchy process. Instead of directly assigning importance weights to each objective, experts within the company can make pairwise comparison of the objectives using a scale (for example 1-3-5-7-9) and derive the importance weights.

Step 2: Identify the factors to be used for screening the spare parts for AM suitability
Some factors, which can be used for screening the spare parts for AM suitability can be as follows: It is important to note that the company should have data on the following factors to be considered for screening.

If such data is not available or available in different IT systems or in physical form ( i.e. drawings with dimensions and material specifications), it will be difficult to include those at the screening stage.

1)Demand and demand uncertainty of the parts
Parts with low and uncertain demand are more suitable for spare parts

2) Overall part size
(Part must fit build volume of the equipment) As the equipment for different AM technologies have build size restrictions, part-size is usually a restriction, if it cannot be fitted in the build chamber. Sometimes changing the orientation of the part may be needed to fit it into the build chamber, but it might require additional support structures and may increase the overall production time. Decisions about part orientation in the build chamber can be taken only for the selected parts and not at this stage. The company can specify the upper limit of size or a range of sizes of parts, which they would like to consider

3) Materials
Not every material can be printed. Thus, parts whose specified materials can be printed can be screened when a company is starting on its journey of producing spare parts using AM. Later, alternate materials which closely match the specifications can be explored.

4) Supply lead time
Parts with long supplier lead time for conventional manufacturing can be more suitable for producing using AM as overall lead time can be reduced. A company may decide to specify either the lower limit or range of lead time, which they would like to consider.

5) Purchase price or unit cost
High priced parts, in general, can be suitable for AM but some high valued parts may be infeasible because of their size or materials etc.

6) Value of inventory across all locations
Usually parts with high inventory value can be good candidates to be produced using AM. High value of inventory can be due to high number of parts in stock or due to high prices. Sometimes, if a lot of stock is available, a company may decide not to proceed with producing such a part using AM. Some other company may decide that they would like to get rid of that stock and hence it may be worth producing those using AM. A company may also decide to include either price or inventory value as a screening factor. 



Figure 1: Step-by-step approach for selecting spare parts suitable for AM

Step 3: Identify the factors to be used for scoring the spare parts for AM suitability
Some potential factors for scoring spare parts in terms of their suitability for AM can be as follows:
• Lead-time;
• Unit cost;
• Criticality in terms of influence of the part for equipment shutdown; 
• Demand predictability measured as standard deviation of demand; 
• Supply risk in terms of number of suppliers; 
• Minimum order quantity. 

The screened parts can be scored based on the above factors and how the factors are related to the objectives. Parts with long lead time, high cost, high criticality, low demand predictability, high supply risk and high minimum order quantity in the existing manufacturing process will be more suitable for AM.

Step 4: Choose appropriate method for scoring the parts
There can be multiple approaches to score the spare parts in terms of suitability to be produced using AM.

We mention some basic guidelines below:

1. Multi-criteria decision making approach (MCDM) –scoring parts on factors and linking factors to be objectives (suitable for less number of factors and less number of parts).

2. Logic decision diagrams, cluster analysis and fuzzy inference system (suitable for large number of parts, medium number of factors but strong interrelationships of factors and objectives). If the factors are dependent on each other and how they influence the objectives depends on the levels of the factors i.e. low, medium and high, logic decision diagrams can be built to score a part using different decision rules. In terms of disagreement between experts about the relationships, the relationships can be expressed as fuzzy numbers and fuzzy inference system can be used to score the parts.

3. Cluster analysis and MCDM approach for ranking of part clusters and within cluster ranking of parts (large number of diverse parts, limited to medium number of factors and independence of factors) If the factors considered are independent but the spare parts portfolio is very diverse, it will be necessary to cluster the parts, rank the clusters and then rank the parts within the clusters.

4. Bottom-up expert driven selection using a questionnaire or selection protocol (no data available or not possible to do quantitative analysis) If no data is available in digital form to score the parts, it is prudent to use the expert judgment of the service personnel or maintenance technicians. But, as those persons may not be aware of AM technologies, it is important to demonstrate to them what is possible using AM through workshops and ask them to suggest spare parts, which they find most difficult to get. Companies can also potentially run internal competitions to identify some spare parts to start with.

Step 5: Assess the impact of AM produced spare parts on performance objectives
Once some spare parts are identified, the next step is to develop business cases for the parts. But, those can be done once the appropriate technologies and equipments are identified. For that purpose, a database of all available AM equipments, the technologies they use, the materials they use, the build chamber of the equipments, the surface finish achievable need to be documented.

For the spare parts which are scored high in terms of their suitability for AM, the most appropriate technologies and equipments can also be shortlisted. Then total-cost of ownership based cost models need to be developed to understand the economic implication of producing the parts using AM. The relevant cists which need to be considered are materials, production, energy, labour as well as the savings in inventory and transportation costs over the lifecycle of usage of the parts for different demand scenarios.

Quotes from different service providers can also be used as a reference for taking the final decision. Once a company has done this exercise and identified a few feasible parts, some machine learning techniques can be used to identify the patterns amongst the most suitable parts so that the process can be automated when new parts are added to the portfolio.