CMMM高成熟主任评估师谈智能制造能力成熟度模型四级评估

“ 智能制造能力成熟度标准符合性评估（以下简称“CMMM评估”）是基于GB/T 39116-2020《智能制造能力成熟度模型》、GB/T 39117-2020《智能制造能力成熟度评估方法》两项国家标准，对产品制造过程的智能化程度开展系统性评估。截至2022年底，全国已有400多家制造企业获得相应等级的标准符合性证书，其中“四级(优化级)”是国内制造企业目前达到的最高级别，通过评估的企业也是国内高端制造业的典型代表。”

那么企业如何识别能否达到四级？建成四级企业需要具备哪些能力？四级评估相比与其他等级评估，在评估准备时需关注哪些要点？针对大家关心的问题，笔者从十余家四级企业评估活动进行总结，分享下体会。希望在企业智能化能力识别、能力建设、评估准备等方面，对申请四级评估的制造企业有所帮助。

一、四级企业综述

四级(优化级)：企业应对人员、资源、制造等进行数据挖掘，形成知识、模型等，实现对核心业务活动的精准预测和优化。

–GB/T 39116-2020《智能制造能力成熟度模型》

四级能力是企业在达到三级（集成级）的基础上，即数字化的生产、检验、物流设备联网集成，各应用信息系统间信息集成，实现企业各生产环节互联互通、数据共享、业务协同的基础上，广泛应用各生产环节、业务职能采集的数据，基于实时数据的深层次分析应用，辅助企业人员做出更为科学的决策，或自动对生产过程进行优化调整。

截止2022年底，国内已有29家企业获得CMMM四级证书，占所有已获证企业的7.25%；其中制造大省的江苏、广东、山东通过四级评估的制造企业分别为9家、8家、4家，北京、安徽、福建、陕西、湖南、广西、辽宁等省市也有当地龙头企业通过四级评估。

从行业分布来看，部分行业龙头企业已达到国际领先水平。例如，以小天鹅、海尔为代表的家用电器制造行业，以深南电路为代表的印刷电路板制造行业，以徐工为代表的工程机械行业，以华星光电、京东方为代表的液晶面板制造行业，以中车为代表的轨道运输设备制造行业，以及其他石油炼化、水泥、医药制造行业。同时，有近50%的四级企业，当地并无政策奖励支持，为企业自发申请评估。由此可见，越来越多的行业龙头企业，希望自身的智能化建设成果得到权威认可，或在评估中寻求继续提升的方向。

从目前四级企业评估结果来看，企业都存在一定的能力弱项,大部分企业还处在四级的初级阶段，虽然在部分业务单点上构建了分析模型，但模型应用效果还需进一步验证，四级的能力建设还需要很长的路要走。

二、申请四级评估常见误区

在众多申请四级评估的企业中发现，一些企业对申请四级评估并未充分了解，对四级评估考核的维度、难点、要求，未进行对标判断。企业盲目申请四级，导致最终评估结论降级。针对四级能力识别，经常出现如下误区。

误区一：大企业/行业品牌知名度高就能通过四级

此误区较为普遍，因为已获证企业确实行业的知名企业居多。行业知名企业确实在产销量、营业额、市场占有率、品牌形象上处于领先地位，且企业资金雄厚，智能化投入也较多。但CMMM评估考核的是实际的产品制造过程智能化应用情况。有些企业的品牌市场宣传大于其实际制造水平、智能化水平；还有些品牌依靠外部代加工，没有制造工厂，无法进行评估。

曾经遇到一家业内知名的非标装备制造企业，申请四级评估，生产工艺类型只有装配，产品组装环节采用全人工方式，且装配过程打螺丝还是采用普通电动螺丝刀，不能实现扭矩参数的数据采集，其他智能化应用较少，最终预评估对标下来，评估组认为企业达到三级还有些困难，整体处于二级水平。企业无法接受作为行业龙头而评估结论是二级，最后终止评估。

误区二：已获得智能制造荣誉或权威认可就能通过四级

有些企业已获得省部级、国家级，甚至国际性的智能制造荣誉，这是很好的智能化建设基础。但由于各种政策、荣誉评选的评价指标不同，不能说获得某个荣誉，就可以等同通过CMMM四级。

特别是有些企业反而容易借此荣誉，产生盲目自信心理，而忽视CMMM评估调查取证的全面性、客观性、严谨性。比如，企业承担了国家重点智能化攻关建设项目，已经由权威专家把关验收，则认为通过四级没有问题；但四级评估中，该项目可能只作为四级考核中一个子域的几个条款的有效举证，无法以偏概全代表企业整体智能化水平。

通常的企业智能化成果申报，申报材料都是尽可能展示企业好的一面，而对弱项避而不提。而对于CMMM评估，企业应基于实际业务情况确认评估范围、评估内容，应评尽评的条款不可随意裁剪、回避，不能避重就轻，应基于统一标准开展CMMM评估。一个完整的CMMM四级评估包括预评估和正式评估两个阶段，在企业现场至少6天时间，评估组利用充分的时间，完整、细致的验证企业各业务职能的智能化应用情况与应用效果。这里的调查取证包括企业获得优势、建设亮点的一面，也包括对标后的能力不足和建设缺失一面。

误区三：新建成的智能工厂就能通过四级

这里“新建成”特指生产正式运行1年左右的工厂。很多企业投资几亿、十几亿元构建新型号产品制造工厂，且工厂在设计、建造之初就按照行业最高标准建设，吸取以往经验，各种智能化技术一步到位，打造行业最先进智能工厂。不否认这些工厂的先进性，但通常新建成工厂的产量、销量还在慢慢爬坡过程，制造产品的种类也有限，特别是智能化应用堆积了很多新技术，工厂整体运行需要一段时间磨合、消化、验证，短期较难发挥智能工厂的特色能力。

此外，四级评估要点中，大多数内容是审查基于数据分析模型的在线优化、预测，由于工厂建成时间短，很多工厂的业务数据累积量不够，一些制造规律难以发现，并通过分析模型去表达、验证。例如，【设备管理】子域四级要求的“设备预测性维护”，需要对历史发生的故障做故障知识积累与分析，而新建工厂是新设备，短期故障发生率也较少，也较难提出预测、优化的需求。四级【能源管理】能力子域要求的精细化“节能模型”，新工厂运行之初重点考虑产量、质量、效率，因此智能化建设会侧重这些方面，短期很少考虑节能降耗举措，当工厂生产稳定，为进一步降低制造成本时，才会由此需求进行智能化建设。当然也有些新建工厂，在建厂之初就综合考虑了之前的节能经验，已经采用了节能化设备和方案，接下来就需要进一步挖掘新的节能空间。

因此，CMMM四级评估的工厂需要稳定运行一段时间，生产运营模式较为成熟，智能技术较好的应用落地，当工厂运行积累了一定的数据，沉淀了知识，有优化提升的需求和举措，四级的智能化成果也会很好的应用、展现。

误区四：生产运行展示效果好的工厂就能达到四级

很多人去过这样的工厂，工厂整洁明亮、工艺布局有序、设备全自动化运行、机器人协同运作、物流配送无人化、指挥调度大厅大屏看板数据展示，看起来非常先进。有的工厂还指定“红毯”路线、观光车路线，配备专职的讲解员，每到之处讲解亮点。

首先，这种参观方式只能了解工厂的局部，讲解员固定的说辞未必是标准符合性评估需要采集的证据，而一些CMMM评估关注的要点，需要评估组去生产工位详细了解验证，例如产品换型时，设备是否自动切换加工程序。

其次，上面提到的生产高度自动化、大屏看板展示等，大部分是二级、三级要求，四级条款要求的数据分析优化、预测模型，这些底层逻辑也很难用大屏看板展现。还有些大屏看板只是为了“秀”一下，看板上展示数据的时效性、准确性、真实性经不起推敲，一些统计数据很难对生产调度、生产控制起指导支撑作用。随着移动技术的发展，很多实时信息未必要放到大屏上，通过移动终端、手机APP推送信息，特别是异常信息，比大屏更有意义。

因此，在四级评估的生产车间现场巡视时，更多查证显性的三级要求；而对于四级条款，一些实时的生产过程优化控制，要到总控的后台“大脑”或者数据分析平台去查证，还有些决策模型已集成到设备运行的底层控制逻辑中。生产现场展示效果好的工厂未必能达到四级。

误区五：高端智能化终端产品的制造工厂容易达到四级

高端、智能化产品蕴含的高新技术多，产品研发周期长，部分制造工艺环节较为复杂，企业在产品研发和工艺研发智能化应用投入较多。随着市场需求的旺盛，也驱动企业在生产环节智能化应用提升。对于结构复杂产品，产品制造的社会化分工较明显，各家工厂分别负责其中关键零部件的制造，或者几道关键工序的加工，再由终端整机厂商把所有零部件装配。

但是，由于终端厂商的品牌曝光度高，容易误解为一切技术都出自终端厂商。然而，在某些行业会发现，企业制造技术水平、智能化应用水平，部分关键零部件制造厂商要比终端产品制造厂商高，例如，液晶面板制造厂商与显示器、电视机制造厂商。

当然，这里不是说绝对，很多终端产品制造厂商的智能化水平也较高，整机的核心零部件也是自家生产，大型复杂产品的装配过程也需要较高水平的智能化技术支撑。

针对以上常见误区，企业需保持“归零”心态，对照标准，认真、客观的总结智能化优势能力。

三、四级企业识别方向

为了准确识别企业四级能力，需要对照标准条款，判断企业智能化应用的符合性。当然，企业还可以从其他几个侧面初步识别四级能力，这里提供几点方向。

（一）产品制造的复杂性

从智能化需求测出发，驱动企业在制造过程使用智能化技术的源动力，在于制造过程的复杂性。这里复杂性包括多个方面，如产品结构的复杂性、制造工艺的复杂性、产品种类的多样性、市场需求的多样性、生产组织的复杂性，这些复杂性的解决一方面促使企业职能分工越来越细，岗位人员增加，另一方面也促使不断投入智能化建设减少人力。可以说，企业面对制造问题的复杂程度，解决复杂性问题的能力，代表了企业智能化应用的潜力，这也表现为每个行业独有的制造特点。

从这个角度来看，四级评估有些挑行业。有些企业虽然已经做到细分领域的行业龙头，但是产品不复杂，生产工艺不复杂，工厂生产的产品有限几种，市场需求较为固定，生产管理也较为简单，这样也就不需要多高的智能化建设，行业整体智能化程度也大致相当。企业也有提升的意愿，只是本身制造环境“复杂度”不够，一些智能化技术用不上，没有“需求”，通常这些行业的市场竞争也比较激烈。

（二）行业创新引领能力

四级企业一般位于行业前列，为了保证其行业地位，在危机意识下会主动拉开与同行企业的差距，迫使自己不断提升，但是因为没有可对标的参照物，没有可借鉴的案例，必须不断激发自己的创新能力。四级评估要求的很多数据分析优化、预测模型，企业最终的应用场景都与行业制造特点密切相关，且解决的问题多是在行业内存在、但一直很难解决的，是行业内第一家攻关的。

