EnglishRSSApp/Android客户端iPad客户端Kindle版手机版天猫旗舰店

三项原则助力硬件产品开发规模化敏捷 数字孪生实现车身工程和制造敏捷性

作者:Christoph Weber 2022-02-20 22:51:00 0

图片9

汽车制造商正在转型成为科技公司。软件开发人员利用敏捷方法每隔几周就能交付一次新产品的迭代,但硬件工程师仍然依赖瀑布式项目管理方法,开发周期需要2至4年。在我们过去的项目中,我们已经确定了实现硬件产品开发敏捷化的三项原则。本文将描述它们如何应用,分享汽车车身制造的实际案例,并解释数字孪生技术是如何发挥作用的。我们旨在激励来自任何传统行业的管理者,敢于向更高的灵活性和以客户为中心为目标迈出下一步。

汽车制造商面临着快速推出新款电动车型、满足精通技术消费者的期望以及赢得新工作时代人才争夺战的压力。尽管如此,车身工程等传统领域可能认为数字化和敏捷颠覆不适用于他们自己的领域,因为交付周期和对模具、冲压线和焊接机器人的大量投入似乎决定了前期规划的事无巨细。事实是,当今数字孪生技术已经准备好,即使是应对复杂的硬件产品,也可以实现全面的数字化和敏捷转型。

图片1

图1:汽车车身工程和制造必须变得更快、更便宜和以客户为中心

在过去三年与汽车制造商和供应商的合作项目中,我们已经确定了三个精益-敏捷原则,帮助传统硬件生产商能够一定程度变得敏捷:点对点负责,基于工作系统优化各项环节和跨职能团队。一个普通汽车制造商可以通过车身工程和制造的数字化和敏捷转型,将生产开始时间加快6个月,每年节省超过1000万美元,并激发更高的客户和员工参与度。

图片2

图2:实现硬件产品开发敏捷性的三个原则

1)点对点负责

在开发一个复杂的产品,如新款车型时,需要多个部门和供应商一同协作。然而,在瀑布式组织架构中,每个部门都有自己的时间节点和KPI。每个部门可能使用不同的工具来制定产品设计和生产概念。不同的工具--例如软件、电子表格、历史数据或个人经验--是基于对同个现实的不同模式和理解。

由于这种组织和技术上的脱节,工程师专注于他们的目标,可能没有充分考虑其他部门的限制和目标。缺乏相互接受的数据减缓了跨部门决策的速度。自发的信息反馈是不可能的。这些职能部门之间的交接导致了浪费和延误--比如过度的工程设计,整合问题的延迟发现,错过了成本节约和轻量化的潜力,因此导致了预算和时间进度的超支。

图片3

图3:职能之间的隔阂导致浪费和延误

敏捷的企业打破了职能上的隔阂,建立了跨职能的团队,对每个价值流承担点到点的责任。每个价值流提供一个模块,如车身或动力系统,之于整个解决方案,即汽车。每个价值流和支流都需要明确定义的接口,并应尽可能地相互独立。团队承担从设计到生产的点到点责任,以优化流程,实现系统思维,而不是局部优化。一款集成软件工具的数字化点到点平台应该支持每个团队。规模化敏捷框架(Scaled Agile Framework)为这种方法提供了进一步的指导。

图片4

图4:价值流组织优化了流程和系统思维

一家德国汽车制造商已经成功地将车身冲压和装配之间职能上的隔阂衔接起来。传统上,大多数汽车制造商对单件冲压和车身装配进行分别优化,这导致了额外的校正轮次。通过新方法,冲压和装配工程师共同承担点对点的责任,以优化整个车身系统。他们整合冲压和装配的数字工艺模型,以优化车身装配的单件形状。通过这种系统优化的方法,这家领先的汽车制造商将一个引擎盖装配的生产准备时间从12个月缩短至6个月。

