为期两天的ECIO 2020华东CIO大会在上海顺利召开。本次大会两天共五大会场。主会场上大咖云集共享共论“数字化智能化新未来新探索”。分会场上各类行业大咖汇聚，就“流程与变革”、“新技术与创新”、“智造与互联”、“安全与基础架构”四大主题进行交流探讨。华东CIO大会www.ecio.com.cn
本次大会云集了江浙沪等东部沿海地区的CIO参加，总计超过3000人注册报名，是一次规模巨大的行业盛会，促进了疫情后期的行业交流。
原福耀玻璃副总裁、CIO夏乐冰受CXO UNION、华东CIO大会组委会邀出席了本次峰会，并发表了主题演讲。
以下为原福耀玻璃副总裁、CIO夏乐冰在本次会议上分享的部分节选：
一、企业的本质是什么？
从1912年到现在，一百多年过去了，企业的本质其实是没有变化的。制造业追求的是什么？是质量、成本、效益。既不是大数据，也不是工业互联网，实际上从头到尾这三者是没有变化的。那么我们今天的路应该怎么走？从这三个指标出发，项目做完了会给IT和CIO一个什么样的评价？
企业是一个价值链的平台，为什么？因为从工艺设计、制造到交付，再到社会服务就是一个价值创造、价值传递、价值认同的过程。哪一个企业在价值链上做的越多就越有竞争力。对于价值链，只要里面有一个环节做的好，这个企业就真正做的好，就可以做的很好。但是如果一个企业想把每一个环节都做的很好，可能什么都做不好。
二、新技术or忽悠？华东CIO大会www.ecio.com.cn
我见过很多企业经受过各种各样新技术的忽悠，也在这上面花了很多的时间，包含工业互联网、区块链，云计算等等，用了一阵时间后会发现这些新技术并没有什么特别大的用处。
在制造业里，从现场设备、设备互联、车间执行的AMS到计划研发，工业制造业基本上所有的软件都用了。如果是商业企业的话，他可能只用其中一个部分，但制造业开始到后面大数据整个一层一层的，层层叠加上来就全部用上，但这些系统用上去完以后，会发现一百套系统以上的就是各种各样的系统。
现在的互联网叫工业互联网，为什么成千上万家都上了，但没几家用得好的？因为工厂制造现场波动的因素太多，没有一个计划能够覆盖那些因素，所以说当按照计划去执行的时候，不管是物料也好，机器也罢，各种不确定性会把整个厂线给淹没掉。所以究竟是新技术还是忽悠，还是需要对其进行一个判断，避免走一些冤路。
三、打通数字链条
为了打通数字链条，很多企业花了很多时间，做了很多的系统建设，也投入了很多钱，甚至连数据治理都要弄进去。但讲到现在最底层的那些数据，电商的数据可以做到99.997%，制造业是几位数的差错率。华东CIO大会www.ecio.com.cn
之前有看到一家很先进的工厂用了一千多台的AGV，全部无人厂，大屏幕显示的齐套率只有60%多，这些系统上了意义大吗？其他地方很多是制造，但现场根本找不到那个调。所以工单下去以后根本无法执行的，就在现场调来调去，再按计算推送物料拿来拿去整个混乱一盘。以前我们以为说AI是为了增强人的技能，我们原来在讲说要符合工种，要符合什么样的功能，让他能够增加柔性化。我们现在讲流程化、数字化、智能化。但流程化不单单是整个IT流程的问题，甚至包含整个组织的架构，也是从课程式组织往流程化组织转化的问题，更是如何往前走的问题。所以说企业流程化、数字化、智能化还存在很多的问题。企业还有哪些未解之迷，我们今天还没有发现的。我们IT不能以上系统、上线为目的，而应该去思考企业有哪些关键性因素是还没有解决的。所以其实系统并不是越复杂越好，而是要解决人的问题，利用上下集成、横向、纵向的简单的系统，而且对于我们制造业来说要从研发设计开始把整个数字链条打通。
四、智能制造五个层次
1、状态感知
比如很多人会做蛋糕，蛋糕需要很多材料，要配料、混合、配淀粉、巧克力。首先要把材料混合起来，再进行烘培，烤完后加巧克力酱。它的工艺虽然很简单，但仍可能会出现各种各样的质量问题，可能太干了很难嚼等等。华东CIO大会www.ecio.com.cn
2、适时分析
例如在做蛋糕的过程中出现的各种质量问题，在各种各样系统里面都会有很多信息，但有用的信息很少，我们很难从海量的信息里面获取这些价值。我们现在讲AI，看到的只是事实分析，但系统要的是成因分析。所以说我们要去了解这个搅拌系统，去整体分析它为什么是这个样子的？搅拌之前先放糖还是先放鸡蛋。放了多少糖出现了这些问题，批次是什么？
3、自主决策
我们其实想要的是一个专家型的系统，即事先能够知道出现什么问题，采用什么过程，跟生产过程的方方面面，从一道道蛋糕生产过程的当中去感知问题的存在，然后自主的去进行一个决策。
4、精准执行
真正的智能制造的AI应该是能够从质量上面洞察发现问题的，能够精确的进行产品向前和向后的双向跟踪，在整个时间轴里面查看整个供应链，根据采购订单等进行序列号的跟踪，能够分析到设备状态、传感器参数、操作工等。
5、深度学习
现在有很多平台跟技术手段已经提供了这些东西让我们去学习，能够不断的去提升，可能开始做只有40分，慢慢的60分、80分，不断地自我学习，自我提升。
五、仓储系统与智能制造
01 我们今天想做的事情，其实是可以从EDI开始。因为整个是从零部件的链条开始的，从EDI开始到订单，到供应链的整个计划，再到整个采购需求，到制造，到整个物联，整个无人化福耀现在都可以做到。从EDI开始无人化创建订单，无人化创建生产工单采购单，再到采购，最后到制造。华东CIO大会www.ecio.com.cn
02 现场作为最关键的一块，其实应该要有一个比较特别的就是现场实时的感知，有一个成本的控制。
03 仓储系统在智能制造里面是很重要的一环，如果仓库管理不好，效率很低的话，所有其他东西都不要去想。现在先进的企业仓库管理系统底层有各种各样无人化的机器人、机械臂、叉车，以及各种各样的打包机等。还有像电子标签、工业像机、穿梭车等技术。它底下就一个系统叫RCS或者叫RMS。这些东西现在做的比较好的，像海康威视现在在这一层上的比较好的仓库管理的叫WES，包含搬运、分检等这些都是无人化的。最后上的是一个WMS系统。两年前我看WMS系统的时候，我用过erp厂商的等等，反正福耀仓库管理系统用了十几家，当时说他们做的还不错。后来去看了其他的系统，专业的wms系统，你才会知道什么叫仓库管理，我在讲的智能制造，你一定要用那些细分行业排名前三的那些领导者。华东CIO大会www.ecio.com.cn
04 今天的仓库管理是大量应用人工智能算法、智能理货、商品动态调整等，而且结合移动机器人最大化的提升运营效率。所以几万平米的仓库只要一个人就可以管理。
05 入库智能模块包括入库、配送、出库三个环节。
智能入库 ：指整个东西都可以通过物人感知、自动卸货、自动扫码，最后那个东西自动处理出来是什么物品，体积是多少。
智能配送 ：指配送环节怎么根据订单，根据生产现场精益拉动整个流程，这个是由货到人、是柔性搬运。
智能出库 ：其实在运行一套系统之前，或新设计一个流程之前，我们一定要通过一个仿真算法，这个就是我讲的，原来的投资可能就是说做智能制造可能投资很大，现在投资可能两到三年就可以完全回收。华东CIO大会www.ecio.com.cn
华东CIO大会www.ecio.com.cn
翻译：
The two-day ECIO 2020 East China CIO Conference was successfully held in Shanghai. There are five venues in the two days of this conference. In the main venue, celebrities gathered to discuss “New future and new exploration of digital intelligence”. Major figures from various industries gathered in the forum to exchange and discuss four themes: “Process and chang. “New technology and innovatio.”. “smart manufacturing and interconnection”, and “Security and infrastructure”.
This conference gathered CIOs from Jiangsu, Zhejiang, Shanghai and other eastern coastal areas to attenA. A total of more than 3,000 people registered for registration, is a huge industry event, promote the industry exchange in the late stage of the epidemic.华东CIO大会www.ecio.com.cn