从标准条款的引导方面，二级、三级很多条款要求比较明确，给企业智能化建设鲜明的指引，且应用场景适用于大多数行业，很多优秀案例可以跨行业互相借鉴学习。而四级条款，有很多是方向性的指引，描述的较为简要，这也给每家企业有更多的创新发挥空间。比如【仓储配送4b】“仓储模型实现库存优化”，这里库存优化可以是物料库存水平优化，或是出入库效率优化，或是库存容积承载能力优化；同时，结合生产业务过程，优化实现的部位既可在原料仓库，也可在半成品仓库、成品仓库。

（三）数据应用环境

四级企业信息系统很多，有的是集团统一推进，有的是企业、事业部自行采购，动辄几十个信息系统。四级企业已意识到上系统不是最终目的，更重要的是整合好系统内的数据，尤其应用好关联的数据；同时，企业在三级建设中进行数据治理，规范数据编码规范、数据交换格式，也为四级数据应用打基础。

四级企业开展业务数据应用分析，需要构建数据应用环境。面对众多系统数据源，首要建立数据的汇聚平台（数据仓库、数据中台、数据湖），实现各方数据的整合，明确数据来源、数据格式、数据同步频率等等。根据生产业务需求，数据平台可能是1个或者多个，且数据来源多样，有的来自业务应用系统，有的来自数字化生产设备；用于后续数据分析的数据，还需在平台内进行必要的清洗与转换。通过数据分析应用工具，发现数据的关联与量化关系；再建立生产数据优化、预测模型部署于制造系统。

（四）智能化人才

四级智能化能力还体现在企业具备的智能化人才规模和能力。四级企业应用的信息系统多，企业IT硬件、软件维护人员数量少则几十人，多则上百人。除了系统运维，上面提到数据治理、数据分析工作都需要相应岗位的人员配备。

此外，对智能化人才的要求不只是IT人员，还有来自各业务部门技术骨干，这些人员能够将平时的业务难点转化为智能化需求，具备智能化能力和意识，四级条款要求的生产优化预测模型构建，都离不开生产业务人员的探索和实践。

综上，随着更多的企业通过CMMM四级评估，对四级能力的识别会更加清晰。企业较难判断时，可以邀请参与过CMMM评估的评估师，特别是参与四级的评估师进行就绪检查，对照GB/T 39116-2020《智能制造能力成熟度模型》四级条款调研企业智能化实施成效。

四、四级能力的建设主体在企业自身

企业在过往三级建设时，可以通过资金投入，外部实施力量快速实施项目，大多数企业在完成信息系统、数字装备的部署、验证后，都能看到明显成效。然而，四级的难点就在于可借鉴的成功案例少，很多企业在这一阶段开始迷茫，不确定方向，方向有了又不确定怎么实现，实现了又不好验证效果或难以达到预期效果。

由于企业所处行业不同，战略侧重点不同，智能化需求不同。而外部解决方案供应商，很难深刻理解企业自身的“复杂性”（见上篇，产品制造的复杂性），外部更多的是提供思路、方法、工具、系统框架；而对“复杂性”的探究，企业有更深刻的理解，因此四级建设的能动性就到了企业手里，借助外部工具，吸收转化为企业内部能力。

企业成为智能化建设主体，重要一步就是企业智能化人才体系建设。这里可能存在误区，认为智能化人才主要是企业IT人员，特别是对于一些企业从二级、三级能力建设过来，IT一直担当智能化建设的主力军，但建设到一定阶段，IT人员会感觉智能化建设长期停留在原地，日常就是不断维护系统，或者认为企业各方面建设都很完善，没有新的建设项目；有些企业还会主动从各业务部门收集智能化建设需求，但发现各业务部门也提不出需求。企业向四级发展出现瓶颈。

解决问题的关键是将IT人员与制造业务人员的能力相融合。IT人员需要下沉懂业务，各职能人员需要懂智能化；但事实上也发现，一些IT人员平日忙于维护系统，很少有时间沉下去了解业务的复杂性，而且企业的各方业务也很多；因此，更需要业务职能人员懂智能化，培养智能化意识。

在评估过程中经常发现，智能化建设好的企业，业务部门人员更懂智能化，甚至一些人员具备软件开发能力。在四级评估答辩时，业务人员将生产优化预测模型讲的更为清晰，包括模型的机理、影响因子、边界条件等，而IT人员更多是描述系统框架、技术应用、功能实现。

当然，业务人员的能力培养过程较为复杂，需要企业建立氛围环境、平台工具、组织活动，甚至建立管理制度予以保障，有效激发员工的智能制造意识。企业IT人员也需要向业务人员传递智能化技术的实现原理，这样业务人员遇到具体问题时，才知道智能化技术能否解决问题，需要具备哪些条件，进而通过智能化技术将日常的知识和经验显性化，提升制造系统的智能化水平。【人员技能3a\3b\4a\4b】提出人员能力提升的多项举措。

在人员能力建设过程需要注意两点。第一是管理者的表率作用，各职能部门主管要深入一线懂智能化，部分企业把操作设备、使用软件作为关键用户的事，管理者高高在上，不使用也不操作，开会还拿线下报表，这样应用智能化的氛围就难以形成。

第二是培养人员的“数据敏感度”，生产大部分时间都是正常运行的，但“与众不同”的数据更值得关注与研究；比如，哪些设备故障的频率较高，哪些设备故障的维修时间较长，哪些设备故障对整体产出的影响较大？这样，对于异常数据的收集、监控、分析，以及触发一些潜在问题的洞察，也会引发业务人员提出智能化建设需求。

五、四级能力建设是一个长期过程

时常听到企业说“我们今年目标是通过三级，明年目标通过四级”。管理者有这样的目标是好的，确实也有企业实现了，但即使通过四级评估，也不代表企业四级建设的终结。

通过CMMM四级评估的企业，四级条款不一定是完全实现，仍存在能力弱项，特别集中在计划与调度、设备管理、能源管理等能力子域。我们也看到，上面提到的组织建设、人才培养也是一个长期持续的过程。此外，四级建设的长期性还表现以下几个方面：

（一）伴随业务变革的智能化升级

制造企业发展经历着变化，比如：产品定位变化（常规大众产品，创新差异产品），产品线扩充（低端、中端、高端），生产工艺上下游延伸（光缆、光纤、预制棒），产线增加（几条线、十几条、几十条），生产模式变化（大批量生产、大规模定制），这些变化都会导致管理“复杂性”增加。而这些变化是受市场需求潜移默化的影响，以及企业发展战略不断调整引发，最终企业发觉现有的智能化能力不够用了，于是“复杂性”的提升导致新的智能化需求提出，触发企业智能化全面升级。但由于企业面临的环境复杂，业务管理模式变化的过程不会很快，且以往的智能化能力都有一定兼容、扩展能力，加之智能化升级的决策会有一定滞后，因此，智能化全面升级需要较长的时间。

（二）源头问题的梳理和探究

智能化建设需要很多的资金投入，进口装备、设计软件需要投入的更多，那是否只要企业有资金、供应商愿意提供技术，大家的制造水平就相近呢？

企业差异化能力体现在管理水平，除了对人的管理，也包括更好的应用已购买的资源，而对于四级企业应再“下沉”一层，探究其制造本源和机理，于是从产品质量提升到工艺参数优化，从生产工艺优化到装备技术改造，从买来应用到完全自主。四级的很多条款就是引导企业不断探究更深层次的机理，【计划与调度4b】对于优化调度的探究，【生产作业4b】对于优化设备参数的探究、【设备管理4a\4b】对于降低设备故障率、提高效率的探究，这个过程伴随着影响因子的识别、量化关系的确定、优化模型的验证等一系列过程。企业能把资源应用好就需要一个过程，再去寻求深层次的突破更需要一个循序渐进的过程。

（三）智能化环境迭代完善

很多企业会发现，在较长一段时间智能化水平并未发生变化，装备和信息系统不断更新，却又功能相近。曾经评估一些企业，ERP软件品牌更换2、3次。企业每次选择的时候，可能只是看中其中一部分功能，后来随着不断的功能追加、定制开发，渐渐对真正的需求越发清晰，最后当平台的技术架构都无法满足时更换平台，或者参与供应商的装备和平台研发，或者企业自己独立构建。

虽然这一过程会导致有些三级建设的内容会推翻重来，但因此平台具备了更强的功能性、扩展性、易用性，也因此一些制造企业在总结工业经验基础上，构建了工业互联网平台。这个过程是在不断夯实智能化底座，为四级建设打好基础，也是促使企业在后续智能化建设中目的更明确、策划更全面、要点更清晰。

六、评估目的

时常有企业咨询，想申请四级评估，但是了解了评估流程后，认为评估时间较长，准备内容多，而企业自身比较大，不好调动各职能部门人员，询问评估过程能否简化，比如只进行正式评估，不进行预评估，或者正式评估能否2天评完等等。CMMM评估遵循统一的评估过程管理规范，因此，这个答案是否定的。

由此引发思考，企业是否正确理解和认识CMMM评估目的？很多企业只当评估是“认定”作用，以此作为申请政策奖励，或彰显智能化水平的荣誉，而忽视了CMMM评估过程对企业智能化建设引导的价值和作用。

CMMM评估更像是一次企业智能化管理咨询，一方面是“评估”验证企业的能力水平，一方面是“建设”，针对存在弱项的改进提升。企业经过多年的智能化建设，应取得了很多成果，可将评估过程作为审视自己能力水平，争取下阶段更大提升的契机。从企业长远发展看，知道能力不足并明确前进道路和方向要比其他面上的荣誉更重要。

（一）申请准备

根据《CMMM标准符合性评估管理规范》规定，对于首次申请CMMM评估的企业，最高等级建议为三级，针对获得过省部级及以上智能制造荣誉的企业，依情况可申请四级。这里建议，所有想评估四级的企业，可以先进行一次三级评估，一方面是了解评估流程和评估难度（三级评估也有相当难度），另一方面，三级评估里涉及的二级、三级条款，有很多是智能制造的基础关键环节，值得企业关注建设，为后续建设打下坚实基础。

曾经有企业在四级评估前，进行了一次三级评估，就是为了检验企业当前智能化建设是否有所遗漏，针对四级建设的基础环境是否扎实，并且在继续四级建设时，能够持续获得人力和资金投入，得到高层领导的支持。当企业三级评估完成，总分超过2.95分（满分3分），说明能力弱项较少，就可以着手四级评估，否则建议先补充三级能力短板，再进行四级建设和评估。

企业确定申请四级评估后，可以向任意一家CMMM Ⅰ类评估服务机构提交《评估申请表》，并签订评估合同。由于四级企业可能涉及的产品类型较多，评估时需选择其中一种产品，即按照《GB/T 4754—2017 国民经济行业分类》中的小类，4位代码，其中一个小类下的产品；并结合企业涉及的业务范围，至少评估15个能力子域，即必评的14个能力子域，至少加1个可选能力子域。