图片5

图5:系统优化减少过度工程和浪费

2) 基于工作系统优化各项环节

传统的产品开发方法是在2至4年内通过数个瀑布式的项目规划阶段进行的,然后以 "大爆炸"的方式向市场推出一款新车型。因此,客户的反馈只能针对每款车型每隔几年被纳入考虑。此外,制造和集成问题只能在生产升级阶段发现。相比之下,敏捷产品开发每隔几周在常规的冲刺中测试和提交小的产品迭代。尽早并尽可能频繁地采纳客户的反馈,发现制造问题减少了产品和可行性失败的风险。

软件产品的迭代可以在几秒钟内通过点击鼠标按钮进行测试。然而,像汽车车身这样的硬件产品需要几个月的准备时间,并在模具、冲压线和焊接机器人方面进行大量投入。这就是数字孪生--现实生活中的物体或过程的虚拟模型--可以取代物理试制和测试。物理驱动的数字孪生技术帮助工程师能够在几个小时内测试产品性能和可制造性,即使是复杂的硬件产品在每一次产品迭代中也是如此。这确保了每一个产品迭代从一开始就有质量保证。对于设计或制造概念的任何变化,数字孪生技术提供关于对质量、成本和时间影响的即时反馈。内部和外部客户可以在此基础上提供有意义的反馈。

图片6

图6:物理驱动的数字孪生实现预测和自我修正

一家日本汽车制造商通过系统地探索数以百计的替代产品设计和生产过程,并通过数字孪生的快速迭代测试,实现了客户价值最大化。例如,他们通过数字工艺孪生分析了具有替代材料厚度和形状的铝制发动机罩部件的冲压过程。他们排除了不可行的设计,然后选择了一个重量最轻、成本最低的设计--每个零件的重量减少了7%,材料成本减少了1欧元。

3)跨职能团队

我们的客户调查显示,50%的冲压工艺工程师没有系统地参与到冲压模具调试。工程师和车间团队之间这种令人震惊的脱节,基本上切断了工程设计的 "数字孪生"和实际制造的"物理孪生"之间的联系。工程团队可能不了解生产的实际情况,缺乏建立一个准确的数字孪生模型所需的反馈。车间团队可能会收到设计不良的模具,并以试错的方式重新设计流程,而不是在现有的流程知识基础上进行改造。

解决方案是在一个点对点的平台上连接设计师、工程师和车间工人,在一个数字工艺孪生平台上协同工作。工程师必须对车间的质量目标负责,并基于数字工艺孪生测试所有的KPI。例如,工程师应通过进行工艺能力分析来确保稳定的生产。车间工人应将实际的制造数据输入系统,以使数字和物理孪生保持一致,达到准确的预测。在此基础上,物理驱动的工艺孪生计算并提供指导,以最系统的方式达到一个稳健的生产过程。

图片7

图7:数字工艺孪生在点对点平台上连接团队

一家中国模具供应商利用跨职能的团队合作,为一个具有挑战性的A级铝制引擎盖项目减少质量检测轮次。如今,中国大多数汽车和模具制造商需要超过10个质量检测轮次来交付具有高表面和尺寸质量的铝板--这导致了大约一年时间上的延迟和巨大的模具再加工费用。在这个咨询项目中,我们与每个部门和供应商密切跟进,建立了一个精确的数字工艺孪生模型,包括材料测试数据和工艺能力分析。因此,我们可以准确地预测和补偿尺寸回弹,并在第一次调试时提供良好的零件。

结论与实施

汽车和其他行业的硬件产品开发必须走向敏捷,这样才可以变得更快、更便宜、更以客户为中心。当今数字孪生技术实现了全面的数字化和敏捷性转型。以下三个原则助力硬件产品的开发,实现规模化敏捷。

