故事 · 一句话改变了制造业
Story · One Idea That Rewired Manufacturing
1950 至 1970 年代，新乡重夫在丰田反复琢磨一个问题：为什么换模非得停那么久的机器？
他蹲在机床旁逐个动作拆解，发现工人在停机期间做的大量事 —— 找工具、搬模具、调位置 —— 其实完全可以在上一批还在生产时就准备好。
他把这类“运行时可做”的叫外部作业，把“非停不可”的叫内部作业，喊出口号：个位数分钟换模(Single-Minute)。
某次他把一台螺栓压机的换模从 4 小时压到了 3 分钟。这套方法，是后来小批量、混流生产能成立的地基。
Between the 1950s and 1970s, Shigeo Shingo kept circling one question at Toyota: why must changeover keep the machine down for so long?
He squatted next to the press and broke every motion down — and realized that most of what operators did while the machine was idle — fetching tools, hauling dies in, lining up positions — could just as easily have been done while the previous batch was still running.
He named the "can-be-done-while-running" work external setup, the "machine-must-be-stopped" work internal setup, and rallied the slogan: changeover in single-digit minutes.
On one bolt-forging press he drove a four-hour changeover down to three minutes. This method became the foundation on which small-lot and mixed-model production were eventually built.
1 换模时间轴：红=停机内部，绿=运行外部 Changeover Timeline: Red = Internal (Stopped), Green = External (Running)
优化前Before上排是优化前：所有作业都挤在停机段里串行排队。点「优化」后看下排 —— 大部分外部作业被挪到机器还在运行的时段提前做，剩下的内部作业再并行压缩，红色停机条肉眼可见地缩短。 Top row is the baseline: every task crammed into the downtime block in serial order. Hit "Optimize" and watch the bottom row — most of the external work gets moved into the window when the machine is still running, the remaining internal work is parallelized, and the red downtime bar visibly shrinks.
2 停机时间：优化前 vs 优化后 Downtime: Before vs After
两条横条直观对比真正的停机时长。外部化把作业搬离停机段、并行与简化再压缩剩余内部作业 —— 二者叠加，就是 SMED 把几小时变几分钟的全部秘密。 Two side-by-side bars compare actual downtime. Externalization lifts work out of the stopped window; parallelization and simplification compress what remains. Stack those two moves and you have the entire trick that turns hours into minutes.
3 现实里的 SMED SMED in the Real World
快速换模：注塑、冲压、印刷换版，换模从小时级降到分钟级，是小批量生产的入场券。
Quick changeover: injection molding, stamping, print plate changes — driving changeover from hours to minutes is the ticket of admission to small-lot production.
小批量经济性：换模便宜了，经济批量随之变小，库存和 EPE（上一页）一起下降。
Economics of small batches: cheap changeover shrinks the economic batch size, pulling inventory and EPE (previous page) down together.
设备 OEE 提升：停机少了，设备综合效率里的“换模损失”直接变小，可用产能增加。
OEE goes up: less downtime means the changeover-loss bucket inside OEE shrinks directly, freeing up available capacity.
新乡重夫方法：F1 进站换胎只要 2 秒，正是 SMED 思想的极致 —— 一切能提前的全提前。
The Shingo method: a 2-second F1 pit stop is SMED taken to the limit — everything that can be done beforehand is done beforehand.
一句话In One Line
SMED 的威力不在“干得更快”，而在重新定义什么必须在停机时做。
第一步把作业分成内部/外部，第二步尽量把内部转成外部（提前到运行时做），第三步把剩下的内部作业并行化、简化、消除调整。
换模一便宜，小批量就有了经济性，EPE 才缩得短、库存才压得下、混流和拉动才跑得动 —— SMED 是整个精益生产体系隐形的承重墙。
SMED's power isn't "do it faster" — it's redefining what must be done while the machine is stopped.
Step one: classify every action as internal or external. Step two: convert internal to external whenever possible (do it ahead of time, while the machine still runs). Step three: parallelize, simplify and eliminate adjustment on what's left.
Cheap changeover makes small batches economical, which lets EPE shrink, inventory drop, and mixed-model pull actually run. SMED is the invisible load-bearing wall of the entire lean production system.
常见误用Common Mistakes
只想着让工人手脚更快。SMED 先重排作业（内转外），不是逼人加速。
Trying to make the operator's hands move faster. SMED resequences the work (internal → external) first — it isn't about pressuring people to hurry.
停机后才去找工具、搬模具。这些外部作业应在上一批运行时就备齐到位。
Hunting for tools and hauling the next die only after the machine is stopped. Those are external tasks — stage them while the previous batch is still running.
换模快了却不缩批量。SMED 的红利要靠减小批量兑现，否则白改一场。
Faster changeover, same big batches. The SMED dividend only pays out when you actually shrink the batch — otherwise the whole project was wasted effort.