此外，《评估申请表》内容应着重体现企业四级能力建设情况，有些企业在申请表中罗列了很多常规的信息系统，这其实是二级要求，未把企业四级亮点展现出来。

（二）人员准备

对于CMMM四级评估，人员准备非常关键。很多企业想申请四级评估，并不了解四级标准的难度，或者对标准理解有偏差。最好的方式，先不要着急确定申请等级，企业先选派人员参加为期3天的国家标准系统性学习，并考试通过获得结业证书，然后再确定企业申请评估等级。同时，三级、四级评估也要求企业在正式评估阶段，至少配备1名获得结业证书的企业学员，作为“企业内部评估师”加入评估组。

建议企业四级评估最好准备2名学员参加培训，一位来自IT部门，对企业整体智能化情况较为了解，作为评估组的内部评估师；由于内部评估师在答辩问询中不能回答问题，因此，还需再补充一位来自制造部门学员，对企业生产工艺、设备比较了解，作为联络人，对接评估组和企业职能部门。同时，2名经过培训的学员可以形成互补，对企业内部关键人员进行标准宣贯培训，准确解读标准条款要求，做好评估举证准备。还有企业在四级评估前，组织企业全体中层以上员工，完整学习成熟度模型国家标准，后续将CMMM模型在全集团内贯彻应用。

除了内部评估师准备，更关键的是各能力子域的答辩人员确定，因为大部分人不了解标准条款涉及哪些内容，这也是比较难的一步。都说CMMM评估是开卷考试，国家标准也是公开的，评估的就是这些条款，但关键看企业人员能否举出证据，证明这些条款实现了。企业经常感觉，这个条款我们应该实现了，但是怎么也找不到举证的案例，这种情况多半是没有找对人，未能把相应职能的人召集对、召集全。

然而，申请四级的企业，内部业务职能分工有时很细，每个人只负责其中一块职能，加之评估范围广，考核点较细，甚至部门主管都无法清晰回答所有问题。比如说“工艺设计”能力子域的工艺仿真举证，有企业就涉及机加工仿真、焊接仿真、装配仿真，不同工艺类型，不同专业化仿真软件，举证时就需要多个工艺专业的仿真人员，才能分别把仿真过程、仿真模型、仿真目标、仿真方案、仿真结果说清楚。

曾经有企业咨询，能否IT人员全程代为所有能力子域答辩？这个答案是否定的。首先IT人员应用系统的能力，不代表业务部门实际应用系统的能力，部分软件可能未完全导入，业务部门人员用不起来或者不想用；其次，在评估中发现，三级及以下条款，IT人员通常可以清晰回答，但是四级条款涉及很多数据分析模型，IT人员更多了解软件技术架构、运行方式，并不能把分析模型的机理和应用效果讲清楚，而平时功能应用最多的相应业务部门人员，能把模型的机理、输入、输出数据，约束条件配置，以及功能如何操作讲的清晰。

据统计，部分职能分工较细的企业，四级评估要涉及60~80人，平均每个条款1~2个人。由于CMMM四级涉及的业务人员众多，建议最开始，按能力子域对应的业务职能部门分工，部门的主管和技术骨干作为初期选定人员，在举证材料准备时，再结合条款要求，按每个条款分配参与答辩人员。企业在人员组织上做好心理准备，组织协调能力要强，这时评估发起人的选定起到比较重要的作用，由于评估涉及企业几乎所有业务部门，高管有时只负责部分业务条线，因此，评估发起人的选择就很重要，必要时评估发起人需要企业一把手担当。

（三）举证材料准备

举证材料准备是评估准备的重要工作，四级评估重点关注如下几方面。

1、有效证据

提及“举证”材料准备，涉及到一个问题，什么是真正有效的举证？一些初步了解过成熟度评估的企业，可能认为是准备每个能力子域答辩PPT。其实，对于CMMM评估并没有强制要求做PPT，只是PPT形式方便图文并茂的介绍企业智能化实施成效，每页主题鲜明的对应每个标准条款，罗列好每个条款举证哪些证据，重点的功能、特性、数据等要点描述清晰，也方便评估组记录。当然企业为了简便，也可以用Excel电子表格。这会造成一种误解，以为准备好PPT就是准备好答辩材料了。特别在预评估阶段，评估组没有较多时间查证每个系统功能，企业会误以为正式评估采集“证据”也是汇报PPT，把PPT讲明白就能通过。

CMMM评估所要求的证据，是举证企业实际推进智能化建设的应用成果，既包括一些文件化信息，如智能制造规划、异构系统集成规范、统一的数据编码规范等，也包括信息系统功能的操作演示，特别是四级要求的信息系统内数据分析模型应用。文件化信息需要提供原始的文档材料，而不是文件的截图；同样，系统演示是登录系统正式环境，演示功能操作和功能历史数据。从这点来看，PPT作用是有，但没那么重要，特别在正式评估阶段，PPT只相当于一个举证清单、提纲，是信息展示的工具，是让评估组先行了解，接下来要展示这些文件，演示这些功能，把重点内容列出来，防止展示时有所遗漏或表达不清晰。有些企业即使PPT做的很好，但真实系统功能还没有实现，这也不能作为评估的有效证据。

建议企业评估准备时，可将文件化的举证材料复制到网络共享服务器，以备评估查验。同时，将需要演示的信息系统，提前安装好客户端，访问地址保存在收藏夹，准备好演示登录账户（账户权限访问功能不同），并准备好演示的记录数据项ID；比如，想展示哪份采购订单由企业下推至供应商管理系统，将采购订单编号准备好。建议每个条款至少准备2个案例，一是为了保证评估验证的充分性，二是当其中一个案例举证不具效力时，还有其他案例，不至于临时准备仓促。

2、业务背景介绍

有些企业认识到举证较为关键，答辩上来就开始介绍每个条款的证据，但缺少整体业务背景介绍，这样也会影响答辩效果，可能导致评估老师无法理解举证的案例，未明白采用此种智能化解决方案、数据分析模型的作用和意义，甚至难以判断系统功能实现的正确性。例如，企业的生产模式，是按成品库存生产，是按客户订单生产，还是两者都有，这会影响生产计划自动生成的逻辑、输入数据的判断，可能还会影响物料需求计划、销售发货等环节评估。再比如，生产线是24小时连续生产，还是只有白班生产，会影响生产调度方式、设备运维管理等。

建议企业在评估首次会，企业整体智能化介绍时，先将企业的主营产品系列、关键工序、产线布局、生产模式、核心优势讲解清楚。在每个能力子域答辩，介绍本部门或本业务涉及的业务背景，包括不限于：

业务管理概况，如【装备】企业关键工序是哪些，【能源管理】企业能耗种类和高能耗设备届定标准。

完整的业务流程，如【计划与调度】计划编制完整流程（从主计划到各级计划编制），【采购】物料采购完整流程（从物料需求计算到物料入库）；

组织分工，如【产品设计】产品设计分成哪些专业，或哪些功能模块进行任务分工；

业务管控要点与难点，如【仓储配送】物料拉动与配送方式，【生产作业】生产现场关键工艺控制指标等。

业务背景介绍会对后续举证有代入感，但介绍内容也不宜过多，尽可能围绕条款涉及的业务背景，毕竟评估答辩时间有限。

3、举证案例描述

四级条款涉及的系统功能要求相对复杂，而且每家企业的智能化应用实现又不尽相同，很多数据分析优化模型又有很强的业务逻辑，因此，企业在举证描述时应抓住重点、逻辑清晰。

不少企业对于系统内的数据分析模型不知道怎么介绍，或是只简单介绍系统的功能操作“点下按钮就自动生成了”。具体分析模型应用到哪些生产环节，优化了哪些目标也没有介绍。

这里建议模型应用从如下方面介绍：

1.模型应用解决什么问题，预期效果是什么；

2.模型相关的数据指标是如何发现、确定的；

3.模型的输入数据包括哪些，数据的来源于哪些系统、设备，数据更新频率，实时还是历史；

4.模型的运算逻辑或公式，以及需要提前配置的参考变量、约束条件、判别边界条件；

5.模型的执行逻辑是手动触发，还是定时自动触发，或是依靠何种逻辑触发；

6.模型输出结果是什么，以何种形式展现，比如是单次还是连续，是数据列表或是趋势曲线；

7.以上功能如何在信息系统中操作实现。

特别注意最后一点，有很多企业理解了条款要求，业务原理讲的也很清晰，技术方案也行得通，但评估要求的是已实现的成果，最终还是要落实到软件系统中实现，而不是纸上谈兵的方案。

4、建设成果有效性

相比于其他等级，四级有些智能化项目是从研究出发，开创性、尝试性的，其结果可能未必发挥出效果，还需要反复迭代验证，企业有必要对已取得的应用效果做介绍。

曾经有家企业实施设备故障预测性维护，根据预测曲线，判断设备将在半年后出现故障，即很长一段时间内设备都是正常工作的；如按此逻辑，设备故障的频率不高、故障影响也不大，那么企业构建预测模型的投入和意义在哪里？然而，实际在设备管理系统查看此类型设备的历史故障记录，却发现每月设备故障率确实较高，因此判定模型考虑的影响因素缺失，判别精度不够，预测的准确性有待提升。

当然，还是鼓励企业做这种尝试，模型还需要迭代验证过程，当前可能还未成功应用或是效果不佳，评估组在打分上会有所权衡。

综上，企业准备举证材料需要付出一定精力，从另一个角度看，这也是企业一次很好的阶段性总结，很多企业通过评估准备的梳理，恍然发现原来自己有这么多优势能力，增强了企业智能化建设信心，这也是CMMM评估的另一种意义。

5、预评估后整改

预评估结束后，企业会有一段时间整改，为正式评估做好准备。部分企业认为预评估过程比较顺利，正式评估也不会出大问题，但实际从评估查证的细致、严谨程度来讲，正式评估需要准备的东西更多。

预评估阶段，企业大致了解评估组是否认可PPT描述的案例作为有效证据，以及可能的弱项在哪些方面。针对这些问题，不是简单的修改PPT就完事了。正式评估不是查证PPT，而是PPT介绍背后对应的实现成果。

因此，建议将PPT提到的智能化应用，模拟操作演示一遍。有些信息系统功能如果需要完善，还需要一定开发建设，如果企业不具备开发能力，还要尽快得到解决方案供应商的技术支持，因为整改时间有限，需要明确功能完善的优先级。还有些PPT里提到的文件材料截图是临时从其他材料摘抄过来，应找到原始文件，以备正式评估查验。

最后，企业通过四级评估不是终点，已通过四级的企业也有弱项需要持续改进。通过现场评估，我们也发现了很多优秀的CMMM服务商，为企业提供智能制造咨询以及软件、装备、系统解决方案。下一步，CMMM评估工作组秘书处将发挥评估优势，提炼一批具有成功实践经验的CMMM服务商资源库，供企业精准选择适配。