1)点对点负责:围绕价值流建立团队,为完整的系统优化单一部件

2)基于工作系统优化各项环节:通过由物理驱动的数字孪生测试快速优化轮次实现客户价值最大化

3)跨职能团队:在点到点平台上连接工程和车间团队,在一个数字工艺孪生平台上协同工作。

数字化和敏捷转型是一项领导任务。在制定战略目标和审查业务结构和流程、软件和技术、人员和文化方面,需要C级的支持。

下图概述了最近与中国一家汽车制造商合作的转型项目的目标和关键措施。

图片8

图8:数字化和敏捷转换是一项领导任务

Christoph Weber 担任AutoForm中国区总经理 – 全球所有主要汽车制造商都选择AutoForm软件。他是经过认证的Scaled Agile Framework® 5项目顾问,帮助汽车制造商通过车身工程和制造的数字化转型加速生产的启动。

-----------------

图片9

Automakers are turning into tech companies. While software developers leverage agile methodologies to deliver new product increments every few weeks, hardware engineers still rely on waterfall project management with 2-4 years development cycles. In our past projects, we have identified the three principles that make hardware product development go agile at scale. This article describes their application, shares real-life cases from automotive car body manufacturing and explains how digital twin technology serves as enabler. We seek to inspire managers from any traditional industry to dare taking the next step towards higher agility and customer centricity.

Automakers are under pressure to quickly launch new electric models, meet tech-savvy consumer expectations, and win the war for talents in the new work age. Still, traditional disciplines like car body engineering may believe that digital and agile disruption would not apply to their own field, since lead times and heavy investment in tooling, press lines and welding robots seem to dictate detailed upfront planning. The truth is that today’s digital twin technology is ready to enable a full digital and agile transformation even for complex hardware products.

图片1

Figure 1: Car body engineering and manufacturing must become faster, cheaper and customer centric

In projects with automakers and suppliers over the past three years, we have identified three lean-agile principles that allow traditional hardware producers to become agile at scale: end-to-end responsibility, fast feedback loops based on working systems and cross-functional teams. An average automaker can speed up start of production by six months, save over USD 10 Mio/year, and inspire higher customer and employee engagement through a digital and agile transformation in car body engineering and manufacturing.

图片2

Figure 2: Three principles enable agility in hardware product development

1) End-to-end responsibility

Multiple departments and suppliers need to work together when developing a complex product like a new car model. However, each department has its own milestones and KPIs in a waterfall organization. Each department may use different tools to work out product design and production concepts. Different tools – such as software, spreadsheets, historic data or individual experience – are based on different models and understandings of the same reality.

As a result of this organizational and technical disconnect, engineers focus on their subset of targets and may not fully consider other departments’ constraints and goals. The lack of mutually accepted data slows down cross-departmental decision making. Automated feedback loops are impossible. The hand-offs between these functional silos lead to waste and delays – such as over-engineering, late discovery of integration problems, missing cost saving and light-weighting potential, and therefore budget and time schedule overruns.

图片3

Figure 3: Functional silos’ hand-offs lead to waste and delays

Agile enterprises break functional silos and establish cross-functional teams with end-to-end-responsibility for each value stream. Each value stream delivers one module, e.g. the car body or powertrain, of the entire solution, i.e. the vehicle. Each value stream and sub-stream requires clearly defined interfaces and should be as independent from one another as possible. Teams take end-to-end responsibility from design to production in order to optimize flow and enable system thinking instead of sub-optimization. A digital end-to-end platform with integrated software tools should support each team. The Scaled Agile Framework provides further guidance on this approach.

图片4

Figure 4: Value stream organization optimizes flow and system thinking

German automaker has successfully connected the functional silos of car body stamping and assembly. Traditionally, most carmakers sub-optimize single part stamping and car body assembly separately, which leads to additional correction loops. In the new approach, stamping and assembly engineers share end-to-end responsibility to optimize the complete car body system. They integrate the digital process models of stamping and assembly in order to optimize single part shapes for the car body assembly. As a result of this system optimization, the pioneering automaker cut production ramp-up time from twelve to six months for a hood assembly.

