一、核心摘要
（一）数字化理解
研发数字化本质：研发数字化是一场革命性的研发流程变革。
（二）研发项目综述
1、整车产品研发类别：包含小改款、年度款、大改款、升级换代和全新构架项目，是以满足用户需求为根本目的的针对性产品迭代。
2、研发方式理解：逆向研发是学习和积累的必要手段，正向研发是学成之后产出的结果。
（三）研发数字化的技术应用
1、协同研发平台：协同研发平台是研发内部与外部的协同，是敏捷开发机制的共建。
2、虚拟现实：可减少一次性开发成本，缩短项目周期，以虚拟的方式在现实中获得最优模型，打破物理空间和时间的限制。
3、数字孪生：利用强大的复现能力大幅度减少物理样机的试验次数，保证产品设计的可追溯性、系统性和经济性，收敛潜在问题，聚焦软硬在环。
4、云上数据反哺：利用后端数据分析反哺研发是大数据造就的核心价值。
（四）研发数字化落地举措
数字化前期的注意事项：
1.数字化转型自上而下的过程；
2. 数据孤岛需未雨绸缪；
3.人才是数字化中的珍稀血液；
4.数字化是对于流程的再造和变革；
5.数字化部门的确立是保证全面协同的基石。
二、研发数字化的理解
（一）何为研发数字化？
利用数字化技术在研发周期缩短、平台化和虚拟验证能力的基础上利用数据流动实现研发流程的变革大多数观点认为以虚拟化和数字化的形式代替或辅助传统汽车研发的业务环节，实现时间、成本的节约和质量的提升为数字化的核心价值。然而，艾瑞认为，在消费者需求快速变化和柔性化生产的背景下，降本增效仅可作为研发数字化的表层价值看待，而其深层价值是利用数字化工具缩短整车开发周期、实现平台的复用和建立虚拟验证能力，其核心在于三维数模和超级BOM。三维数模达成可延展、可控制、可追溯、可复现的分析；超级BOM可实现平台一体化、产品数据管理，在柔性化生产基础上实现适合小批量、个性化生产的组件集合。部分车企可将研发周期从36个月降低至18个月甚至更短，开发后期的设计修改减少50%，原型车制造和试验成本减少50%，投资收益提高50%。而究其根本，研发数字化的核心价值体现在研发周期缩短、平台复用和软硬件一体化虚拟验证能力的基础上，利用数据缩短决策链，能够围绕用户迅速给与支持和响应，也有能力按照消费者要求的时间、方式、配置、价格提供消费者期望的车型，是一场革命性的研发流程变革。
（二）新冠疫情对于研发数字化进程的驱动
以数字化转型缓解黑天鹅事件带来的潜在风险。
1、远程办公：各类远程办公、线上会议软件在疫情间发挥了巨大的作用。然而，一方面汽车研发需要大型软件和高算力进行虚拟仿真，个人电脑和家庭网络则难以带动；另一方面由于研发工作和数据的高度保密性，使得部分企业员工只能通过公司内网才能登陆办公账号提取关键数据，因此远程办公在汽车研发环节的价值难以真正体现。
2、车企与供应商的协同：汽车产业的横向跨度较大，与供应商间协同的重要性不言而喻。然而疫情下供应商同样难以将重要的非标准化、定制化数据及时与主机厂同步，导致部分关键决策的停滞。
3、物流与供应链：在车企全球化布局的背景下，疫情难以保证零部件和其他实体物料的及时供应，导致部分物理实验被迫停滞，车企被迫应用更多的虚拟验证手段来应对线下供应链的断裂。
三、整车研发项目综述
（一）整车产品研发类别以满足用户需求为根本目的的针对性产品迭代。
不同企业对于研发项目定义和类别有所差别，但均依据不同目的进行针对性改动。全新构架项目一般为迎合行业新趋势，打造差异化产品布局的战略目标而建立；需改动车辆钣金及下车体的项目通常因为造型过时、舒适度及体验难以满足大部分用户需求；年度改款通常针对市场同类产品竞争和用户需求进行针对性改动；而小改款则根据用户痛点快速迭代。究其根本，均为满足日益变化的消费者需求。其中，优良的车灯模具为千万级别，前保模具为百万级别，因此车辆的更新换代涉及大量资金投入同时面临着激烈的市场竞争和消费者接受度风险，导致较大的项目决策极为谨慎。目前，现有车型更新换代周期约为16个月，从研发到销售的周期甚至更短，对于提升研发效率的需求和难度也日益增加，真正可以做到完全自主的全新构架产品考验着我国汽车工业的硬实力。
（二）整车产品逆向研发流程逆向研发是快速追赶世界水平的必然手段。
逆向研发源于逆向思维，将原有产品进行拆解、测量、分析得出数据，在原有的结构、造型上进行再次开发从而得到全新产品，降低研发成本和风险。我国汽车工业起步时期经历了从零到一的艰难过程，更新换代缓慢，车辆性能、造型、耐久等指标难以与大众、通用等巨头媲美，因此逆向研发成为了短时间内快速提升我国研发水平的重要手段。然而，在中国汽车工业萌芽时期，部分项目的逆向研发可悲哀的等同于逆向抄袭，并在我国市场并未进入完全竞争的条件下获得了喜人的成绩。在不断的技术积累下，逆向抄袭的现象逐渐减少，而逆向研发正在逐渐转变为系统性分析，可靠的研发理念和思考问题的方式，并在吸收先进经验的基础上进行创新和改造。
（三）整车产品正向研发流程结构化、标准化的指导产品开发过程的文件体系。
整车开发项目是多模块、多系统、多目标之间的平衡，而一辆优质的汽车也是外部和内部各项因素和指标取舍、平衡的产物。通用汽车的GVDP和福特汽车的GPDS等均蕴含的丰富的经验和科学的设计，为各大车企所广泛沿用。目前正向开发周期在36~48个月不等，并随着经验的积累和技术的进步逐渐缩短。
虽然不同企业的开发流程会按照企业自身情况进行不同程度的裁剪，阀点设置、周期和具体工作等或有所不同，但都与门径管理思想和集成开发体系一致，基于时间轴和里程碑进行明确的划分，底盘系统、电气系统等跨部门、跨系统的并行开发保障着节点交付物的质量和项目管理的周密性。
（四）不同研发方式的理解与选择逆向研发是学习和积累的必要手段，正向研发是学成之后产出的结果。
由于国外技术的封锁和我国技术的落后，我国对汽车结构设计长期以来处于懵懂阶段。2008年左右我国汽车结构化能力提升，此后多家车企宣布了正向研发的计划。而特斯拉的出现对车企产生了降维打击，其底盘结构、一体化冲压方式、高压接插件的连接方式、集中式EEA等又给我国车企带来了巨大的进步空间。因此，我国汽车主要研发手段呈现了由逆向转为正向，又转变为逆向的起伏过程。
然而，逆向研发并非代表着技术落后；逆向研发是学习和积累的必要手段，正向研发是学成之后产出的结果。同时，研发方式与车型、制造工艺等息息相关。由于经济车型材料利用率高、设计技术相似度高、技术复杂度低，大部分正向研发从经济车型入手。而豪华车腰线和裙线较为复杂，部分设计存在冲压负角，制造难度大，材料利用率低，导致正向研发难度较大。因此，主要研发方式的选择基于了不同历史时期下行业环境、技术能力、市场环境和产品定义等因素。
三、研发数字化技术应用和价值观
（一）研发数字化的核心技术架构以数据流为核心资产的全生命周期数字孪生。