CMMM评估是企业智能化建设的其中一面镜子，镜子只照到有限的一面，企业还有更多的目标去攻关。希望企业用好CMMM，在智能制造建设过程中不断创新，脚踏实地，也希望有更多的企业通过CMMM四级评估。

七、四级能力建设的核心思想

 企业四级能力建设有哪些方向可以引导呢，这里总结为16个字“降低复杂、动态平衡、整合资源、抵御风险”。

（一）降低复杂

随着企业发展，制造“复杂性”都将从各维度提升，四级企业面临的问题更加突出，如果复杂性未能有效解决，就会降低企业运行效率。因此，降低复杂就是应用智能化，使我们的制造过程更简单、更顺畅、更敏捷，不管是产品研发，还是生产排程、生产控制等方面。这就需要根据复杂场景，进行知识与经验总结，辅助业务活动快速处理。比如，【产品设计4a】面对系列化产品的设计复杂性，采用模块化、参数化设计加快产品研发，【工艺设计4b】面对典型特征的工艺相似性，利用工艺知识为工艺设计与规划提供决策，【生产作业4a】面对多品种产品混线生产，根据产品与程序匹配实现生产指令自动下发，【计划与调度4a】面对多任务、多资源、多约束的复杂性，实现作业计划自动优化排产调度；以及在【集成4a】面对信息系统间点对点集成的复杂性，采用ESB、ODS等技术路线，使集成运维工作更为敏捷。

（二）动态平衡

企业生产过程一直受到外部环境、内部资源变化影响，不可能一直按既定的路线、流程、模式等正常环境生产出产品，指标异常就在“不确定”的时候出现，于是整个生产过程就处于一方面尽可能保持正常，一方面又尽快的在异常中恢复，这样交叉进行，智能化要解决的就是这种动态平衡能力。

有时这些异常有明确的指标判断，显性异常，比如产品质量指标超差，还有的是隐形异常，比如工艺指标趋势表现偏离。不论是哪种异常，首先都需要建立指标参照体系，监控哪些指标，指标在何种范围是正常的，是结果性的，是趋势性的，还是关联性的。在国标三级条款里就是要建立这些指标并进行监控，如【设备管理3b】设备OEE指标、【能源管理3a】高能耗设备能耗评价指标、【计划与调度3c】生产进度指标，指标制订的合理性本身也是数据分析的过程。

四级的动态平衡就是要针对这些指标的异常结果做出反馈，消除这些“不确定”带来的复杂性，进行敏捷响应，实现短周期平衡；而制造系统经过一定时间运行，指标值不断验证，指标类型及其边界需要重新校正调整，优化使其更为合理，进行长周期平衡闭环。

短周期动态平衡，比如【设备管理4b】对设备OEE异常的设备进行分析，确定影响运行效率的因素（单元供料不及时、前后工序节拍不平衡、设备工艺参数不合理），自动进行工艺和计划调节。【计划与调度4b】面对物料短缺、质量异常、设备宕机等各类异常情况，自动进行生产调度调节，使其恢复到正常生产状态。长周期动态平衡，比如【采购4c】在积累了供应商历史供货表现后，优化供应商评价模型，使其更能反应供应商供货绩效；【物流4c】结合运力资源能力、运送节点、运输任务优化资源装载能力和运输路径，降低运输成本，提高物流效率。

（三）整合资源

这里“资源”的范围较广，不只是企业的物料、设备、工装、能源、软件、数据等，还包括企业内部的组织、岗位、业务活动。【组织战略3b】提到对岗位结构的优化和调整，就是结合企业智能化建设，进行相应岗位的职能调整与岗位结构的整合。

一方面，通过平台围绕产品全生命周期整合数据，比如【数据4a】企业级的数据中心，【产品设计4b】实现设计模型在内的全面数据归档，【产品服务4c】将产品使用、运行过程的数据整合，并优化产品运行性能。另一方面，通过智能化应用实现各职能岗位的业务活动协同、传递信息，整合组织内部资源。比如，【产品设计4d】整合产品研发与生产、物流、销售、服务业务协同需要的BOM等资源，【销售4b】整合线上与线下销售渠道与需求资源等。

此外，相比在三级（集成级）重点关注的软件系统和硬件设备间的数据集成，在四级里更强调实时的“软硬一体”无缝整合。比如，【仓储配送4a】基于生产调度指令，实现物料从仓储区域到生产区域的自动配送，这一过程涉及生产管理系统、排程调度系统、仓储系统、配送系统等软件系统与数字化仓储设备、数字化配送设备、生产单元设备等硬件系统的联动交互，交互的每个动作包含着指令下达、收到回应、驱动执行、执行完毕等消息传递，软、硬件频繁交互也将生产组织和设备运行紧密结合一起；同样，【计划与调度4b】对于生产异常的实时优化调度，需要软硬一体的模式实现。

注：文中出现【计划与调度4b】表达形式，代表国家标准GB/T 39116-2020《智能制造能力成熟度模型》中，“计划与调度”能力子域，四级，b条款，即“应实时监控各生产要素，系统实现对异常情况的自动决策和优化调度”

（四）抵御风险

相比“动态平衡”消除已发生的异常，抵御风险是赋予制造系统的预知能力。提前知道风险，防范潜在问题发生，将损失降到最低，即四级标准提到很多的“预测”。

为了构建风险预测模型，首先需要找到表征风险的指标，这需要基于日常经验与知识的积累，也就是三级条款提到的“知识库”，同时要保证“指标”是可实时采集或度量的。其次，需要给出指标趋势变化的边界范围和各级临界值，量化风险大小。然后，针对风险采取的措施，建立解决方案对策模型，将风险控制在一定范围内。最后与实际情况对比，验证模型对风险预测的准确性，并不断优化模型。

有些指标可以直接预示风险，如监控质量特性指标的趋势变化；有些与风险产生存在关联关系，间接预示风险，如通过工艺指标的趋势变化预测产品质量的变化，有些关联还由多重因子导致，比如【设备管理4a】设备故障预测性维护，可能涉及设备关键部件多指标运行表现、外界环境影响、产品加工参数等。

四级条款强调的就是企业抵御风险，建立预警预测模型的能力。比如，【采购4b】采购风险模型、【生产作业4c】质量预测模型，【安全环保4a】安全风险的动态识别、【安全环保4b】环保排放风险，【仓储配送4c】罐区冒罐风险，以及【信息安全4c】工业信息安全风险，提升主动防御能力。同时，在【产品服务4a\4b】将风险预测延伸到产品运行使用环节，感知产品运行的潜在风险，比如对电梯、发动机产品建立远程运维监控与优化性能服务，在抵御运行安全风险的同时，增加了产品附加值。

综上，企业四级能力建设是一个长期过程，其智能化建设成果代表了企业深耕行业的看家本领，也是区别于其他企业的优势能力，其应用成效也会确保企业处于行业竞争领先地位。

作者简介

杨卓峰，工学博士，CMMM®高成熟主任评估师、培训讲师；《智能制造能力成熟度模型》相关标准主要起草人之一，曾主持航空航天、工程机械、压力容器、汽车、电子等行业多家企业信息化整体规划与智能制造解决方案设计工作，具有丰富的业务分析、流程优化及企业两化融合工作经验，精通企业信息管理系统设计、开发和实施。

翻译：

[May • Online] Notice on the development of intelligent manufacturing capability Maturity Model training course

“Intelligent manufacturing capability maturity standard compliance assessment (hereinafter referred to as” CMMM assessment “) is based on GB/T 39116-2020 “Intelligent Manufacturing Capability Maturity Model”, GB/T 39117-2020 “Intelligent Manufacturing capability Maturity Assessment Method” two national standards, Systematically evaluate the degree of intelligence of the product manufacturing process. By the end of 2022, more than 400 manufacturing enterprises in the country have obtained the corresponding level of standard compliance certificate, of which “Level 4 (optimization level)” is the highest level currently achieved by domestic manufacturing enterprises, and the enterprises that have passed the evaluation are also typical representatives of domestic high-end manufacturing.

So how do enterprises identify whether they can reach Level 4? What capabilities do you need to build a Level 4 enterprise? What are the key points to focus on when preparing for a Level 4 assessment compared to other levels? In response to the problems that we are concerned about, the author summarizes the evaluation activities of more than ten four-level enterprises and shares the following experience. It is hoped that in the aspects of enterprise intelligent ability identification, capacity building, evaluation preparation, etc., it will be helpful to manufacturing enterprises applying for Level 4 evaluation.

Overview of four-level enterprises

Level 4 (optimization level) : Enterprises should conduct data mining for personnel, resources, manufacturing, etc., to form knowledge, models, etc., to achieve accurate prediction and optimization of core business activities.

–GB/T 39116-2020 “Intelligent Manufacturing Capability Maturity Model”

Level 4 capability refers to the extensive application of data collected by various production links and business functions on the basis of achieving level 3 (integration level), namely digital production, inspection and logistics equipment networking integration, information integration among application information systems, and interconnection, data sharing and business collaboration among all production links of the enterprise. Assist enterprise personnel to make more scientific decisions, or automatically optimize the production process.

By the end of 2022, 29 domestic enterprises have obtained CMMM Level 4 certificate, accounting for 7.25% of all certified enterprises; Among them, Jiangsu, Guangdong and Shandong, which are major manufacturing provinces, have 9, 8 and 4 manufacturing enterprises that have passed the four-level assessment respectively, and Beijing, Anhui, Fujian, Shaanxi, Hunan, Guangxi, Liaoning and other provinces and cities also have local leading enterprises that have passed the four-level assessment.

From the perspective of industry distribution, some leading enterprises in the industry have reached the international leading level.

For example, the household appliance manufacturing industry represented by Cygnet and Haier, the printed circuit board manufacturing industry represented by Shennan Circuit, the construction machinery industry represented by Xugong, the LCD panel manufacturing industry represented by Huaxing Optoelectronics and boe, the rail transportation equipment manufacturing industry represented by CRRC, and other petroleum refining, cement, and pharmaceutical manufacturing industries. At the same time, nearly 50% of fourth-level enterprises have no local policy incentive support, and apply for assessment for enterprises spontaneously. It can be seen that more and more leading enterprises in the industry hope that their own intelligent construction achievements are recognized by the authority, or seek to continue to improve the direction in the evaluation.

Judging from the current assessment results of Level 4 enterprises, enterprises have certain weak capabilities, and most enterprises are still in the initial stage of Level 4. Although analysis models have been built on some business single points, the application effect of the models needs further verification, and the capacity building of Level 4 still needs a long way to go.

Common misunderstandings in applying for Level 4 assessment

In many enterprises applying for the four-level assessment, it is found that some enterprises do not fully understand the application for the four-level assessment, and do not judge the dimensions, difficulties and requirements of the four-level assessment. Enterprises blindly apply for Level 4, resulting in a downgrade of the final assessment conclusion. For the identification of four levels of ability, the following misunderstandings often appear.

