Lessons Learned from Implementing the Lean Principles in a Single High-Mix Low-Volume Make-To-Order Compressor Parts Machining Cell

Acknowledgement

This article is based on work that I did when I was the Director of IE Research at Hoerbiger Corporation of America from 2012-2014. After an academic career that spanned 22 years, this was my first industry job! It was a golden opportunity given to me by Hannes Hunschofsky, who was then the President of Hoerbiger Corporation of America. He invited (and challenged!) me to pilot the implementation of JobshopLean in their Houston, TX, facility.  This industry job gave me the opportunity to put into practice the research on JobshopLean that I did at The Ohio State University.  I have learned the hard way that the only way to learn the unique IE that Toyota pioneered is through hands-on implementation. Thank you, Hannes!  

Scope of this Article

This article describes the lessons I learned from being part of an implementation team that implemented a high-mix low-volume machining cell in a Make-To-Order manufacturing facility that produced compressor parts for field repairs. Unlike a repetitive high-volume automobile manufacturer like Toyota, small and medium-size custom manufacturers (SME) usually make a variety of products.  Most of these products do not have the volume and repetitive demand to justify producing those products on assembly lines. However, by grouping products with similar/identical routings into part families, the aggregate workload on key machines needed to produce a part family could be sufficient over the long term. This would justify the implementation of a manufacturing cell that will produce any and all parts in a part family.     

The Principles of Lean

The Principles of Lean are a five-step thought process for guiding the implementation of Lean:

1. Principle #1: Specify value from the standpoint of the end customer by product family.
2. Principle #2: Identify all the steps in the Value Stream for each product family, eliminating whenever possible those steps that do not create value.
3. Principle #3: Make the value-creating steps occur in tight sequence so the product will flow smoothly toward the customer.
4. Principle #4: As flow is introduced, let customers pull value from the next upstream activity.
5. Principle #5: As value is specified, Value Streams are identified, wasted steps are removed, and Flow and Pull are introduced, begin the process again and continue it until a state of perfection is reached in which perfect value is created with no waste.

Group Technology and Cellular Manufacturing

Group Technology (GT) is a method for classification of manufactured parts (or assemblies) into groups (or families) that have similar manufacturing routings, design attributes, materials, quality specifications, etc. A part family is “a collection of parts which are similar either because of geometric shape and size or because similar processing steps are required to manufacture them”. Knowledge of these groups can be exploited in a variety of activities such as facility layout, machine design, fixture design, tooling standardization, variant design, cost estimation, feature-based process planning, etc. Since the 1960s, GT has been practiced around the world for many years as part of sound engineering practice and scientific management.       

Cellular Manufacturing (CM) is an application of GT to factory reconfiguration and shop floor layout design. A manufacturing cell is “an independent group of functionally dissimilar machines, located together on the shop floor, dedicated to the manufacture of a part family”. Since the 1960s, cells have been implemented in high-mix low-volume manufacturing facilities, including job shops, to gain most of the operational benefits of flowline production. Due to the diversity of machines in it, a cell is capable of completing all operations required by a family of products, usually with minimum (or no) reliance on outside resources. If a cell is self-sufficient, it produces complete products ordered by customers. This motivates the team of employees working in the cell because they see their work earning revenue for their company. Due to the proximity between different machine operators, the cell fosters team work, facilitates quick resolution of quality issues, promotes visual control of WIP (work-in-process) and guarantees shortened lead times. But there is a down side when a complete Value Stream is compressed into a cell! In order for a cell to be dedicated to the production of a single part family, it must have the necessary equipment, flexibility to produce a variety of parts, cross-trained employees, autonomy to make personnel decisions, reliable quality and deliveries from vendors, and complete (plus timely) response from other in-house departments, such as Maintenance, Purchasing, Production Control, etc.

A Manufacturing Cell is Analogous to a Value Stream

Any manufacturing cell implemented in a high-mix low-volume Make-To-Order manufacturing facility automatically embodies the Principles of Lean because:

1. The cell has focus. It produces a family of parts (or assembled products) to fulfill orders received from one or more customers.
2. The cell is (i) Flexible enough to produce any and all orders for parts that belong in a specific part family and (ii) Lean because it easily eliminates two of the Seven Types of Waste in its daily operations – Moving and Waiting.[1] 
3. The cell employees have to work as a team that is accountable for meeting performance expectations set by management. 
4. The cell can be autonomous if it has all required resources (including inspection) available at point of use.[2] 

Case Study

At Hoerbiger Corporation of America (HCA-TX), they were loosely organized into seven cells – five in the machine shop and two in the molding department. Except for the QRC (Quick Response Cell), the other cells were not self-contained because some of the operations in every cell were external and done in other areas of the facility. Of the five existing machining cells in the facility, the Manual Packings Cell (MPC) was the best candidate for demonstrating the viability of JobshopLean in our high-mix low-volume facility. Unlike the MPC, the other four machining cells were in flux for a variety of reasons, such as changes in their product mix, technology upgrades, reduction or replacement of vendors, etc.  Since HCA-TX management was interested in exploring the merits of JobshopLean, they gave their full support for a pilot project to re-design the MPC. 