研发数字化的技术构架主体包含协同研发平台、虚拟现实、数字孪生和云上数据流，其本质是多技术融合的数字化生态，以打通物理世界和虚拟世界的壁垒。整体数字化研发架构可被理解为以云服务为研发环境，以满足协同设计要求的仿真软件为基础，以虚拟现实为展现形式，以数据为流动资产的全生命周期数字孪生。
（二）协同研发平台对于研发效率的提升实现上传下达和部门间实时沟通的数字底座设计建模。
基于AUTOCAD和CATIA等软件提供的接口，做深度定制化的二次开发，建立账号体系和平台，员工登录平台后便可使用对应软件进行三维建模、有限元分析、运动仿真分析等。账号系统与项目管理系统深度绑定，由管理者在PLM中发布设计、校核等任务。由于软件的深度二次开发，在做数据校对和整车数据匹配时可便捷地抽取轻量化数据，有利于整车层面所需的大量校对工作。
同时，由于CAD等软件与企业级IT系统的融合，可对图形设计者和设计版本进行追溯。无纸化评审：评审过程中时常出现相似评审间的冗余问题，即不同评审专家提出相同问题，整改时需付出额外的沟通时间，甚至做出大量的无效整改。而基于协同办公的评审将问题在系统中实时记录，进行冗余问题的分类后推送至各整改人。冗余部分可被推翻、覆盖或仲裁，实现利用少量人工跟踪整个项目，降低出错风险，减少无效工作。
（三）协同研发平台对于研发流程的优化实现业务间的横向拉通，打开决策者的信息牢笼。
在传统研发流程中，部分企业的研发部门与其他部门独立办公，信息相互隔离，导致在研发环节较少考虑到制造工艺、成本和用户抱怨，甚至不同研发部门间也存在着信息和物理隔离；同时中层决策者也埋没在大量的研发日报检查和汇报材料的准备中，导致其无力倾听制造分析、用户需求和外部环境的变化，一定程度上导致了企业内的重要决策者在信息化时代的信息牢笼，被动过滤掉了大量有价值的信息。而协同开发平台可以实现信息的透明化，数模的改动可反映至制造端并判断其产生的影响；售后产生的车辆问题清单也可帮助研发进行数模的修改和评审决策。因此，协同研发平台的重要意义在于优化研发内部流程和横向拉通制造、售后等其他版块，实现信息在核心业务流程中的透明化和实时性，打破流程僵化给部分决策者带来的信息牢笼。
（三）虚拟现实对于研发效率的提升减少设计所需时间，以虚拟的方式在现实中获得最优模型。
虚拟现实以各种数据为基础构建虚拟汽车模型或驾驶环境，取代对实际零件进行测试和组装的汽车开发过程。设计师可以尽可能快地修改和检查设计方案，并快速地识别和改进在真实世界原型车中难以验证的错误，达到缩短研发周期的目的。在造型评审时可将原本16个月的周期缩短50%以上。同时，虚拟验证可降低油泥模型的制作和反复实物试验所带来的巨大成本消耗，在低成本、短周期的环境下有效评估设计与产品性能间的适应性以便获得最优的模型参考。
（四）虚拟现实对于研发流程的优化优化造型评审，打破物理空间和时间的限制。
专业部门评审时通常会用到油泥模型判断设计策略与外观的一致性，最终确定造型方案。通常造型评审会做5-6次油泥模型，其制作耗时且昂贵，难以观测汽车的动态变化。而虚拟现实将评审流程简化，通过VR设备可清晰的展现效果图中未能清晰展现的部位，同时可提前判断姿态表现甚至内饰设计。虽然部分评审项目已全部将油泥模型更换为VR模型，但部分企业考虑到开模、制造工艺和可行性等，目前VR暂时难以完全替代油泥模型，因此部分项目在初审和终审阶段仍需使用油泥模型，而在中间的审核阶段使用VR模型。
（五）数字孪生对于研发效率的提升数字模型和复现保证产品设计的可追溯性、系统性和经济性。
数字孪生实现项目前期的虚拟化验证，具有无限次、可变参数、可加速的复现特性，可验证产品的适应性和系统性表现，实现基于需求、功能、逻辑、物理的全过程仿真验证，加强主机厂在软件定义汽车时代的核心能力，造就了数字孪生在汽车研发环节的重要价值。在造型阶段，数字孪生解决数据滞后和难以回溯的问题；在虚拟仿真阶段解决传统分布式仿真的系统性和协同性验证问题；在样机试验阶段解决物理实验的故障风险和经济性问题。总体而言，数字孪生在相同试验标准下可节省约20%左右的时间和40%-45%的研发成本，其多维度、多领域的虚拟验证方式令其成为研发数字化关键的技术之一。
（六）数字孪生对于研发流程的优化协同性优化收敛潜在问题，耦合性优化聚焦软硬在环。
协同性：传统而言，内饰、外饰和动力总成开发相互独立，在月度同步节点进行模型拼合和总布置的断面评估，由于沟通即时性较低导致问题易被时间放大。而数字孪生可将数字模型在后台进行相对地实时拼合，干涉后的报警保证设计和调整的实时性，在源头即可弱化问题放大的风险，同时可观测数模改变后对产品功能的影响，大幅节省沟通和测试时间成本。耦合性：日益增加的智能化功能需要软件带动硬件实现，在设计环节则需要软件和硬件统一评审，然而传统而言主机厂并不具备完整的软件开发能力。在数字孪生的加持下，主机厂可选择定期复盘，判断软件迭代开发对于整车功能和表现的影响，不仅节省时间及成本，更加增强了软件定义汽车背景下主机厂的话语权。
（七）数据类型与反哺路线利用后端数据分析反哺研发为大数据造就的核心价值。
数据采集的渠道和分析方法多种多样，但将后端数据反哺至研发环节，使研发能够更加贴近生产需求和客户需求为研发数字化最大的价值所在。用户使用车辆的数据及整车数据可上传至云平台，为提升研发水平提供重要参考。尤其对于新能源汽车而言，2016年发布的《新能源汽车生产企业及产品准入管理规定》中指出，新能源汽车生产企业应当建立新能源汽车产品运行安全状态监测平台，按照新能源汽车产品用户协议对已销售的全部新能源汽车产品运行安全状态进行监测。而生产端、车辆端和售后端产生的数据都可针对汽车产品的功能改进、工程改良和研发流程提供全面指导，存在着巨大的潜在价值。
案例一：生产质量问题反哺生产数据云上流通提升研发与其他环节间的协同性和敏捷性。
部分车企的生产和研发基地分散在全国各地，某一具体车型在生产环节发现问题时，要求研发人员快速修改数模并传递给各地的生产和研发中心使其可调用正确的数模，同时判断数模修改对于整车性能、供应链及财务的影响，实现远程、多程的数据同步和实时传输。在共平台生产的背景下，也能够更加敏捷地防止问题扩大，减少由于信息壁垒导致的开发质量问题，有效缩短研发周期。
案例二：生产合理性问题反哺提升研发环节在生产层面的耦合性、经济性和可行性。
研发人员通常从满足设计逻辑合理性角度出发进行数模设计，而生产人员则需考虑制造工艺、可装配性和维修性等；同时，目前的数模装配仅供检查干涉，但无法从装配角度入手判断装配可行性。因此，部分研发人员会设计出复杂的，需特殊装配工具或较长工时才能完成装配工作的零部件，如视觉冲击感强、棱角鲜明的组合，需较长杆件装配的螺接等。但部分组合方式难以保证间隙面差，繁琐的装配难以达成预定的生产节拍，导致后期需工艺、工位和装配流程上的调整，因此此类设计是生产人员希望极力避免的要素。通过云上数据流通实现的紧耦合研发方式，允许生产人员在设计初期介入，在初版数模发布后结合软件、产线、工位、夹具、装配方式等提前校核出可装配性和工艺复杂程度等，在问题初显时将其解决，减少后期数据校核和修改所需时间。