Misconception 1: Large enterprises/industries with high brand awareness can pass the level 4

This misunderstanding is more common, because most of the well-known enterprises in the industry have been certified. The well-known enterprises in the industry are indeed in a leading position in production and sales, turnover, market share, and brand image. And the enterprises have strong funds and intelligent investment. However, the CMMM assessment evaluates the actual application of intelligent product manufacturing process. The brand marketing of some enterprises is greater than their actual manufacturing level and intelligent level. Some brands rely on external processing, there is no manufacturing plant, can not be evaluated.

I have encountered a well-known non-standard equipment manufacturing enterprise in the industry. Applying for four-level assessment, the type of production process is only assembly. The product assembly process uses a full manual way, and the assembly process is made of screws or ordinary electric screwdrivers, can not achieve the data acquisition of torque parameters. Other intelligent applications are less, the final pre-evaluation is marked down. The evaluation team believes that it is still difficult for enterprises to reach the third level. And the overall level is at the second level. The enterprise can not accept as the industry leader and the evaluation conclusion is second level. And finally terminate the evaluation.

Misconception 2: You can pass Level 4 if you have obtained the honor or authority of intelligent manufacturing

Some enterprises have won the honor of intelligent manufacturing at the provincial and ministerial level, national level, and even international. Which is a good basis for intelligent construction. However, due to the different evaluation indicators of various policies and honor selection. It cannot be said that obtaining a certain honor can be equivalent to passing CMMM level 4.

In particular, some enterprises are easy to take this honor and produce blind confidence. While ignoring the comprehensiveness, objectivity and rigor of CMMM evaluation investigation and evidence collection. For example, the enterprise undertook the national key intelligent research construction project, has been checked and accepted by authoritative experts. It is believed that there is no problem through the fourth level. However, in the four-level assessment. The project may only be used as effective proof of several provisions in the subdomain of the four-level assessment. And it cannot be used to represent the overall intelligence level of the enterprise.

The usual enterprise intelligent results declaration, declaration materials are as far as possible to show the good side of the enterprise, and do not mention the weak.

For CMMM evaluation, enterprises should confirm the scope of evaluation and evaluation content based on the actual business situation. And the terms that should be evaluated should not be cut or avoided at will. And should not avoid the heavy and light, and CMMM evaluation should be carried out based on unified standards. A complete CMMM four-level assessment includes two stages of pre-assessment and formal assessment. At least 6 days in the enterprise site, the assessment team takes full time to verify the intelligent application and application effect of the various business functions of the enterprise in a complete and detailed manner. The investigation and evidence collection here includes the advantages and construction highlights of the enterprise. As well as the lack of capacity and construction after the standard.

Misconception 3: Newly built smart factories can pass Level 4

Here, “newly built” refers to a factory that has been in operation for about one year. Many companies have invested hundreds of millions of dollars to build a new model of product manufacturing factory. And the factory is built in accordance with the highest standards of the industry at the beginning of design and construction, drawing on past experience, various intelligent technologies in one step, and creating the most advanced smart factory in the industry. Do not deny the advanced nature of these factories. But usually the output and sales of newly built factories are still slowly climbing the process. And the types of manufacturing products are limited, especially the accumulation of many new technologies in intelligent application. The overall operation of the factory takes a period of time to run in, digest, and verify, and it is difficult to play the characteristics of smart factories in the short term.

In addition, most of the four-level assessment points are to review online optimization and prediction based on data analysis models.

Due to the short construction time of factories, the accumulation of business data in many factories is not enough. And some manufacturing laws are difficult to find and express and verify through analysis models. For example, the “predictive maintenance of equipment” required by the subdomain level 4 of [equipment management] requires the accumulation and analysis of fault knowledge of historical failures. And the newly built factory is new equipment, and the short-term failure rate is less. And it is difficult to put forward the demand for prediction and optimization.

Level 4 [energy management] capacity subdomain requirements of the refined “energy saving model”. The new factory operation at the beginning of the focus on production, quality, efficiency. So intelligent construction will focus on these aspects, short-term rarely consider energy saving measures. When the factory production is stable, in order to further reduce manufacturing costs. It will be intelligent construction demand. Of course, there are also some new factories. At the beginning of the construction of the factory has taken into account the previous energy-saving experience, has adopted energy-saving equipment and programs. And then need to further explore new energy-saving space.

Therefore, the factory assessed by CMMM Level 4 needs to run steadily for a period of time. The production and operation mode is more mature, and the application of intelligent technology is better. When the factory runs and accumulates certain data and precipitates knowledge. There are needs and measures to optimize and improve. The intelligent results of Level 4 will also be well applied and displayed.

Misconception 4: The factory with good production operation and display effect can reach level four

Many people have been to such a factory, the factory is clean and bright, the process layout is orderly. The equipment is fully automated, the robot collaborative operation, the logistics distribution is unmanned. And the command and dispatch hall is a large screen Kanban data display, which looks very advanced.

Some factories also specify “red carpet” routes, sightseeing car routes, equipped with full-time docent, each place to explain the highlights.

First of all, this kind of visit can only understand the part of the factory. The fixed speech of the docents may not be the evidence that needs to be collected for standard conformity assessment. And some of the main points of CMMM evaluation need to be verified by the assessment team to the production station in detail. Such as whether the equipment automatically switches processing procedures when the product is changed.

Secondly, the above mentioned high automation of production, large screen Kanban display, etc. Most of them are second – and third-level requirements, data analysis optimization and prediction models required by fourth-level clauses. And these underlying logic are also difficult to show with large screen Kanban. There are also some large screen Kanban just to “show”, the timeliness, accuracy and authenticity of the data displayed on the Kanban can not withstand scrutiny. And some statistical data is difficult to guide and support production scheduling and production control. With the development of mobile technology, a lot of real-time information does not necessarily need to be put on the large screen. And pushing information through mobile terminals and mobile apps, especially abnormal information, is more meaningful than the large screen.

Therefore, more explicit three-level requirements are verified when the production workshop is inspected in the four-level assessment. For level 4 clauses, some real-time production process optimization control must be verified by the background “brain” or data analysis platform of the general control. And some decision models have been integrated into the underlying control logic of equipment operation. The factory with good production site display effect may not be able to reach level 4.

Misconception 5: The manufacturing plant of high-end intelligent terminal products is easy to reach level 4

High-end and intelligent products contain more high-tech, long product development cycle. Some manufacturing processes are more complex. And enterprises invest more in product development and intelligent application of process research and development. With the strong market demand, it also drives enterprises to improve the intelligent application in the production process. For complex products, the social division of labor in product manufacturing is more obvious. And each factory is responsible for the manufacturing of key parts, or the processing of several key processes. And then the terminal manufacturer assembler all parts.

However, due to the high brand exposure of the terminal manufacturer. It is easy to misunderstand that all technologies come from the terminal manufacturer. However, in some industries, it will be found that the level of enterprise manufacturing technology and intelligent application level. And some key component manufacturers are higher than terminal product manufacturers. Such as LCD panel manufacturers and display and TV manufacturers.

Of course, this is not absolute, many end product manufacturers of intelligent level is also high. The core parts of the whole machine is also their own production. Large complex product assembly process also needs a high level of intelligent technology support.

In view of the above common misunderstandings, enterprises need to maintain a “return to zero” mentality, compare the standards. And carefully and objectively summarize the advantages of intelligent capabilities.

Four enterprise identification direction

In order to accurately identify enterprise level 4 capabilities, it is necessary to judge the conformity of enterprise intelligent application against standard provisions. Of course, enterprises can also preliminarily identify level 4 capabilities from several other aspects, and several directions are provided here.

(1) The complexity of product manufacturing

Starting from the intelligent demand measurement. The source power that drives enterprises to use intelligent technology in the manufacturing process lies in the complexity of the manufacturing process. The complexity here includes many aspects. Such as the complexity of product structure, the complexity of manufacturing processes. The diversity of product types, the diversity of market demand, and the complexity of production organizations. On the one hand, the solution of these complexities promotes the functional division of enterprises to become more and more detailed. And the number of post personnel increases, and on the other hand. It also promotes the continuous investment in intelligent construction to reduce manpower. It can be said that the complexity of manufacturing problems faced by enterprises and the ability to solve complex problems represent the potential of intelligent application of enterprises. Which is also reflected in the unique manufacturing characteristics of each industry.

From this point of view, the level 4 assessment is somewhat selective. Although some enterprises have achieved the segmentation of the industry leader. But the product is not complex, the production process is not complex, the factory produces a limited number of products. The market demand is more fixed, the production management is relatively simple. So that it does not need much high intelligent construction, the overall intelligent degree of the industry is roughly equivalent. Enterprises also have the will to improve, but the “complexity” of their own manufacturing environment is not enough. Some intelligent technology is not used, there is no “demand”. Usually the market competition in these industries is also relatively fierce.

(2) Industry innovation leading ability

Level 4 enterprises are generally at the forefront of the industry. In order to ensure their industry status, in the sense of crisis will take the initiative to open the gap with peers, forcing themselves to continue to improve, but because there is no reference to the target. There is no case for reference, must constantly stimulate their own innovation ability. Many of the data analysis optimization and prediction models required by the level 4 evaluation. And the final application scenarios of the enterprise are closely related to the characteristics of the industry manufacturing. And the problems solved are mostly existing in the industry, but have been difficult to solve, and are the first company in the industry to tackle.

In terms of the guidance of standard provisions. Many provisions at the second and third levels are more clear, giving clear guidance to the intelligent construction of enterprises. And the application scenario is applicable to most industries, and many excellent cases can learn from each other across industries. The four-level clauses, there are many directional guidelines. The description is relatively brief, which also gives each enterprise more room for innovation. For example, [Warehousing and Distribution 4b] “Storage model to achieve inventory optimization”. Where inventory optimization can be the optimization of material inventory level. Or the optimization of inbound and outbound storage efficiency. Or the optimization of inventory capacity. At the same time, combined with the production business process. The optimized parts can be in raw material warehouses, semi-finished products warehouses and finished products warehouses.

(3) Data application environment

There are many four-level enterprise information systems. Some of which are promoted by the group in a unified way. And some of which are purchased by enterprises and business divisions on their own, often using dozens of information systems. Level 4 enterprises have realized that the system is not the ultimate goal, more important is to integrate the data in the system, especially the use of related data. At the same time, enterprises carry out data governance in the three-level construction, standardize data coding specifications and data exchange formats, and also lay the foundation for the application of four-level data.

Level 4 enterprises need to build a data application environment to carry out business data application analysis. In the face of many system data sources, the first thing is to establish a data aggregation platform (data warehouse, data center, data lake), to achieve the integration of data from all parties. And to clarify data sources, data formats, data synchronization frequency, and so on. According to the needs of the production business, the data platform may be 1 or more. And the data sources are various, some from the business application system. Some from the digital production equipment. The data used for subsequent data analysis also needs to be cleaned and converted within the platform as necessary. Discover the correlation and quantification of data through data analysis application tools. Then the production data optimization and prediction model are established and deployed in the manufacturing system.