Main Objective

The overarching objective for designing the new cell was to co-locate all of its manufacturing equipment so it could operate autonomously. We collected the routings of all the parts that were produced in the MPC during a 5-day week and sorted them in order of decreasing production volumes (P-Q Analysis). Next, we selected the top few parts and mapped their routings on the current layout of the cell to produce the Flow Diagram shown in Figure 1. The figure clearly shows the chaotic flows of products into and out of the cell prior to its re-design. In the figure, the flows shown in red represent large values in the From-To Chart and the flows shown in green represent low values. In stark contrast, Figure 2 shows the Flow Diagram for the new cell layout that was implemented. 

Figure 1 Material Flows in the Current Layout for the Manufacturing Cell

Figure 2 Material Flows in the Proposed Layout for the Manufacturing Cell

What Our Team Did to Implement the Cell

1. Used software tools like PFAST and STORM to map the material flows and design the cell layout.
2. Hired a full-time IE intern who worked full-time on the project and engaged daily with the cell employees.
3. Involved the in-house Tiger Team, a group of employees who were responsible for all Lean projects.
4. Involved the in-house IE who provided valuable time study data.
5. Involved the cell employees to implement 5S and ergonomics-related improvements at their machines.
6. Accepted input from the cell employees about the changes that were made in their work place.
7. Justified the implementation of the cell to upper management. 

Benefits of Implementing the Cell

• The external resource requirements for the cell were reduced to (a) the delivery of raw materials to the cell from the Receiving department and (b) the removal of finished product from the cell to the Shipping department. This significantly reduced the inter-operation transportation delays and queuing delays experienced when several operations were done external to the cell. It was estimated that labor time wasted every year in material transport for all orders processed in the cell was reduced by 51 hours.
• The Line Of Sight Efficiency (LOSE) improved from 0.286 (See Figure 1) to 0.714 (See Figure 2). One machine operator did not have to walk far to communicate a quality issue or tooling problem to another operator. Or, if one operator wanted to take a break, they could inform the operator of the previous machine to not over-produce beyond a certain number of units during his absence.
• Two pairs of machines in the cell could each be tended by a single operator. Each operator would spend significantly less time walking between the two machines he was operating.
• The distance travelled by any order processed in the cell was reduced from 618 ft. to 368 ft. This reduced operator motion waste and allowed for smaller batches of parts to be moved between consecutive machines.
• Order Flow Times, which were as high as16 days, could be reduced to ≤5 days. .
• The Standard Lead Time quoted to customers could be reduced from 20+ days to 10 days. 
• The cell footprint was reduced from 2816 sq. ft. to 1410 sq. ft. The freed-up floor space could be used to either (1) add a new cell to bring in new business for the company or (2) co-locate the Receiving and Shipping departments adjacent to each other.

Lessons Learned from Implementing the Cell

Lean Principle #1 provides a sound strategy for implementing Lean: Lean Principle #1 is “Specify value from the standpoint of the end customer by product family”. A cell is a factory-within-a-factory whose performance can be measured using the same metrics (Quality, Cost and Delivery) that are used to measure overall company performance.  Except that a cell is designed to produce a single product family extracted from the entire product mix. 

Bad data will scuttle good software …. always! A major take-away from this project was that computer software is necessary for implementing JobshopLean. However, the primary role of the software should be to enhance the effectiveness of the decisions made by the team ex. it should guide them to generate and evaluate the alternative layouts that were designed for the cell. This new layout was designed using the PFAST and STORM software tools. Unfortunately, as the preliminary software outputs indicated, the routings for the different parts in the family that we extracted from the ERP (Enterprise Resource Planning) system were inaccurate. So we did what Lean Thinkers constantly advise, “Go to the gemba!” The cell employees and machine shop supervisor corrected the routings and even explained to us the differences between the parts. Collecting data in person is a form of learning!     