案例三：车身异常数据反哺异常数据的实时推送提升车企发现和解决问题的敏捷性。
车联网平台以小时或分钟为单位回传车辆数据，预判潜在问题。车辆数据如油耗、制动、油门等经过提前设定好的规则和分类回传至各地数据库，基于不同平台、车型、子系统进行交叉匹配和数据分析，以可视化的方式定期推送至责任工程师，基于问题的严重性（召回性问题或一般性问题）、规模化程度（一方面指问题出现的车辆数；另一方面指问题零部件的通用性，如特殊零部件、跨平台零部件或跨架构零部件）、零部件重要性（转向轴、底盘、刹车等）和问题的必然性（必然发生或偶然发生）判断风险等级和对应的回流措施。严重问题可通过召回将正确的零部件传递给消费者；轻微问题可在年度改款车型中进行覆盖。以上车身异常数据的反哺可帮助车企快速的发现问题并及时更正，避免问题的规模化，做到对于车辆的预测性维护，提升对于客户的安全保障。
案例四：用户驾驶数据反哺多维度把握目标用户群体，精准提升车辆操控性和舒适性。
车企针对不同用户群体，基于不同车型打造差异化的操控和舒适体验，以把握不同类型消费者对于操控和舒适的需求。而大部分企业通过前期的市场调研难以得到对于研发人员精准的、可直接参考的目标参数。通过用户驾驶数据监测，从CAN总线中提取精准的动态数据并将其细化至参数指标之后，研发人员结合前期调研和目标参数可精准地改善整车性能参数标定，提升目标客群所需的操控性和舒适性，达成研发和用户需求之间的强关联，如通过驾驶风格定义不足转向度、扭矩变化梯度和制动力曲线等。
一方面有助于下一代车型的性能迭代，另一方面可将用户偏好的标定数据进行打包、分类，推送给用户，使得用户可以自行选择符合自己驾驶习惯的功能参数包，以个性化的方式为客户提供定制化的驾乘体验。案例五：售后维保数据反哺售后数据弥补研发缺陷，提升个性化问题的发现和解决能力目前我国汽车售后问题反馈手段较为落后，仍以4S店和质保热线为主，部分企业的售后问题清单仍需通过从CRM中人工导出，导致售后问题发现缓慢，整改不及时。虽然也出现了以第三方直播或记者线下走访进行问题反馈的模式，但仍然治标不治本。而建立多渠道的以数据驱动的反哺系统，可以从APP、售后等渠道实时回传用户抱怨和维保信息，经过数据分析后按照时间周期识别出不同车型、品牌的趋势性问题，帮助工程师快速的根据问题迭代数模或相关软件，从而提升发现和解决用户个性化/规模化问题的能力。同时可帮助研发人员快速进行知识积累，弥补研发缺陷，实现售后维保数据对于研发环节的价值提升。
案例六：用户网络数据反哺验证市场调研准确性，帮助下一代车型的迭代更新。
在研发前期市场调研阶段产生的结果难以在市场中进行精准验证，除销量外的其他数据难以证明当时用户画像和需求定义的准确性。通过抓取和挖掘论坛、网络媒体、私域流量中的消费者评论，从大数据的角度指导具体品牌车型的表现和用户需求的变化趋势。一方面可以证明前期市场调研结果与目前主要客群间的偏差，以便优化市场调研的手段；另一方面分析结果可以帮助到小改款或年度车型的开发，使其更加贴近消费者需求。
（八）研发流程变革趋势总结研发、生产、用户需求协同的流程变革。
研发内部和外部均存在相对孤立的情况将被数字化进程逐渐打破。更多的在研发前期引入市场调研的PDCA流程将有助于车企基于上一次的经验教训，更精确的对品牌、车型、车身等进行再规划和再定位。通过数据反哺考虑上一代车型的设计缺陷、消费者需求等问题，可在设计开发阶段避免类似问题的产生，逐渐培养研发人员的生产思维和用户需求导向。重要的是，随着软件定义汽车的趋势，汽车产品的差异化逐渐从传统硬件转变为软件，更多的主机厂会选择在G4阀或SOP前将车辆对外发布，并基于样车执行新一轮的用户调研，在量产前给与了主机厂给更多的软件优化机会，以便快速响应个性化的、多变的用户需求。
如何把握数据反哺和安全隐私间的平衡？保证数据安全的前提下基于所需内容收集和处理数据是做到平衡的关键在汽车智能化时代的背景下汽车数据的收集和使用不可避免，但也出现了安全和隐私泄露的风险。欧洲ADAC汽车协会针对33个品牌的237款车型进行了安全测试，结果显示99%的车辆能够被黑客解锁开走，整个过程最短仅需18秒。随着《汽车数据安全管理若干规定（试行）》和《中华人民共和国个人信息保护法》等文件的出台，对于数据的合规使用提出了更严格的要求。因此车企在采集和利用数据时，在保证数据安全的前提下，需要基于用途最小化的原则进行数据收集，例如采集用于自动/辅助驾驶的感知数据时避免车内数据的收集；用于驾驶员疲劳监测的数据尽量将图片或视频转换为2D/3D点云图像等，如此才能做到合理地采集数据并实现数据反哺。
（九）技术是否是实现数字化的唯一途径？
如果将技术的应用当做目的，就偏离了问题本身我们毋庸置疑技术的重要性，但在传统观念和近年来技术发展路径的依赖下，更加需要被明确的概念是数字化转型并不是信息技术和工具的简单叠加，而需要兼顾技术、行业知识、组织、业务和流程的变革。报告开篇提出，研发数字化的核心价值在于研发流程的变革，利用数据的流动缩短研发环节上的决策链。然而，数据的流动并非是技术的专利，业务流程和组织构架的转变同样可以实现研发流程的变革，同时节约大量成本。但部分流程上的变革有悖于传统认知或牵涉多方利益，导致部分项目负责人以技术为由掩盖自身对于数字化认知的捉襟见肘。因此在技术发展日新月异的时代下，有能力兼顾技术应用和组织力量的企业才能在竞争中屹立不倒。
四、研发数字化落地举措
（一）数字化前期的关键因素和相关建议：凡事预则立，不预则废
1、顶层设计：将顶层设计和管理层的支持作为充分必要条件，自上而下、小步快跑地推进数字化进程，切勿单一模块、单一部门推进。需立足企业自身的资源禀赋，结合业务现状、组织现状、行业趋势、技术成熟度等进行全面构架来确认数字化转型的远景目标。
2、系统协作：庞大的车厂体系包含部门间、业务间甚至海内外的沟通协作，在系统建立前要充分考虑多主体间原有系统和新系统间的协同，和不同数据格式间的互认，避免出现数据孤岛。同时注意将系统进行盘整，从不盲目扩散系统逐渐转变为精简合并，后期在数据类型变更时快速增加系统的数据采集和分析功能，以数据流带动生产、销售等各个环节对于研发的数据反哺。
3、人才引进：车厂体系通常以硬件和部分电气电器件的设计制造较为擅长，缺乏数字化相关人才，导致现有人才难以和服务商对话甚至过渡依赖服务商，以至于重过程轻结果。同时部分中层由于事务繁杂在推行数字化时往往忽略掉了某些不该忽略掉的因素，难以看到数字化整体蓝图，也难以带给高层数字化战略思维，导致无法形成专业团队进行数字化的推进。