图片5

Figure 5: System optimization cuts over-engineering and waste

2) Fast feedback loops based on working systems

The traditional product development approach is to proceed through several waterfall project planning phases over 2-4 years before launching a new car model with a “big bang” to market. As a result, customer feedback can only be considered once per model every few years. Furthermore, manufacturing and integration issues may only be discovered in the production ramp-up phase. In contrast, agile product development tests and delivers small product increments in regular sprints every few weeks. Early and frequent incorporation of customer feedback and surfacing of manufacturing issues reduces the risk of product and feasibility failure.

Software product increments can be tested within seconds by the click of a mouse button. However, hardware products like a car body require several months lead time and heavy investment in tooling, press lines and welding robots. This is where digital twins – virtual models of real-life objects or processes – can step in to replace physical prototypes and testing. Physics-driven digital twin technology enables engineers to test product performance and manufacturability even for complex hardware products in every product iteration within a matter of hours. This ensures build-in quality from the beginning in every product increment. For any change in design or manufacturing concept, the digital twin provides immediate feedback on the impact on quality, cost and time. Internal and external customers can provide meaningful feedback on this base.

图片6

Figure 6: A physics-driven digital twin enables prediction and self-correction

Japanese automaker maximizes customer value by systematically exploring hundreds of alternative product designs and production processes tested in fast iterations by digital twins. For instance, they analyzed the stamping process of an aluminum hood component with alternative material thicknesses and shapes by a digital process twin. They excluded unfeasible designs and then selected the one with lightest weight and lowest cost – reducing weight by 7% and material cost by EUR 1 per part.

3) Cross-functional teams

Our customer survey revealed that 50% of stamping process engineers do not systematically attend the stamping tool tryout. This shocking disconnect between engineering and shop floor teams essentially cuts the connection between the engineered “digital twin” and the actually manufactured “physical twin”. The Engineering team may not be aware of production realities and lack the feedback needed to build an accurate digital twin model. The shop floor team may receive poorly designed tools and re-engineer the process in a trial-and-error approach instead of building on the existing process knowledge.

The solution is to connect designers, engineers and shop floor workers on an end-to-end platform to work collaboratively on one digital process twin. Engineers must be held accountable for quality targets on the shop floor and test all KPIs based on digital process twins. For instance, engineers should ensure a stable production by conducting a process capability analysis. Shop floor workers should enter actual manufacturing data into the system, in order to align the digital and physical twins and reach accurate predictions. Based on this, the physics-driven process twin calculates and provides guidance on how to reach a robust production process in the most systematic fashion.

图片7

Figure 7: A digital process twin connects teams on end-to-end platform

Chinese tool supplier leveraged cross-functional teamwork to cut quality loops for a challenging A class aluminum hood project. Today, most auto and toolmakers in China take over ten quality loops to deliver aluminum panels with high surface and dimension quality – resulting in circa one-year time delay and significant cost for tool re-milling. In this consulting project, we followed up closely with every department and supplier to model and manufacture an accurate digital process twin, including material test data and process capability analysis. As a result, we could accurately predict and compensate dimensional springback and deliver good parts in the first tryout.

Conclusion & Implementation

Hardware product development in automotive and other industries must go agile in order to become faster, cheaper and more customer centric. Today’s digital twin technology enables a full digital and agile transformation. The following three principles make hardware product development agile at scale:

1) End-to-end responsibility: Establish teams around value streams to optimize single parts for the complete system

2) Fast feedback loops based on working systems: Maximize customer value through fast optimization loops tested by physics-driven digital twins

3) Cross-functional teams: Connect engineering and shop floor teams on end-to-end platform to work collaboratively on one digital process twin

Digital and agile transformation is a leadership task. C-level support is required to set strategic goals and review business structures and processes, software and technology, people and culture.

Below overview shows the goals and key measures from a recent transformation project with an automaker in China.

图片8

Figure 8: Digital and agile transformation is a leadership task

Christoph Weber  is General Manager of AutoForm in China – AutoForm’s engineering software is used by all major carmakers worldwide. He is certified Scaled Agile Framework® 5 Program Consultant and helps automakers speed up start of production through digital transformation in car body engineering and manufacturing.

相关阅读

已有0人发表了评论

哈佛网友评论