Total Productive Maintenance (TPM) is the best known framework for organizing maintenance. In this post I would like to dig deeper into its ideas and how it works. The principles of TPM are often arranged as “eight pillars,” although there are different views of what these eight pillars are. Let me give you an overview of these. In this post I will also look at the general problem of maintenance. In subsequent posts I will be going deeper into the individual pillars.
I will try to not only give a brief explanation how all the pillars work, but also give my opinion on which pillars I find useful and which one I don’t. There are some excellent elements in this framework that everybody should use on the shop floor, but there are also additional pillars that to me serve little purpose. In any case, I look forward to hearing your opinion in the comments.
The General Problem with Maintenance
Maintenance is sometimes neglected in manufacturing. The first problem is that the relation between maintenance and the effect of maintenance is ill understood. If you change the oil of your car, does it really improve the performance or prevent breakdowns? Sure, if you never change the oil, then eventually it will be bad for the vehicle. But what maintenance interval is sensible? Should you exchange every 15,000 kilometer, or every 40,000 kilometer? Or even less? The cost is often easy to determine based on labor and parts. For your car this would be the oil, the oil filter, etc., as well as the labor for exchanging the oil.
But the benefit is much harder to figure out. Let’s take your car as an example. How often will your car have problems? What kind of problems? Is it just a higher fuel consumption, or will the entire engine have to be scrapped due to a lack of oil. Does the breakdown happen when you just by chance drive to the mechanic anyway, which would be less of a hassle, or does the breakdown happen when you are on your way to an interview for your dream job, which would be extremely detrimental?
Given sufficient data points, you can eventually describe these relations statistically. Some industries have better data, like automotive or aviation, simply because they collect plenty of data and accidents may be extremely expensive. But getting sufficient data points takes time, and a small company may never have enough data to understand a machine before the machine is replaced by a new one. Hence, a cost-benefit analysis is difficult to to for maintenance.
The second problem is the delay between maintenance (or lack thereof) and the improved uptime (or lack thereof). This can easily be years. For example, if you buy a used car, you probably should look at the service history. Not only does this delay make the relation more difficult to understand, but in industry the person paying for the maintenance and the person benefiting from it may be different. It is a dirty trick by one manager to look good by saving money on maintenance but leaving the problem of deteriorating equipment to his successor three years later. At the same time, you can’t blame him either, since the relation is also difficult to understand. It does not help that maintenance consultants almost always favor more maintenance, even though they are (often) right. Overall, determining the right amount of maintenance is usually a rough and error-prone guesswork at best.
Overview of the Pillars
Now, let’s have a look at the eight pillars of TPM. Frameworks are often arranged as houses with pillars, foundations, and so on. There is the house of lean, with three columns and the three pillars of SCRUM. There are some graphs with the four pillars of Six Sigma, and so on. Pillars are used to give these ideas weight, although sometimes it feels more like a sales tactic to me.
TPM also has pillars, and an unusually high number of them too. In the English literature, there are usually eight pillars of TPM. Originally there were only seven, but the the Japan Institute of Plant Maintenance (JIPM) expanded them to eight. Additionally, 5S is also often listed as the foundation of TPM. Let me give you an overview of the most common version of the pillars:
- Focused Improvement (meaning Continuous Improvement)
- Autonomous Maintenance
- Planned Maintenance
- Early Equipment Management (somewhat similar to Design for Maintenance)
- Quality Maintenance (which in lean would be called Jidoka)
- Education and Training (which in lean would be part of standardization)
- Administrative & Office Maintenance
- Safety, Health, and Environment
In Japanese, these pillars would be as follows:
- Focused Improvement is Kobetsu-Kaizen (個別 改善 for individual; separate; personal; case-by-case and betterment; improvement).
- Autonomous Maintenance is Jishu Hozen (自主 保全 for independence; autonomy; self-reliance and preservation; integrity; conservation; maintenance).
- Planned Maintenance is Keikaku Hozen (計画 保全 for plan; project; schedule; scheme; program; and preservation; integrity; conservation; maintenance).
- Early Equipment Management is Shoki Kanri (初期 管理 for early; initial stage and control; management).
- Quality Maintenance is Hinshitzu Hozen (品質 保全 for quality and management).
- Education and Training is Kyōiku Kunren (教育 訓練 for education; schooling; training; instruction; teaching; upbringing and training; drill; practice; discipline).
- Administrative & Office Maintenance is Kanri kansetsu bumon katsudō (管理 間接部門活動 for control and internal division; back-office section; indirect department).
- Safety, Health, and Environment is Anzen and Kankyō Kanri (安全 for safety; security, 環境 管理 for environment and control; management).
Variants of the Pillars
However, there are also other versions of the eight pillars. They often differ not only in sequence, but also in the names and the content. For example, what in English is called Focused Improvement, is in German Continuous Improvement (which for me as a lean guy feels more normal). The Early Equipment Management focuses on design for maintenance, but this is sometimes replaced by Ramp Up, focusing more on ramp up than on machine design.
In other sources you can also find only seven pillars of TPM. Some of these dropped the safety, health, and environment aspect (and I am always unhappy if safety is overlooked), others dropped the early equipment management. There are even sources with only five pillars, dropping the administrative part, and putting continuous improvement and safety as foundations instead of pillars (where I like them much better). If you want more pillars, this is not a problem either. There is a variant of TPM with nine pillars also available, adding logistics. There are even completely different structures and pillars that claim to be TPM, with pillars like “To reduce forced deterioration” or “To reduce natural deterioration.”
There are many more versions of the pillars of TPM. If you’re curious, just do a google image search for the “Pillars of TPM.” The points I’m getting at are that there is no fixed canon of pillars, and everybody likes drawing houses. However, despite a lack of structure, and besides an overlap of some of these topics, TPM contains lots of useful ideas worth exploring. In my next posts I will go deeper into the different pillars of TPM. Now, go out, think about how you can improve your maintenance, and organize your industry!
PS: Many thanks to Leandro Barreda and Gil Santos for nudging me to write about maintenance.