4、流程梳理：需认清数字化本质，引进数字化功能配置并非引进了技术本身，而是利用技术能力促进研发流程的变革和再梳理，业务流程在线下没有理清的状态下只能让数字化理念变得更加苍白，因此需要将僵化的业务流程灵活化才能发挥出数字化的最大潜力。
5、组织构架：数字化的进程结合了业务、IT等多部门，跨部门的长期协作需要更高级别的数字化部门进行支撑，特别要包含懂业务、懂技术、懂战略的复合型人才，既保证集团内部的全面协同，也保证数字化落地后的精准赋能。
（二）数字化转型的路径抉择以整体布局、试点先行的策略全面推进数字化进程。
数字化转型是企业至关生死的长远规划，在过程当中切勿单点规划逐个推进，否则后期系统间协同、利益平衡、流程机制等问题将随着转型维度的扩大而被无限放大，此类系统性问题的叠加将导致数字化周期延长甚至在日益激烈的竞争中失去竞争地位。因此，在数字化的源头要先做好数据治理和业务流程的梳理，不清晰的线下流程对于线上则毫无意义。在数据和流程基础夯实后，根据各企业的资源禀赋和发展战略等进行数字化顶层设计，构建内外部协同、目标兼容、纲举目张的整体构架。但其并不意味着通过宏伟的顶层设计解决一切问题，而要在能够尽快看到成果的部分以协同的方式进行试点建设，在顶层设计的全局思维下看到局部成果并触类旁通后进行全面推广，以实现企业的高质量发展。
（三）数字化服务商选型标准深刻的行业认知和成功案例是服务商选型的关键标准。
主机厂在云平台和数字孪生等技术上的应用仍处于探索前进的阶段，因此决策较为谨慎，会优先考虑其对于汽车行业的认知，和相关性高、可移植性高的成功案例。在云服务领域，多样的部署方案、安全性和连通性为主要的考虑因素。数字孪生领域对于现有软件的兼容、互认，和整车级别的虚拟化协同仿真能力成为选型主要标准。研发平台则更多的被国外厂商垄断，其包含了数学、物理等基础学科的应用，因此对底层能力要求较高。而主机厂对于虚拟现实的know how则相对较少，更加注重包含软硬件在内的整体能力。
（四）数字化服务商能力图谱专业软件以国外巨头为主，国产自主化脚步亟需加速。
大部分研发软件和数字孪生等优秀产品均来自于国际巨头，而在国产化、自主化趋势下主机厂同样希望得到全流程自主可控的研发环境，以便在愈加复杂的国际环境下维持核心竞争力，因此我国服务商若在底层技术、商业模式、知识沉淀上有所积累，或将结合数字经济的发展加快核心软件国产化的进程。
（五）车企与服务商合作时的主要矛盾与建议以双方协同共建的方式探索数字化的落地策略。
目前，部分数字化服务商对于汽车行业的研发业务流程/痛点认知不足，难以打动客户，加之定制化能力的相对缺乏和较为封闭的生态导致了车企与服务商间合作的两大矛盾。其一，主机厂业务流程与产品功能不匹配，导致车企需要被动进行流程变更或产品需要被动打散重组，耗费沟通成本，延长数字化落地周期。其二，不同服务商的产品难以打通，使得在不同业务环节上的云平台、数字孪生等被迫独立部署，难以发挥出最大效能。因此，服务商需增强对汽车行业的认知，提供兼容性强、定制化程度高的解决方案，以更加开放的心态看待自身与车企和其他服务商之间的合作。
重要的是，车企数字化转型离不开基础设施的云化，云化的第一步便是底层数据的治理。而部分主机厂由于人才缺乏，理念模糊而并不擅长进行数据的治理。因此，数字化供应商或可协助车企在数字化前期共同进行数据的治理，厘清数据和流程间的触点，共建数字化执行策略，同时在此过程中逐渐弥补自身不足，共同探索数字化的落地方案。
翻译：
I. Core abstract
(1) Digital understanding
Nature of R & D digitization: R & D digitization is a revolutionary R & D process change.
(II) Research and development project overview
- Vehicle product research and development categories: including minor changes, annual changes, major changes, upgrades and new architecture projects, targeted product iteration with the fundamental purpose of meeting user needs.
- Understanding of research and development methods: Reverse research and development is the necessary means of learning and accumulation, while forward research and development is the result of output after learning.
(3) Research and development of digital technology applications
- Collaborative R&D platform: Collaborative R&D platform is the collaboration between internal and external R&D, as well as the joint construction of agile development mechanism.
- Virtual reality: it can reduce one-time development costs, shorten project cycle, obtain the optimal model in reality in a virtual way, and break the restrictions of physical space and time.
- Digital twin: The use of strong recurrence ability to greatly reduce the number of physical prototype test, ensure the traceability, systematization and economy of product design, convergence of potential problems, focus on the soft and hard in the ring.
