锻造模具寿命的监测与预警｜Monitoring and Predicting Forging Die Life

摘要：模具寿命数据多数工厂只有总数没有明细。讲单模档案要记什么、失效前的可测信号、从记录到换模计划的算法路径。
Abstract: Most shops know total die spend but not per-die history. What a single-die record should hold, measurable signals before failure, and the path from records to a replacement schedule.

正文： 模具是锻造车间第三大成本，但问起来一副模具到底打了多少件、坏在哪一腔，多数答不上来。寿命管理的前提是每副模具有一条自己的档案。
Main text: Dies are the forge shop’s third-largest cost, yet ask how many strokes a die has thrown or which cavity failed first, and silence follows. Life management starts with a per-die record.

一、单模档案记什么
1. What the per-die record holds

• 身份与履历：模号、钢种、热处理批次、翻新次数。
  Identity and history: die number, steel grade, heat-treat lot, rework count.
• 服役数据：打件数、生产的产品批、班次、异常停机次数。
  Service data: strokes, product batches, shifts, abnormal stops.
• 失效记录：失效模式（磨损、热裂、塌陷、断裂）与失效位置照片。
  Failure records: mode (wear, heat checking, sink, breakage) and location photos.

二、失效前的可测信号
2. Measurable signals before failure

锻件飞边厚度缓慢变大是模膛磨损的影子；同一模膛锻件错模量漂移是导向或模座出问题；锻件粘模频率上升对应表面粗糙度劣化。这三个信号都能从在线检测或人工抽检里拿到，提前一到两个班次预警。
Slowly growing flash thickness mirrors cavity wear; drifting mismatch points at guides or die seats; rising sticking frequency tracks surface degradation. All three are catchable via inline or manual sampling — one to two shifts of warning.

三、从记录到换模计划
3. From records to a schedule

积累半年数据就能画出每套模具的寿命分布：同类模具取八十分位寿命做预警线，配合订单计划提前翻修或备模。备模充足的车间，模具成本反而更低——因为失效不再发生在压机上。
Six months of data yields a life distribution per die family: set the warning line at the eightieth percentile and schedule rework or spares against the order book. Shops with spares ready actually spend less on dies — because failure no longer happens at the press.