(4) Intelligent talents

The four-level intelligent ability is also reflected in the scale and ability of intelligent talents possessed by the enterprise. The number of IT hardware and software maintenance personnel is as few as dozens and as many as hundreds. In addition to system operation and maintenance. The above mentioned data governance and data analysis work all require personnel staffing for corresponding positions.

In addition, the requirements for intelligent talents are not only IT personnel. But also technical backbone from various business departments, these personnel can convert the usual business difficulties into intelligent needs, with intelligent ability and awareness, and the construction of production optimization and prediction models required by the four-level clauses can not be separated from the exploration and practice of production business personnel.

In summary, as more enterprises pass the CMMM Level 4 assessment, the identification of Level 4 capabilities will become clearer. When it is difficult for enterprises to judge, they can invite appraisers who have participated in CMMM evaluation. Especially appraisers who have participated in level 4 to conduct readiness checks. And investigate the effectiveness of intelligent implementation of enterprises against the level 4 provisions of GB/T 39116-2020 “Intelligent Manufacturing Capability Maturity Model”.

The main body of the construction of the fourth and fourth level capacity is the enterprise itself

In the past three-level construction, enterprises can quickly implement projects through capital investment and external implementation forces. And most enterprises can see obvious results after completing the deployment and verification of information systems and digital equipment. However, the difficulty of level 4 is that there are few successful cases for reference. Many enterprises in this stage began to be confused, unsure of the direction, the direction is not sure how to achieve. And it is difficult to verify the effect or achieve the expected effect.

Because the enterprises are in different industries, the strategic focus is different, and the intelligent needs are different. External solution providers, however, can hardly deeply understand the “complexity” of the enterprise itself (see the complexity of product manufacturing in the previous part). And external solutions are more about providing ideas, methods, tools, and system frameworks. The exploration of “complexity”, enterprises have a deeper understanding. So the initiative of four-level construction into the hands of enterprises, with the help of external tools, absorption into the internal capabilities of enterprises.

An important step for enterprises to become the main body of intelligent construction is the construction of enterprise intelligent talent system.

There may be misunderstandings here that intelligent talents are mainly enterprise IT personnel. Especially for some enterprises from the second and third level capacity building. IT has been acting as the main force of intelligent construction, but to a certain stage of construction. IT personnel will feel that intelligent construction has stayed in place for a long time, daily is to constantly maintain the system, or think that all aspects of enterprise construction are perfect. No new construction projects. Some enterprises will also take the initiative to collect intelligent construction needs from various business departments. But find that the business departments can not provide the needs. The development of enterprises to the fourth level appears bottleneck.

The key to solving the problem is blending the capabilities of the IT staff with those of the manufacturing operations. IT personnel need to understand business, and all functional personnel need to understand intelligence. But in fact, IT is also found that some IT staff are busy maintaining the system on weekdays. And have little time to sink down to understand the complexity of the business. And there are many businesses in the enterprise. Therefore, it is more necessary for business function personnel to understand intelligence and cultivate intelligent consciousness.

In the evaluation process, it is often found that the intelligent construction of enterprises, business department personnel are more intelligent. And even some personnel have software development capabilities. In the defense of Level 4 evaluation, business personnel will explain the production optimization prediction model more clearly. Including the mechanism of the model, impact factors, boundary conditions, etc. While IT personnel will describe the system framework, technical application, and function realization.

Of course, the ability training process of business personnel is more complex

Of course, the ability training process of business personnel is more complex. And enterprises need to establish an atmosphere, platform tools, organizational activities. And even establish management systems to ensure that employees’ awareness of intelligent manufacturing is effectively stimulated. Enterprise IT personnel also need to transfer the implementation principle of intelligent technology to business personnel. So that when business personnel encounter specific problems, they know whether intelligent technology can solve the problem. What conditions are required, and then through intelligent technology to make daily knowledge and experience explicit, improve the intelligent level of manufacturing system. [Personnel skills 3a\3b\4a\4b] Put forward a number of measures to improve personnel ability.

Two points need to be noted in the process of personnel capacity-building. The first is the role of the manager, the head of the functional department should go deep into the front line to understand intelligence, some enterprises to operate equipment. The use of software as a key user, the manager is high, do not use and do not operate. The meeting also take offline reports, so the application of intelligent atmosphere is difficult to form.

The second is to cultivate personnel’s “data sensitivity”. Production runs normally most of the time, but “different” data is more worthy of attention and research. For example, which equipment has a higher frequency of failure, which equipment failures take longer to maintain. And which equipment failures have a greater impact on the overall output? In this way, the collection, monitoring and analysis of abnormal data. As well as the insight that triggers some potential problems, will also cause business personnel to propose intelligent construction needs.

Capacity-building at level 4 is a long-term process

It is often heard that enterprises say, “Our goal is to pass level 3 this year. And our goal is to pass level 4 next year.” It is good that managers have such a goal, and indeed some enterprises have achieved it. But even if they pass the four-level assessment, it does not mean the end of the four-level construction of enterprises.

For enterprises that have passed the CMMM four-level assessment. The four-level provisions are not necessarily fully realized, and there are still capacity weaknesses. Especially in the sub-fields of planning and scheduling, equipment management, and energy management. We can also see that the organizational construction and personnel training mentioned above are also a long-term and sustained process. In addition, the long-term nature of the four-level construction is also manifested in the following aspects:

(1) Intelligent upgrades accompanied by business changes

The development of manufacturing enterprises is undergoing changes. Such as: Changes in product positioning (conventional mass products, innovative differentiated products), product line expansion (low-end, mid-range, high-end), upstream and downstream extension of production processes (optical fiber, optical fiber, prefabrication rods), increase in production lines (several lines, dozens, dozens). And changes in production modes (mass production, mass customization) will all lead to an increase in management “complexity”.

These changes are caused by the subtle impact of market demand. As well as the continuous adjustment of enterprise development strategy. And finally the enterprise finds that the existing intelligent ability is not enough. So the improvement of “complexity” leads to new intelligent needs, triggering the comprehensive upgrade of enterprise intelligence. However, due to the complex environment faced by enterprises, the process of business management model change will not be very fast. And the previous intelligent capabilities have certain compatibility and expansion capabilities. And the decision of intelligent upgrade will lag behind, so the comprehensive intelligent upgrade will take a long time.

(2) Sorting out and exploring the source problems

Intelligent construction needs a lot of capital investment, imported equipment, design software needs to invest more. Then whether as long as the enterprise has funds, suppliers are willing to provide technology, everyone’s manufacturing level is similar?

The differentiation ability of enterprises is reflected in the management level, in addition to the management of people. But also includes the better application of purchased resources. And for fourth-level enterprises should “sink” a layer, explore the manufacturing source and mechanism. So from product quality improvement to process parameter optimization, from production process optimization to equipment technology transformation, from purchase application to complete autonomy.

Many provisions of Level 4 are to guide enterprises to continue to explore deeper mechanisms, [planning and scheduling 4b] to optimize scheduling, [production operation 4b] to optimize equipment parameters, [equipment management 4a\4b] to reduce equipment failure rate and improve efficiency. This process is accompanied by a series of processes such as the identification of impact factors. The determination of quantitative relations, and the verification of optimization models. It takes a process for an enterprise to apply its resources well. And then it needs a gradual process to seek a deep breakthrough.

(3) Iteration and improvement of intelligent environment

Many enterprises will find that the level of intelligence has not changed in a long period of time. And equipment and information systems are constantly updated, but they have similar functions. Once evaluated some enterprises, the ERP software brand changed 2 or 3 times. Every time an enterprise chooses, it may only take a fancy to some of the functions. Later, with the continuous function addition and custom development, it gradually becomes clearer and clearer to the real needs. Finally, when the technical architecture of the platform cannot be met. It will change the platform, or participate in the equipment and platform research and development of the supplier. Or the enterprise will build it independently.

Although this process will lead to some three-level construction content will be overturned,. But therefore the platform has stronger functionality, scalability, ease of use. And therefore some manufacturing enterprises in the summary of industrial experience on the basis of building an industrial Internet platform. This process is to continuously consolidate the intelligent base, lay a good foundation for the four-level construction. And also promote the enterprise to have a clearer purpose, more comprehensive planning and clearer points in the subsequent intelligent construction.

Purpose of evaluation

Often, some enterprises want to apply for the four-level assessment. But after understanding the assessment process, they think that the assessment time is long and the preparation content is much. And the enterprise itself is relatively large, it is difficult to mobilize the staff of various functional departments. They ask whether the assessment process can be simplified. Such as only formal assessment, no pre-assessment, or whether the formal assessment can be completed in 2 days. The CMMM evaluation follows a uniform evaluation process management specification, so the answer is no.

This leads to thinking, does the enterprise correctly understand and understand the purpose of CMMM evaluation? Many enterprises only regard the evaluation as the “identification” function. As an application policy reward, or to highlight the honor of the intelligence level. And ignore the value and role of the CMMM evaluation process to guide the intelligent construction of enterprises.

CMMM evaluation is more like an enterprise intelligent management consulting. On the one hand, it is “assessment” to verify the ability level of the enterprise. On the other hand, it is “construction”, and the improvement and improvement of the existing weak items. After years of intelligent construction, enterprises should have achieved a lot of results. And the evaluation process can be used as an opportunity to examine their own ability level and strive for greater improvement in the next stage. From the perspective of the long-term development of the enterprise, knowing the lack of ability and making clear the way forward and direction are more important than other aspects of honor.

(1) Application preparation

According to the provisions of the “CMMM standard conformity assessment management Specification”. For the first time to apply for the CMMM assessment of the enterprise, the highest level is recommended to be three. For the provincial and ministerial level and above intelligent manufacturing honor enterprises, depending on the situation can apply for four. It is suggested here that all enterprises that want to evaluate the four levels can first carry out a three-level evaluation. On the one hand, to understand the evaluation process and evaluation difficulty (three-level evaluation is also quite difficult). On the other hand, the three-level evaluation involved in the second and third level clauses. Many of which are key links in the foundation of intelligent manufacturing, worthy of enterprise attention and construction. And lay a solid foundation for subsequent construction.

There have been enterprises in the four-level assessment, carried out a three-level assessment. In order to test whether the current intelligent construction of the enterprise is missing. For the four-level construction of the basic environment is solid. And in the continuation of the four-level construction, can continue to get human and capital investment, get the support of senior leaders. When the three-level assessment of the enterprise is completed and the total score exceeds 2.95 points (full score: 3). It indicates that there are fewer weak points, and the four-level assessment can be started. Otherwise, it is recommended to supplement the three-level shortcomings first, and then carry out the four-level construction and assessment.