- Data feeding on the cloud: Using back-end data analysis to feed research and development is the core value of big data.
(4) Research and development digital implementation measures
Precautions in the early stage of digitization:
- The top-down process of digital transformation;
- Data islands need to be prepared for a rainy day;
- Talent is the rare blood in digitization;
- Digitalization is the reengineering and transformation of process;
- The establishment of a digital sector is the cornerstone of comprehensive collaboration.
Two, research and development of digital understanding
(A) What is the digitization of R&D?
Based on the shortening of R&D cycle, platformization and virtual verification ability, using digital technology to realize the transformation of R&D process by using data flow. Most views believe that the core value of digitalization is to replace or assist the traditional business links of automotive R&D in the form of virtualization and digitalization, so as to realize time, cost saving and quality improvement. However, iResearch believes that under the background of rapid changes in consumer demand and flexible production, cost reduction and efficiency increase can only be regarded as the surface value of R&D digitization, while its deep value is to use digital tools to shorten vehicle development cycle, realize platform reuse and establish virtual verification capability, whose core lies in 3D digital model and super BOM. Three dimensional digital model to achieve scalable, controllable, traceable, reproducible analysis; Super BOM can realize platform integration, product data management, and component collection suitable for small-batch and personalized production on the basis of flexible production. Some automakers can reduce the development cycle from 36 months to 18 months or less, reduce late-stage design changes by 50%, reduce prototype manufacturing and testing costs by 50%, and increase return on investment by 50%. On the basis of shortening R&D cycle, platform reuse and virtual verification capability of software and hardware integration, the core value of R&D digitization is embodied in shortening the decision chain by using data, which can quickly provide support and response around users, and provide models expected by consumers according to the time, method, configuration and price required by consumers. Is a revolutionary change in the R&D process.
(2) The driving force of COVID-19 on the digitization process of R&D
Digital transformation to mitigate the potential risks of black Swan events.
- Telecommuting: All kinds of telecommuting and online meeting software have played a huge role in the outbreak. On the one hand, however, car development requires large software and high computing power for virtual simulation, which is difficult for personal computers and home networks. On the other hand, due to the high confidentiality of research and development work and data, some employees can only log in the office account through the company Intranet to extract key data, so the value of telecommuting in automotive research and development is difficult to be truly reflected.
- Collaboration between automobile enterprises and suppliers: the horizontal span of the automobile industry is large, and the importance of collaboration with suppliers is self-evident. However, it is also difficult for suppliers to timely synchronize important non-standard and customized data with Oems under the epidemic, leading to the stagnation of some key decisions.
- Logistics and supply chain: In the context of the global layout of automobile enterprises, it is difficult to ensure the timely supply of parts and other physical materials due to the epidemic, which leads to the stagnation of some physical experiments and the application of more virtual verification methods to deal with the breakdown of offline supply chain.
Iii. Overview of vehicle research and development projects
(I) Targeted product iteration of vehicle product research and development category with the fundamental purpose of meeting user needs.
Different enterprises have different definitions and categories of R&D projects, but they are modified according to different purposes. New architecture projects are generally established to cater to the new trends of the industry and create differentiated product layout strategic goals; The projects that need to change the sheet metal and the lower body of the vehicle are usually difficult to meet the needs of most users because of the outdated shape, comfort level and experience. The annual revision is usually targeted at the competition of similar products in the market and the needs of users; Minor changes quickly iterate according to users’ pain points. Fundamentally, they are to meet the ever-changing consumer needs. Among them, the excellent headlight mold is at the level of ten million, and the front protection mold is at the level of one million. Therefore, the replacement of vehicles involves a large amount of capital investment and faces fierce market competition and consumer acceptance risk, leading to a large project decision is extremely cautious. At present, the replacement cycle of existing models is about 16 months, and the cycle from research and development to sales is even shorter. The need and difficulty to improve the research and development efficiency are also increasing. The new framework products that can truly be completely independent test the hard strength of our automobile industry.
(2) Reverse research and development process of vehicle products Reverse research and development is the inevitable means to catch up with the world level quickly.
Reverse research and development comes from reverse thinking. The original product is disassembled, measured and analyzed to obtain data, and the original structure and shape are re-developed to obtain new products and reduce research and development costs and risks. Chinese automobile industry has experienced a difficult process from zero to one in its initial period, with slow renewal and the performance, shape and durability of vehicles are difficult to compete with giants such as Volkswagen and General Motors. Therefore, reverse research and development has become an important means to rapidly improve China’s research and development level in a short period of time. However, in the embryonic stage of China’s automobile industry, reverse research and development of some projects can be sadly equivalent to reverse copying, and achieved good results under the condition that the Chinese market has not entered the perfect competition. With the continuous accumulation of technology, the phenomenon of reverse copying is gradually reduced, and reverse R&D is gradually changing into systematic analysis, reliable R&D concept and way of thinking, and innovation and transformation on the basis of absorbing advanced experience.
(3) The document system that guides the product development process of structured and standardized forward research and development process of vehicle products.
A vehicle development project is a balance between multiple modules, multiple systems and multiple objectives, and a high-quality vehicle is also the product of external and internal factors and indicators of choice and balance. GVDP of General Motors and GPDS of Ford Motor contain rich experience and scientific design, which are widely used by major car companies. At present, the forward development period ranges from 36 to 48 months, and is gradually shortened with the accumulation of experience and technological progress.
Although the development process of different enterprises will be tailored to different degrees according to their own conditions, and the valve setting, cycle and specific work may be different, they are all consistent with the door path management idea and integrated development system, and are clearly divided based on the time axis and milestone. Cross-departmental and cross-system parallel development of chassis system, electrical system, etc., guarantees the quality of node deliverables and thorough project management.
(4) Understanding and selection of different research and development methods Reverse research and development is a necessary means of learning and accumulation, while forward research and development is the result of output after learning.
Because of the foreign technology blockade and the backwardness of our technology, our country has been in the ignorant stage of automobile structure design for a long time. Since 2008, the structural ability of Chinese automobile has improved. Since then, many automobile enterprises have announced their plans to conduct positive research. However, the appearance of Tesla has produced a blow to the automobile enterprises to reduce the dimension, and its chassis structure, integrated stamping mode, high voltage connector connection mode, centralized EEA, etc., have brought great room for progress to Chinese automobile enterprises. Therefore, the main research and development method of our car presents the reverse to positive, and reverse to reverse ups and downs.
However, backward development does not mean backward technology; Backward research and development is the necessary means of learning and accumulation, while forward research and development is the result of output after learning. At the same time, the way of research and development is closely related to the model and manufacturing process. Due to the high material utilization rate, high design technology similarity and low technical complexity of economy models, most forward research and development starts from economy models. However, the waist line and skirt line of luxury cars are more complicated, some designs have negative stamping Angle, which is difficult to manufacture and low material utilization rate, resulting in greater difficulty in forward research and development. Therefore, the choice of major R&D methods is based on factors such as industry environment, technological capability, market environment and product definition in different historical periods.