A cell automatically helps to implement Lean Principle #2: Lean Principle #2 is “Identify all the steps in the Value Stream for each product family, eliminating whenever possible those steps that do not create value”. It is well-known that 95% of the time that an order spends in process is comprised of two of the Seven Types of Waste – Transportation and Waiting. Only 5% of that time is spent on the machines that perform the process steps specified in its routing. The obvious solution for reducing, often eliminating, the wastes of Transportation and Waiting is to implement a manufacturing cell to produce a product (or product family)! How to identify and measure the wastes and delivery delays in the timeline of each part being produced in the cell? Please do not rush to produce (1) a Value Stream Map, (1) an Extended Value Stream Map (for steps that must be performed external to the cell), (3) a Process Flow Chart and (4) a Spaghetti Diagram for each and every part being produced in the cell. It will take you forever! Instead, prepare these four charts for the part that has the highest production volume, as indicated in the P-Q Analysis. Then use those charts to educate the cell employees about the wastes that they have “been living with” all these years.  This way you will focus your time and energy on designing and implementing a cell which will automatically eliminate significant amounts of waste.

A cell automatically helps to implement Lean Principle #3: Lean Principle #3 is “Make the value-creating steps occur in tight sequence so the product will flow smoothly toward the customer”. Cells are the ideal option to implement this principle. Recall that the amoeba and paramecium are self-sufficient single-cell organisms that exist in nature? Similarly, a manufacturing cell is capable of operating autonomously! Inside a cell, the employees can use all the Lean tools (cross-training of the cell employees, 5S, SMED, TPM, Poka-Yoke, Standard Work, etc.) to operate the cell without having its operations disrupted by supply of defective materials, machine breakdowns, long setups, employee absenteeism, poor communications with suppliers and customers, etc. 

An ERP system is going to prevent implementation of Lean Principle #4: Lean Principle #4 is “As flow is introduced, let customers pull value from the next upstream activity”. A cell is the nearest equivalent to an assembly line. The new cell design had most of the operations required to produce the part family being done inside the cell. All that was needed to estimate the work load due to orders released to the cell was to ask any of machine operators to guesstimate the work load on their machine due to the batch (or batches) of parts queued at their machine/s. With their years of experience, their guesstimates were really good! Unfortunately, the cell received orders per the production schedule set in the corporate ERP system which (1) assumed infinite capacity and (2) used standard backward scheduling with fixed lead times to determine the release dates to send orders to the cell. Alas, this was not a battle that we ever wanted to fight because we knew at the outset who would win it!

Use the first step in 5S (Sort) as a test of employee buy-in: Lean is about involving employees in achieving the cost reduction and performance improvement goals of their cell. So, before we moved a single machine to implement the cell, we decided to first pluck the low-hanging fruits that required negligible investments. Figure 3 shows the housekeeping that was undertaken by each employee in the cell. 

Figure 3 Workplace Improvements by Cell Employees

Allow employees to show off their efforts (and talents!): Lean can inspire every employee to “take charge of their own work environment”. Luong Dam, who runs the three Cincinnati Mills in the MPC, had been with the company for nearly three decades!  He did not stop with the initial housekeeping effort. Instead, he worked tirelessly with us over a period of two weeks to single-handedly re-arrange his work center, as shown in Figures 4(a) and 4(b). 

Figure 4(a) Milling Work Center Before Cell Implementation

Figure 4(b) Milling Work Center After Cell Implementation

The motivated employees will surely come up with innovative ideas! Figure 5 shows an example of a fixture built by an employee in another cell (QRC) that was replicated to store the tools used on the Le Blonde lathes in the MPC. Phillip Nguyen, one of our senior multi-talented employees, built it and never asked to be compensated for his time and effort!   

Figure 5 Fixture Built for the MPC Cell to Store Tools

Follow up every Lean training session with activity in the gemba: To train members of our Tiger Team and the cell employees, I developed a simple simulation to compare and contrast (1) Batch Flow, (2) Single-Piece Flow and (3) Transfer Batch Flow. Also, I used this YouTube video titled One Piece Flow versus Batch Production – Lean Manufacturing in that training module. Okay, so we all knew that single-piece flow in a high-mix machining cell that operated in a Make-To-Order mode was infeasible. That was because every week the MPC processed a different mix of orders with unique due dates. But had not the cell helped to position inter-related machines in close proximity to each other, as in an assembly line? So we posed this question to the training group, “Can we split up any order batch into (smaller) transfer batches that could be moved one-after-the-other from the “supplier machine” to the “customer machine”?  