After the enterprise determines to apply for the four-level evaluation, it can submit the Evaluation Application Form to any CMMM Ⅰ evaluation service agency and sign the evaluation contract.

Because there are more product types involved in the four-level enterprises. One of the products needs to be selected during the evaluation. That is, according to the sub-category in the “GB/T 4754-2017 National Economic Industry Classification”, the 4-digit code, which is the product under one sub-category; Combined with the business scope involved in the enterprise, at least 15 capability subdomains are evaluated, that is, 14 capability subdomains are required to be evaluated, and at least 1 optional capability subdomain is added.

In addition, the content of the “evaluation application form” should focus on the enterprise level 4 capacity building, some enterprises listed a lot of conventional information systems in the application form, which is actually a second-level requirement, and did not show the highlights of the enterprise level 4.

(2) Personnel preparation

For the CMMM Level 4 assessment, personnel preparation is critical. Many enterprises want to apply for Level 4 assessment, do not understand the difficulty of level 4 standards. Or there is a deviation in the understanding of standards. The best way is not to rush to determine the application level, the enterprise first selects personnel to participate in the three-day national standard systematic study, and passes the exam to obtain the completion certificate, and then determines the enterprise application evaluation level. At the same time, the third and fourth level assessment also requires the enterprise to equip at least one enterprise student who has obtained the completion certificate to join the assessment team as an “internal enterprise appraiser” in the formal assessment stage.

IT is suggested that two students should be prepared to participate in the training for the enterprise level 4 assessment. One is from the IT department, who has a better understanding of the overall intelligence of the enterprise and is the internal appraiser of the assessment team. Since the internal appraiser cannot answer the questions in the defense inquiry, it is necessary to add another student from the manufacturing department who has a better understanding of the production process and equipment of the enterprise and acts as the liaison person to connect the assessment team with the functional departments of the enterprise.

At the same time, two trained students can form complementary, conduct standard publicity and consistent training for key personnel within the enterprise, accurately interpret the requirements of the standard clauses, and prepare for assessment and proof. In some enterprises, before the level 4 assessment, all middle-level and above employees of the enterprise are organized to fully learn the national standards of maturity model, and then the CMMM model is implemented in the whole group.

In addition to internal appraiser preparation, it is more critical to determine the respondent of each competency subarea, because most people do not understand what the standard terms involve, which is also a difficult step.

It is said that the CMMM assessment is an open-book examination, the national standards are also open, and the evaluation is these terms, but the key depends on whether the enterprise personnel can cite evidence to prove that these terms are realized. Enterprises often feel that we should achieve this clause, but how can we not find cases of proof, this situation is mostly not to find the right person, failed to call the corresponding function of the people, call all.

However, for enterprises applying for Level 4, the division of internal business functions is sometimes very thin, and each person is only responsible for one of the functions, coupled with a wide range of assessment, assessment points are relatively detailed, and even the department head can not clearly answer all the questions. For example, “process design” ability of the subdomain of the process simulation proof, there are enterprises involved in machining simulation, welding simulation, assembly simulation, different process types, different professional simulation software, proof requires a number of professional simulation personnel, in order to respectively simulation process, simulation model, simulation objectives, simulation programs, simulation results clearly.

There has been a business consultation, can IT staff on behalf of all the capabilities of the sub-domain defense?

The answer is no. First of all, the ability of IT personnel to apply the system does not represent the actual ability of the business department to apply the system. Some software may not be fully imported, and the business department personnel cannot use it or do not want to use it. Secondly, in the evaluation, IT is found that IT personnel can usually answer the terms of level 3 and below clearly, but the terms of Level 4 involve a lot of data analysis models. It personnel know more about the software technology architecture and operation mode, and cannot explain the mechanism and application effect of the analysis model clearly. It can clearly explain the mechanism of the model, input and output data, constraint configuration, and how to operate the function.

According to statistics, for some enterprises with a finer division of functions, the four-level assessment involves 60 to 80 people, with an average of 1 to 2 people per clause.

Due to the large number of business personnel involved in CMMM Level 4, it is suggested that at the beginning, the division of labor should be based on the business functional departments corresponding to the competence sub-domains, and the department supervisors and technical backbone should be selected as the initial personnel, and the participating defense personnel should be assigned according to each clause in the preparation of evidence materials, combined with the provisions of the clause. When enterprises are psychologically prepared for personnel organization and have strong organization and coordination ability, the selection of the evaluation sponsor plays a more important role. Since the evaluation involves almost all business departments of the enterprise, senior executives are sometimes only responsible for part of the business lines, so it is very important to evaluate the selection of the sponsor. When necessary, the evaluation sponsor needs to be the head of the enterprise.

(3) Preparation of evidentiary materials

The preparation of evidentiary materials is an important work of evaluation preparation, and the four level evaluation focuses on the following aspects.

Valid evidence

The mention of “evidence” material preparation involves the question, what is really effective evidence? Some enterprises that have a preliminary understanding of maturity assessment may think that it is to prepare a defense PPT for each capability subdomain. In fact, there is no mandatory requirement to make PPT for CMMM evaluation, but the PPT form is convenient to introduce the effectiveness of enterprise intelligent implementation, each page has a distinct theme corresponding to each standard clause, lists what evidence is presented in each clause, and describes the key functions, characteristics, data and other points clearly, which is also convenient for the evaluation team to record.

Of course, companies can also use Excel spreadsheets for simplicity. This can create the misconception that having a powerpoint presentation ready means that you have prepared your defense material. Especially in the pre-evaluation stage, the evaluation team does not have more time to verify each system function, and the enterprise will mistakenly think that the formal evaluation and collection of “evidence” is also a report PPT, and the PPT can be passed.

The evidence required by the CMMM assessment is to prove the application results of the actual promotion of intelligent construction by enterprises

The evidence required by the CMMM assessment is to prove the application results of the actual promotion of intelligent construction by enterprises, including some documented information, such as intelligent manufacturing planning, heterogeneous system integration specifications, unified data coding specifications, etc., and also includes the operation demonstration of information system functions, especially the application of data analysis models in information systems required by level 4. Document information needs to provide the original document material, not the screenshot of the file; Similarly, a system demonstration is a login to the formal environment of the system, demonstrating functional operations and functional history data.

From this point of view, PPT role is there, but not so important, especially in the formal evaluation stage, PPT is only equivalent to a list of evidence, outline, is a tool for information display, is to let the evaluation team first understand, and then to show these documents, demonstrate these functions, list the key content, to prevent the display of omission or unclear expression. Some enterprises even if the PPT is very good, but the real system function has not been implemented, which can not be used as effective evidence for evaluation.

It is recommended that enterprises copy the documented evidence materials to the network sharing server when preparing for assessment.

At the same time, the information system that needs to be demonstrated, install the client in advance, save the access address in the favorites, prepare the demonstration login account (the account access function is different), and prepare the demonstration record data item ID; For example, if you want to show which purchase order is pushed down by the enterprise to the supplier management system, prepare the purchase order number. It is recommended to prepare at least two cases for each clause, one is to ensure the adequacy of the assessment verification, and the other is not to prepare hastily when one of the cases is not valid.

Business background introduction

Some enterprises realize that the proof is more critical, and begin to introduce the evidence of each clause in the defense, but the lack of the overall business background introduction will also affect the defense effect, which may lead to the assessment teacher can not understand the case presented, do not understand the use of such intelligent solutions, data analysis model and the role and significance, and even difficult to judge the correctness of the realization of system functions.

For example, whether a company’s production model is based on finished product inventory, on customer orders, or both, this will affect the logic of the automatic generation of production plans, the judgment of input data, and may also affect the evaluation of material requirements planning, sales and shipments. For example, whether the production line is 24-hour continuous production or only day shift production will affect the production scheduling mode and equipment operation and maintenance management.

It is suggested that when evaluating the first meeting and the overall intelligent introduction of the enterprise, the main product series, key processes, production line layout, production mode, and core advantages of the enterprise should be explained clearly.

In each competency subarea, introduce the business background involved in the department or business, including but not limited to:

Overview of business management, such as [equipment] what are the key processes of the enterprise, [energy management] enterprise energy consumption types and high energy consumption equipment standards.

Complete business processes, such as [Planning and scheduling] Complete process of planning (from master plan to planning at all levels), [procurement] complete process of material procurement (from material requirement calculation to material storage);

Organization of division of labor, such as [product design] product design is divided into which majors, or which functional modules for task division;

Key points and difficulties of business control, such as [warehousing and distribution] material pulling and distribution methods, [production operations] key process control indicators on the production site, etc.

The business background presentation will have a sense of inclusion in the subsequent evidence, but the introduction content should not be too much, as far as possible around the business background involved in the terms, after all, the evaluation of the defense time is limited.

Description of evidence cases

The system function requirements involved in the four-level clauses are relatively complex, and the intelligent application realization of each enterprise is not the same, and many data analysis optimization models have strong business logic, therefore, enterprises should grasp the key points and clear logic when presenting evidence.

Many enterprises do not know how to introduce the data analysis model in the system, or simply introduce the functional operation of the system “click the button and automatically generate”. The specific production links to which the analysis model is applied and the targets optimized are not introduced.

It is recommended to introduce the model application from the following aspects:

1. What problems are solved by the model application and what are the expected effects?
2. How to discover and determine the data indicators related to the model;
3. What are the input data of the model, which systems and devices do the data come from, and the frequency of data update, real-time or historical?
4. The operation logic or formula of the model, as well as reference variables, constraints and discriminant boundary conditions that need to be configured in advance;
5. Whether the execution logic of the model is triggered manually, or triggered automatically on a regular basis, or depends on what kind of logic to trigger;
6. What is the output result of the model and in what form, such as single or continuous, data list or trend curve;
7. How to implement the above functions in the information system.

Pay special attention to the last point, many enterprises understand the requirements of the terms, the business principles are clear, and the technical solutions are feasible, but the evaluation requires the results that have been achieved, and the final implementation must be implemented in the software system, rather than the scheme on paper.

Effectiveness of construction results

Compared with other levels, some intelligent projects in level 4 are based on research, pioneering and tentative, and the results may not necessarily play a role, but also need to be iterated repeatedly to verify, and it is necessary for enterprises to introduce the application effects that have been achieved.

There was once a company to implement predictive maintenance of equipment failure, according to the prediction curve, it was judged that the equipment would fail after half a year, that is, the equipment would work normally for a long period of time; If according to this logic, the frequency of equipment failure is not high, and the impact of failure is not large, then where is the investment and significance of the enterprise to build a predictive model? However, by actually checking the historical fault records of this type of equipment in the equipment management system, it was found that the monthly equipment failure rate was indeed high, so the influencing factors considered by the judgment model were missing, the discrimination accuracy was not enough, and the accuracy of the prediction needed to be improved.

Of course, enterprises are still encouraged to make such attempts, and the model also needs an iterative validation process, which may not be successfully applied or the effect is not good at present, and the evaluation team will weigh in on the score.