- Develop digital technology applications and values
(I) Research and development of digital core technology architecture with data stream as the core assets of the full life cycle of digital twinning.
The main technical framework of R&D digitization includes collaborative R&D platform, virtual reality, digital twinning and data stream on the cloud. Its essence is the digital ecology of multi-technology integration to break through the barriers between the physical world and the virtual world. The overall digital R & D architecture can be understood as the digital twinning of the whole life cycle with cloud service as the R & D environment, simulation software that meets the requirements of collaborative design as the basis, virtual reality as the form of presentation, and data as the current assets.
(II) The collaborative R&D platform can improve the efficiency of R&D and realize the design and modeling of the digital base for uploading and transmitting and real-time communication between departments.
Based on the interface provided by AUTOCAD, CATIA and other software, deeply customized secondary development was carried out, and the account system and platform were established. After employees logged in to the platform, they could use corresponding software for three-dimensional modeling, finite element analysis, motion simulation analysis, etc. The account system is deeply bound to the project management system, and the manager releases the design, check and other tasks in the PLM. Due to the deep secondary development of software, it is convenient to extract lightweight data when doing data proofreading and vehicle data matching, which is conducive to a large number of proofreading work required at the vehicle level.
At the same time, due to the integration of CAD and other software with enterprise-class IT systems, graphic designers and design versions can be traced. Paperless review: The redundant problem between similar reviews often occurs in the review process, that is, different review experts put forward the same problem, the rectification needs to pay extra communication time, and even make a large number of ineffective rectification. The collaborative office based review records the problems in real time in the system, classifies the redundant problems and sends them to the rectificators. Redundancy can be overridden, overridden, or arbitrated, enabling the entire project to be tracked with a small number of hands, reducing the risk of errors and reducing ineffective work.
(3) The collaborative R&D platform can optimize the R&D process to realize horizontal pull through between businesses and open the information cage of decision makers.
In the traditional R & D process, the R & D department of some enterprises works independently from other departments, and information is isolated from each other. As a result, manufacturing process, cost and user complaints are less taken into account in the R & D process, and there is even information and physical isolation between different R & D departments. At the same time, middle-level decision-makers are also buried in the preparation of a large number of research and development daily inspection and report materials, resulting in their inability to listen to manufacturing analysis, user needs and changes in the external environment. To some extent, this leads to the information cage of important decision-makers in the enterprise in the information age, passive filtering out a large number of valuable information. The collaborative development platform can realize the transparency of information, and the change of digital model can be reflected to the manufacturing end and the impact can be judged. The list of vehicle problems generated after sale can also help research and development to make mathematical and modular modifications and review decisions. Therefore, the important significance of the collaborative R&D platform lies in optimizing the internal process of R&D and horizontally pulling through other sections such as manufacturing and after-sales, realizing the transparency and real-time performance of information in the core business process, and breaking the information cage brought by the rigidity of process to some decision makers.
(3) Virtual reality can improve R&D efficiency and reduce the time required for design, so that the optimal model can be obtained in reality in a virtual way.
Virtual reality builds a virtual car model or driving environment based on a variety of data, replacing the vehicle development process where real parts are tested and assembled. Designers can modify and review the design plan as quickly as possible, and quickly identify and improve errors that are difficult to verify in the real world prototype to achieve the purpose of shortening the development cycle. During the modeling review, the original 16-month cycle can be shortened by more than 50%. At the same time, virtual verification can reduce the huge cost of making sludge models and repeated physical tests, and effectively evaluate the adaptability between design and product performance in a low-cost, short-cycle environment in order to obtain the best model reference.
(4) Virtual reality for the optimization of the R & D process optimization modeling review, breaking the restrictions of physical space and time.
Professional departments usually use putty model to judge the consistency of design strategy and appearance, and finally determine the modeling scheme. Usually the modeling review will do 5-6 times of mud model, its production time and expensive, difficult to observe the dynamic changes of the car. Virtual reality simplifies the review process, allowing the VR device to clearly show the parts that are not clearly shown in the renderings, and to judge the posture and even the interior design in advance. Although all the sludge models have been replaced with VR models in some review projects, some enterprises consider mold opening, manufacturing process and feasibility, and currently it is difficult for VR to completely replace sludge models. Therefore, some projects still need to use sludge models in the preliminary and final review stages, and use VR models in the intermediate review stage.
(5) Digital twin for the improvement of research and development efficiency Digital model and reproduction ensure product design traceability, system and economy.
Digital twin realizes the virtualization verification at the early stage of the project, has the characteristics of infinite times, variable parameters and accelerated repetition, can verify the adaptability and systematic performance of the product, realizes the whole process simulation verification based on the demand, function, logic and physics, strengthens the core ability of the OEM in the software-defined automobile era, and creates the important value of digital twin in the automotive research and development. In the modeling stage, digital twin solves the problem of data lag and difficult to trace back. In the virtual simulation phase, the problem of systematic and cooperative verification of traditional distributed simulation is solved. The problems of failure risk and economy in physical experiment are solved in prototype test stage. In general, digital twin can save about 20% time and 40%-45% research and development cost under the same test standard. Its multi-dimensional and multi-field virtual verification method makes it one of the key technologies for research and development digitalization.
(VI) Digital twinning for R & D process optimization synergistic optimization convergence potential problems, coupling optimization focuses on soft and hard in the loop.
Synergy: Traditionally, interior, exterior and powertrain development are independent of each other, and model integration and section evaluation of general layout are carried out at monthly synchronous nodes. Due to the low immediacy of communication, problems are easily amplified by time. The digital twin can combine the digital model relatively in real time in the background, and the alarm after interference can ensure the real-time design and adjustment, which can weaken the risk of problem amplification at the source. At the same time, it can observe the impact of the change of the digital model on the product function, and greatly save the cost of communication and test time. Coupling: Increasingly intelligent functions require software to drive hardware implementation, and unified review of software and hardware is required in the design process. However, traditionally, OMs do not have complete software development capabilities. Under the support of digital twinning, Oios can choose to periodically review and judge the impact of software iterative development on vehicle functions and performance, which not only saves time and cost, but also enhances the voice of Oios under the background of software-defined vehicles.
(7) Data Types and feeding routes Using back-end data analysis to feed research and development is the core value of big data.
There are a variety of data acquisition channels and analysis methods, but the biggest value of digital R&D lies in feeding back-end data back to the research and development process, so that the research and development can be closer to the production needs and customer needs. Users’ vehicle data and vehicle data can be uploaded to the cloud platform, providing an important reference for improving the level of research and development. Especially for new energy vehicles, it is pointed out in the New energy vehicle Production Enterprises and Product Access Management Regulations released in 2016 that new energy vehicle production enterprises should establish a monitoring platform for the operation safety status of new energy vehicle products and monitor the operation safety status of all new energy vehicle products sold in accordance with the user agreement of new energy vehicle products. The data generated at the production end, vehicle end and after-sales end can provide comprehensive guidance for the functional improvement, engineering improvement and research and development process of automobile products, which has huge potential value.