Now let me share with you what happened a few days after I had run the simulation and showed the video to the group. Luong Dam, the employee who ran the three Cincinnati Mills in the MPC, called the intern and me over to show us what he had implemented between the Haas Mill and the Cincinnati Mills that he operated. Figure 6 shows his idea to start pulling just enough pieces off the Haas Mill that he could pack on the arbor used on each of his three mills. Thereby, any order could be split up and each batch run in parallel on all three mills. To this day, I discuss Luong Dam’s idea to illustrate the integration of (1) cell implementation, (2) training employees using the single-piece flow simulation and (3) encouraging the employees to come forward with their own improvement ideas!

Figure 6 Transfer Batch based on the Capacity of the Fixture used on the Mill

Implementing Lean Principle #5 Could Take Time! Lean Principle #5 is “As value is specified, Value Streams are identified, wasted steps are removed, and Flow and Pull are introduced, begin the process again and continue it until a state of perfection is reached in which perfect value is created with no waste”. As more and more cells are implemented, among the major challenges that management could face during this transformation are (i) being able to place inside each cell all the equipment it needs to be self-sufficient, (ii) ensuring that the order pipeline for the part family assigned to each cell is sustained (else the cell will die), (iii) cross-training employees to run different machines in their cell, (iv) allowing the cell operators to manage their cell, etc. Figure 7 provides a map of the different tools that pre-date the Toyota Production System or were developed by Toyota that could be used to improve cell performance. It is not necessary (nor advisable) to implement each and every one of these tools when the cell is first commissioned. The choice of tools should be “pull”ed depending on the needs (and ideas) of the cell operators.

Figure 7 Tools of the Toyota Production System to Implement the Lean Principles

I had been teaching IE (Imaginary Engineering) to my students for 22 years: I have many reasons to say this. Some of those reasons are:      

• “ToyotaIE” is a very unique flavor of Industrial Engineering that can only be learned if one is employed at/by Toyota. Alas, employment at Toyota is not something that the typical IE faculty can claim on their resume. How I envy the faculty at the University of Kentucky!
• The standard practice in any undergraduate IE course is to teach from a textbook. IMHO, the undergraduate textbooks that are being used in most IE departments teach archaic (even wrong!) principles and methods. That is because their authors have little, if any, hands-on IE experience to know any better!  Frankly, any IE (Industrial Engineering) faculty who is assigned to teach undergraduates absolutely needs to have worked in industry for 3-5 years. In particular, he/she must have industry experience related to the course he/she is teaching. 
• It is extremely important to have at least one course in the core curriculum that teaches students about employee engagement and leadership skills that influence the success (or failure) of each and every CI (Continuous Improvement) project they will do after the enter industry!
• Academic research journals promote IE research that will never work in industry!

Is management willing to let a cell’s team operate autonomously? Each cell has the potential to become an Autonomous Business Unit (ABU) within the company. But this will require the team of employees in each cell to be empowered to manage day-to-day operations and make decisions about scheduling, assigning work to different operators, deciding who gets cross-trained on which machines, etc. The team would determine the CI projects to eliminate, or at least mitigate, all the constraints that force the cell to send its orders to external resources, either in-house or vendors. Ideally, each cell should communicate directly with their customers concerning changes in delivery dates, questions about part drawings or routers, unforeseen quality problems, etc. It is my personal opinion that Western manufacturers not knowing about the Toyota Management System is one of the primary reasons why Cellular Manufacturing failed to take root in the early 60’s and 70’s.

Cells are not a panacea for high-mix low-volume manufacturing: Job shops are ill-advised to pursue a complete shop reorganization into FLean (Flexible and Lean) cells.  In fact, it is possible that the production volumes and demand stability for many part families simply could not justify dedicating equipment, tooling, and personnel to produce any of those families in a stand-alone cell. In that case, try implementing Virtual Cells!

What Do You Think?

If you are a high-mix low-volume (HMLV) manufacturer and have customized the implementation of Lean in your facility or you would like to know how, let’s get a conversation going! Send me your questions and comments on this article to [email protected]. Thank you.

References

Irani, S. A. (2017, Forthcoming). JobshopLean: Adapting Lean for Small and Medium High-Mix Low-Volume Manufacturers. Sugar Land, TX: Lean & Flexible, LLC. 

[1] “FLean” was coined by Hannes Hunschofsky, my former boss at Hoerbiger Corporation of America.

[2] The design and implementation of Virtual Cells needs (1) the creation of the position of Water Strider/s and (2) a reliable daily production schedule.