In summary, enterprises need to pay a certain amount of energy to prepare evidence materials, from another point of view, this is also a good stage of the enterprise summary, many enterprises through the evaluation of preparation, suddenly found that they have so many advantages, enhance the confidence of enterprise intelligent construction, which is also another meaning of CMMM evaluation.

（5) Rectification after pre-evaluation

After the pre-evaluation, the company will have a period of time to rectify and prepare for the formal evaluation. Some enterprises think that the pre-evaluation process is relatively smooth, and the formal evaluation will not cause big problems, but in fact, from the meticulous and rigorous degree of evaluation and verification, the formal evaluation needs to prepare more things.

In the pre-assessment stage, the enterprise generally understands whether the assessment team accepts the cases described in the PPT as valid evidence, and what the possible weaknesses are. In view of these problems, it is not a simple modification of the PPT. The formal evaluation is not a verification of the PPT, but the corresponding implementation results behind the PPT presentation.

Therefore, it is recommended to demonstrate the intelligent application mentioned in the PPT and simulate the operation.

If some information system functions need to be improved, they also need to be developed, and if the enterprise does not have the ability to develop, it is necessary to get the technical support of the solution supplier as soon as possible, because the rectification time is limited, it is necessary to clarify the priority of functional improvement. Some screenshots of documents and materials mentioned in the PPT are temporarily extracted from other materials, and the original documents should be found for formal evaluation and inspection.

Finally, enterprises passing the level 4 assessment is not the end, and enterprises that have passed the level 4 also have weaknesses that need continuous improvement. Through on-site evaluation, we also found many excellent CMMM service providers to provide enterprises with intelligent manufacturing consulting and software, equipment, system solutions. In the next step, the secretariat of the CMMM Evaluation Working Group will give full play to the advantages of evaluation and refine a batch of CMMM service provider resources with successful practical experience for enterprises to accurately select and adapt.

CMMM evaluation is one of the mirrors of enterprise intelligent construction, the mirror only reflects a limited side, the enterprise has more goals to tackle. It is hoped that enterprises can make good use of CMMM, innovate continuously and be down-to-earth in the process of intelligent manufacturing construction, and hope that more enterprises can pass the CMMM level 4 evaluation.

The core idea of four-level capacity building

What direction can be guided by the four-level capacity building of enterprises? It is summarized here as 16 words “reduce complexity, dynamic balance, integrate resources, and resist risks”.

(1) Reduce complexity

With the development of enterprises, the “complexity” of manufacturing will be increased from all dimensions, and the problems faced by fourth-level enterprises are more prominent. If the complexity is not effectively solved, it will reduce the operation efficiency of enterprises. Therefore, reducing complexity is the application of intelligence, so that our manufacturing process is simpler, smoother and more agile, whether it is product development, production scheduling, production control and so on. It is necessary to summarize knowledge and experience according to complex scenarios to assist business activities in rapid processing.

For example, in the face of the complexity of serialized product design, [Product Design 4a] adopts modular and parametric design to accelerate product development; [Process design 4b] faces the process similarity of typical characteristics and uses process knowledge to provide decision-making for process design and planning; [production operation 4a] faces the mixed production of multi-variety products. Realize the automatic delivery of production instructions according to the matching of products and programs. [Planning and Scheduling 4a] realizes the automatic optimization and scheduling of work plans in the face of the complexity of multi-tasks, multi-resources and multi-constraints; In the face of the complexity of point-to-point integration between information systems in [Integration 4a], technical routes such as ESB and ODS are adopted to make the integration operation and maintenance work more agile.

(2) Dynamic balance

The production process of the enterprise has been affected by changes in the external environment and internal resources, and it is impossible to always produce products in the normal environment such as the established route, process and mode. The abnormal indicators appear in the “uncertain” time, so the entire production process is on the one hand to maintain normal as much as possible, on the one hand, and recover in the abnormal as soon as possible, so that the cross. Intelligence is to solve this dynamic balance ability.

Sometimes these anomalies have clear indicators, explicit anomalies, such as poor product quality indicators, and some invisible anomalies, such as deviations from the trend performance of process indicators. No matter what kind of anomaly, it is necessary to establish an indicator reference system to monitor which indicators and what range of indicators are normal, outcome, trend, or correlation. In the national standard three clauses is to establish these indicators and monitor, such as [equipment management 3b] equipment OEE indicators, [energy management 3a] high energy consumption equipment energy consumption evaluation indicators, [planning and scheduling 3c] production schedule indicators, the rationality of the indicator formulation is itself a process of data analysis.

Level 4 dynamic balance is to give feedback to the abnormal results of these indicators, eliminate the complexity brought by these “uncertainties”, carry out agile response, and achieve short-cycle balance;

After the manufacturing system runs for a certain period of time, the index value is constantly verified, the index type and its boundary need to be readjusted and adjusted, optimized to make it more reasonable, and a long-cycle balanced closed-loop is carried out.

Short-cycle dynamic balance, such as [Equipment Management 4b] to analyze the equipment with abnormal OEE, determine the factors affecting the operation efficiency (unit feed is not timely, the process rhythm is not balanced before and after the process, the equipment process parameters are unreasonable), and automatically adjust the process and plan. [Planning and Scheduling 4b] In the face of material shortage, abnormal quality, equipment downtime and other abnormal situations, automatic production scheduling adjustment, so that it can return to normal production. Long-term dynamic balance, such as [Procurement 4c] after the accumulation of suppliers’ historical supply performance, optimize the supplier evaluation model to make it more reflective of suppliers’ supply performance; [Logistics 4c] Optimize resource loading capacity and transport path by combining transport capacity resource capacity, transport nodes and transport tasks, reduce transport costs and improve logistics efficiency.

(3) Integrate resources

The scope of “resources” here is wide, not only the materials, equipment, tooling, energy, software, data, etc., but also the organization, posts, and business activities within the enterprise. [Organizational Strategy 3b] refers to the optimization and adjustment of job structure, which is to combine the intelligent construction of enterprises and carry out the functional adjustment of corresponding posts and the integration of job structure.

On the one hand, the platform integrates data around the whole life cycle of products, such as [Data 4a] enterprise-level data center, [product design 4b] to achieve comprehensive data archiving including design model, [product service 4c] to integrate data of product use and operation process, and optimize product operation performance. On the other hand, through intelligent application, business activities of various functional posts can be coordinated, information can be transmitted, and internal resources of the organization can be integrated. For example, [Product design 4d] integrates the BOM and other resources needed for product development and production, logistics, sales and service business collaboration, and [Sales 4b] integrates online and offline sales channels and demand resources.

In addition, compared with the data integration between software systems and hardware devices that is focused on at level 3 (integration level), more emphasis is placed on real-time “soft and hard” seamless integration at level 4.

For example, [Warehouse and Distribution 4a] realizes the automatic distribution of materials from storage area to production area based on production scheduling instructions. This process involves the linkage and interaction between software systems such as production management system, scheduling system, storage system and distribution system and hardware systems such as digital storage equipment, digital distribution equipment and production unit equipment. Each action of interaction includes message transmission such as giving instructions, receiving responses, driving execution, and completing execution. Frequent interaction between software and hardware also closely combines production organization and equipment operation. Similarly, [Planning and Scheduling 4b] requires an integrated soft and hard mode to implement real-time optimal scheduling of production anomalies.

Note: The expression form of “Planning and scheduling 4b” appears in the article, representing the national standard GB/T 39116-2020 “Intelligent Manufacturing capability Maturity Model”, “planning and scheduling” capability subdomain, level 4, clause b, that is, “all production factors should be monitored in real time, and the system can realize automatic decision-making and optimal scheduling of abnormal situations”.

(4) Fending off risks

In contrast to “dynamic balancing” to eliminate anomalies that have already occurred, resistance to risk is the predictive ability that gives manufacturing systems. Know the risks in advance, prevent potential problems, and minimize losses, that is, the four-level standard mentions a lot of “forecasting”.

In order to build a risk prediction model, it is first necessary to find indicators representing risks, which need to be based on the accumulation of daily experience and knowledge, that is, the “knowledge base” mentioned in the third level clause, and ensure that the “indicators” can be collected or measured in real time. Secondly, it is necessary to give the boundary range of indicator trend change and critical values at all levels to quantify the risk. Then, according to the measures taken by the risk, the solution countermeasure model is established to control the risk within a certain range. Finally, compared with the actual situation, the accuracy of the model for risk prediction is verified, and the model is continuously optimized.

Some indicators can directly predict risk, such as monitoring the trend change of quality characteristics; Some are associated with risks and indirectly predict risks, such as predicting the change of product quality through the trend change of process indicators, and some are also caused by multiple factors, such as [equipment management 4a] predictive maintenance of equipment failure, which may involve the multi-indicator operation performance of key equipment components, external environmental impact, product processing parameters, etc.

The four-level clause emphasizes the ability of enterprises to resist risks and establish early warning and forecasting models.

For example, [procurement 4b] procurement risk model, [production 4c] quality prediction model, [safety and environmental protection 4a] dynamic identification of safety risks, [safety and environmental protection 4b] environmental emissions risks, [storage and distribution 4c] tank farm tank risks, and [information security 4c] industrial information security risks to improve active defense capabilities. At the same time, in [product service 4a\4b], the risk prediction is extended to the operation and use of products, and the potential risks of product operation are perceived, such as establishing remote operation and maintenance monitoring and optimizing performance services for elevator and engine products, which increases the added value of products while resisting operation safety risks.

To sum up, the four-level capacity building of enterprises is a long-term process, and its intelligent construction results represent the ability of enterprises to deeply cultivate the industry, but also the advantage of other enterprises, and its application results will also ensure that enterprises are in the leading position of industry competition.

About the author

Yang Zhuofeng, Doctor of Engineering, CMMM® high maturity chief appraiser, training instructor; As one of the main drafters of relevant standards of “Intelligent Manufacturing Capability Maturity Model”, he has presided over the overall informatization planning and intelligent manufacturing solution design work of many enterprises in aerospace, engineering machinery, pressure vessels, automobiles, electronics and other industries, and has rich experience in business analysis, process optimization and enterprise integration. Proficient in enterprise information management system design, development and implementation.

本文由数字化转型网（www.szhzxw.cn）转载而成，来源于新工业网；编辑/翻译：数字化转型网宁檬树。

免责声明: 本网站(http://www.szhzxw.cn/)内容主要来自原创、合作媒体供稿和第三方投稿，凡在本网站出现的信息，均仅供参考。本网站将尽力确保所提供信息的准确性及可靠性，但不保证有关资料的准确性及可靠性，读者在使用前请进一步核实，并对任何自主决定的行为负责。本网站对有关资料所引致的错误、不确或遗漏，概不负任何法律责任。

本网站刊载的所有内容(包括但不仅限文字、图片、LOGO、音频、视频、软件、程序等) 版权归原作者所有。任何单位或个人认为本网站中的内容可能涉嫌侵犯其知识产权或存在不实内容时，请及时通知本站，予以删除。