Case 1: Production quality issues feed back into the flow of production data cloud to enhance synergy and agility between R&D and other links.
The production and research and development bases of some automobile enterprises are scattered all over the country. When a problem is found in the production process of a specific model, the R&D personnel are required to quickly modify the digital model and transfer it to the production and research and development centers around the country so that the correct digital model can be called. At the same time, the impact of the digital model modification on the vehicle performance, supply chain and finance can be judged to realize remote and multi-process data synchronization and real-time transmission. In the context of common platform production, it can also be more agile to prevent the expansion of problems, reduce development quality problems caused by information barriers, and effectively shorten the research and development cycle.
Production quality problem data feeds logic
Case 2: The issue of production rationality feeds back into the coupling, economy, and feasibility of R&D at the production level.
Research and development personnel usually from the point of view of satisfying the design logic rationality, while production personnel need to consider the manufacturing process, assemblability and maintainability. At the same time, the current digital-analog assembly can only be used to check interference, but it cannot judge the feasibility of assembly from the perspective of assembly. Therefore, some researchers will design complex parts that require special assembly tools or long working hours to complete the assembly work, such as the combination of strong visual impact, sharp edges and corners, and the screw joint that requires longer rod assembly. However, it is difficult for some combination methods to ensure the gap surface difference, and cumbersome assembly is difficult to reach the predetermined production beat, which leads to the adjustment of the process, station and assembly process in the later stage. Therefore, this kind of design is the element that production personnel hope to avoid. The tight coupling research and development method realized by the data flow on the cloud allows the production personnel to intervene in the early stage of the design, and check the assemblability and process complexity in advance by combining the software, production line, station, fixture and assembly mode after the release of the first version of the digital model. The problems can be solved when they become apparent, reducing the time required for later data checking and modification.
Case 3: Abnormal body data feed abnormal data real-time push to improve the agility of auto enterprises to find and solve problems.
Iot platforms send back vehicle data in hours or minutes to anticipate potential problems. Vehicle data, such as fuel consumption, braking and throttle, etc. are sent back to local databases after pre-set rules and classification. Cross-matching and data analysis are carried out based on different platforms, models and subsystems, and pushed to responsible engineers regularly in a visual way based on the severity of the problem (recall problem or general problem) and scale (on the one hand, the number of vehicles with problems; On the other hand, it refers to the universality of the problem parts, such as special parts, cross-platform parts or cross-architecture parts), the importance of parts (steering shaft, chassis, brake, etc.) and the inevitability of the problem (inevitable or accidental occurrence), judging the risk level and corresponding reflux measures. Serious problems can be recalled to pass the correct parts to consumers; Minor problems can be covered in the annual model change. The feeding of the above abnormal body data can help automobile enterprises quickly discover problems and correct them in time, avoid the scale of problems, achieve predictive maintenance of vehicles, and improve the safety guarantee for customers.
Case 4: User driving data feeding multi-dimensional grasp of target user groups, accurately improve vehicle handling and comfort.
Automobile enterprises create differentiated control and comfort experience based on different models for different user groups, so as to grasp the needs of different types of consumers for control and comfort. However, it is difficult for most enterprises to obtain accurate and direct reference target parameters for R&D personnel through preliminary market research. Through user driving data monitoring, accurate dynamic data CAN be extracted from CAN bus and refined into parameter indicators. By combining preliminary research and target parameters, R&D personnel can accurately improve vehicle performance parameter calibration, improve the controllability and comfort required by target customer groups, and achieve a strong correlation between R&D and user needs. For example, define understeer, torque gradient and braking force curve by driving style.
On the one hand, it is conducive to the performance iteration of the next generation model; on the other hand, it can package and classify the calibration data preferred by users and push it to users, so that users can choose the functional parameter package in line with their driving habits, and provide customized driving experience for customers in a personalized way. Case 5: after-sales maintenance data feeding after-sales data to make up research and development defects, improve the ability to find and solve personalized problems. At present, the feedback methods of automobile after-sales problems in our country are relatively backward, and 4S stores and quality assurance hotlines are still the main ones. Some enterprises still need to manually export the list of after-sales problems from CRM, resulting in slow discovery of after-sales problems and delayed rectification. Although there is also a pattern of third-party live broadcasting or offline visits by journalists to give feedback on problems, it still treats the symptoms rather than the root causes. The establishment of a multi-channel data-driven feedback system can send back user complaints and maintenance information from APP, after-sales and other channels in real time. After data analysis, trend problems of different models and brands can be identified according to the time cycle, helping engineers to quickly iterate the digital model or related software according to the problems, so as to improve the ability to find and solve user personalized/scale problems. At the same time, it can help R&D personnel to accumulate knowledge quickly, make up for R&D defects, and realize the value of after-sales maintenance data for R&D.
Case 6: User network data feeds back to verify the accuracy of market research and helps iteration update of next-generation models.
It is difficult to accurately verify the results produced in the market research stage in the early stage of research and development, and other data except sales volume can hardly prove the accuracy of the user portrait and demand definition at that time. By grabbing and mining consumer comments in forums, online media and private traffic, it guides the performance of specific brand models and the changing trend of user demand from the perspective of big data. On the one hand, it can prove the deviation between the previous market research results and the current main customer groups, so as to optimize the means of market research; On the other hand, the analysis results can help the development of minor changes or annual models to make them more close to consumer needs.
(VIII) The trend of R&D process reform Summarizes the process reform of R&D, production and user demand collaboration.
The relatively isolated situation inside and outside R&D will be gradually broken by the digital process. More introduction of PDCA process of market research in the early stage of research and development will help automobile enterprises to replan and reposition brands, models and bodies more accurately based on the last experience and lessons. By reflecting the data to consider the design defects and consumer demands of the previous generation of models, similar problems can be avoided in the design and development stage, and the production thinking and user demand orientation of R&D personnel can be gradually cultivated. Importantly, with the trend of software-defined automobiles, the differentiation of automobile products gradually changes from traditional hardware to software. More Oems will choose to release vehicles before G4 valve or SOP, and carry out a new round of user research based on sample vehicles, giving Oems more opportunities for software optimization before mass production, so as to quickly respond to personalized and changeable user needs.
How to strike a balance between data feeding and privacy? Under the premise of ensuring data security, the collection and processing of data based on the required content is the key to achieve a balance. In the background of the era of automobile intelligence, the collection and use of automobile data is inevitable, but it also appears the risk of security and privacy disclosure. The European automobile association ADAC conducted security tests on 237 models from 33 brands and found that 99 percent of the vehicles could be unlocked and driven away in as little as 18 seconds. With the promulgation of Several Regulations on the Safety Management of Automobile Data (Trial) and the Personal Information Protection Law of the People’s Republic of China and other documents, stricter requirements have been put forward for the compliant use of data. Therefore, when collecting and utilizing data, automobile enterprises need to conduct data collection based on the principle of minimizing usage on the premise of ensuring data safety. For example, the collection of in-car data should be avoided when collecting perception data for automatic/assisted driving. The data used for driver fatigue monitoring should be converted into 2D/3D point cloud images as far as possible, so as to reasonably collect data and realize data feeding.
(9) Is technology the only way to achieve digitalization?
If the application of technology is regarded as the purpose, it will deviate from the problem itself. We have no doubt about the importance of technology, but under the dependence of traditional concepts and the development path of technology in recent years, it needs to be more clearly defined that digital transformation is not the simple superposition of information technology and tools, but the transformation of technology, industry knowledge, organization, business and process. The report begins by stating that the core value of digitisation of R&D is the transformation of the R&D process, using the flow of data to shorten the decision-making chain in R&D. However, the flow of data is not the sole domain of technology. Changes in business processes and organizational structures can also revolutionize R&D processes and save significant costs. However, some process changes are contrary to traditional cognition or involve multiple interests, leading to some project leaders to cover up their own difficulties in digital cognition on the pretext of technology. Therefore, in the era of rapid technological development, enterprises that have the ability to take into account both technological application and organizational strength can stand up in the competition.
Fourth, research and development of digital implementation measures
(I) Key factors and relevant suggestions in the early stage of digitalization: Everything in advance is successful, and nothing in advance is lost
- Top-level design: Take top-level design and management’s support as sufficient and necessary conditions to promote the digital process from top to bottom, in small steps, rather than a single module or a single department. It is necessary to establish a comprehensive framework based on the enterprise’s own resource endowment, combined with business status, organizational status, industry trend and technology maturity to confirm the vision and goal of digital transformation.
- System cooperation: The huge auto factory system includes communication and cooperation between departments, businesses and even at home and abroad. Before the establishment of the system, full consideration should be given to the collaboration between the original system and the new system among multiple subjects, as well as the mutual recognition between different data formats, so as to avoid data islands. At the same time, attention should be paid to the consolidation of the system, which is gradually transformed from the non-blind diffusion system to the simplification and consolidation system. In the later period, when the data type changes, the data acquisition and analysis function of the system will be rapidly increased, and the data flow will drive production, sales and other links to feed the data of research and development.
- Talent introduction: The auto factory system is usually good at the design and manufacture of hardware and some electrical and electrical parts, but lacks digital-related talents, which makes it difficult for existing talents to talk to service providers or even rely on service providers excessively, so that process is more important than result. At the same time, some middle-level personnel tend to ignore some factors that should not be ignored in the implementation of digitalization due to complicated affairs. It is difficult to see the overall blueprint of digitalization and bring strategic thinking of digitalization to senior management. As a result, professional teams cannot be formed to promote digitalization.
- Process sorting: It is necessary to recognize the nature of digitalization. The introduction of digitalization function configuration is not the introduction of technology itself, but the use of technical ability to promote the change and re-sorting of research and development process.
- Organizational structure: The digitization process combines business, IT and other departments, and the long-term cross-department cooperation requires the support of higher-level digital departments, especially the inter-disciplinary talents who understand business, technology and strategy, so as to ensure the comprehensive coordination within the group and the precise empowerment after digitalization.
(II) Path selection of digital transformation to comprehensively promote the digital process with the strategy of overall layout and pilot first.
Digital transformation is a long-term plan of life and death for an enterprise. During the process, it is not necessary to carry out single point planning one by one. Otherwise, problems such as coordination between systems, balance of interests and process mechanism will be amplified infinitely with the expansion of transformation dimensions. Therefore, at the source of digitalization, data governance and business process sorting should be done first. Unclear offline process is meaningless for online. After consolidating the data and process foundation, the digital top-level design is carried out according to the resource endowment and development strategy of each enterprise, so as to build an overall framework of internal and external collaboration, compatible goals and outline. However, it does not mean to solve all problems through the grand top-level design, but to carry out pilot construction in a collaborative way in the part that can see the results as soon as possible, in the global thinking of top-level design to see the local results and by comparison after comprehensive promotion, in order to achieve high-quality development of enterprises.
Profound industry recognition and successful cases are the key standards for the selection of digital service providers.
The application of cloud platform and digital twin technology is still in the stage of exploration, so the decision is more cautious, giving priority to their knowledge of the automobile industry and successful cases with high relevance and portability. In cloud services, multiple deployment options, security, and connectivity are major considerations. In the field of digital twinning, compatibility and mutual recognition of existing software, as well as vehicle level virtual co-simulation capability become the main criteria for selection. Research and development platform is more monopolized by foreign manufacturers, which includes the application of mathematics, physics and other basic disciplines, so it has high requirements on the underlying ability. Oems, on the other hand, know relatively little about virtual reality and pay more attention to the overall capabilities including hardware and software.
(4) Digital service provider capability graph professional software is dominated by foreign giants, and the pace of domestic autonomy needs to be accelerated.
Most of the excellent products, such as R&D software and digital twin, come from international giants. Under the trend of localization and autonomy, Oios also hope to obtain an independent and controllable R&D environment of the whole process, so as to maintain their core competitiveness in a more complex international environment. Therefore, if Chinese service providers have accumulated on the underlying technology, business model and knowledge precipitation, Or will combine with the development of digital economy to accelerate the process of core software localization.
(v) Main contradictions and suggestions in the cooperation between automobile enterprises and service providers to explore digitalization landing strategies in the way of joint construction by both parties.
At present, some digital service providers have insufficient understanding of the R&D business process/pain points of the automobile industry, which makes it difficult to impress customers. In addition, the relative lack of customization ability and the relatively closed ecology lead to two major conflicts in the cooperation between automobile enterprises and service providers. First, the business process of the OEM does not match the product function, which leads to the need for passive process change or passive product reorganization, which consumes communication costs and prolongs the digital delivery cycle. Second, the products of different service providers are difficult to get through, which makes the cloud platform and digital twin in different business links have to be independently deployed, and it is difficult to give full play to the maximum efficiency.
Therefore, service providers need to enhance their understanding of the automobile industry, provide solutions with strong compatibility and high degree of customization, and view their cooperation with auto companies and other service providers with a more open mind.
Importantly, the digital transformation of automobile enterprises cannot be achieved without the cloud-based infrastructure, and the first step of cloud-based is the governance of underlying data. However, some Ovens are not good at data management due to lack of talents and fuzzy concept. Therefore, digital suppliers may assist automobile enterprises to jointly manage data in the early stage of digitalization, clarify the contact points between data and process, and jointly build digital execution strategies. Meanwhile, in this process, they can gradually make up their own shortcomings and jointly explore digitalization implementation schemes.
本文由数字化转型网（www.szhzxw.cn）搜集而来，作者：艾瑞咨询；编辑：数字化转型网默然。
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。