Xia Lebing, CIO and Vice president of Fuyao Glass, was invited by CXO UNION and the Organizing committee of CIO Conference in East China to attend the summit and delivered a keynote speech.
What is the nature of business?华东CIO大会www.ecio.com.cn
More than a hundred years have passed since 1912, but the nature of business has not changed. What is manufacturing after? It’s quality, cost and benefit. It is neither big data nor the industrial Internet. In fact, all three remain unchanged from beginning to end. So where do we go today? Based on these three indicators, what kind of evaluation will be given to IT and CIO after the completion of the project?
Enterprise is a value chain platform, why? Because from process design, manufacturing to delivery, and then to social service is a process of value creation, value transmission, value identification. The more an enterprise does in the value chain, the more competitive it is. For the value chain, as long as there is a link to do well, the enterprise will really do well, can do very well. But if an enterprise wants to do every link well, it may do nothing well.
New technology or deception?
I’ve seen a lot of companies go through all sorts of new technologies and spend a lot of time on them, including the industrial Internet, blockchain, cloud computing, and so on, and after a while it turns out that these new technologies are not particularly useful.
In manufacturing, virtually all software is used in industrial manufacturing, from field equipment, device interconnection, AMS executed on the shop floor, to planning and development. If it is a commercial enterprise, it may only use one part, but the manufacturing industry begins to use the whole big data layer by layer, layer by layer, but after these systems are used up, it will find that more than 100 systems are all kinds of systems.
Now the Internet is called the industrial Internet, why tens of thousands of people are on it, but few use it well? Because there are too many factors that fluctuate in the manufacturing site of the factory, no plan can cover those factors, so when the plan is implemented, whether it is materials or machines, various uncertainties will drown the whole factory line. Therefore, whether it is a new technology or a fudge, or we need to make a judgment on it to avoid some unjust road.华东CIO大会www.ecio.com.cn
Break through the digital chain
In order to break through the digital chain, many enterprises spend a lot of time, do a lot of system construction, also invested a lot of money, even the data governance. But when it comes to the lowest level of data, e-commerce data can achieve 99.997%, manufacturing is several digit error rate.
I have seen before that a very advanced factory used more than one thousand AGVs, all unmanned factories, and the integration rate of the big screen was only more than 60%. Do these systems have great significance? There’s a lot of manufacturing in other places, but you can’t find that tone on site. So after the work order can not be executed, on the site to transfer, and then push the material to take the whole mess.
,we thought that AI was to enhance people’s skills
In the past, we thought that AI was to enhance people’s skills, but we used to talk about what kind of work should be in line with, what kind of function should be in line with, so that he can increase flexibility. We are now procedure-oriented, digital and intelligent. However, process is not only a problem of the entire IT process, even including the entire organizational structure, but also a problem of transforming from a curricular organization to a process organization, but also a problem of how to move forward. Therefore, there are still many problems in the process, digitalization and intellectualization of enterprises.华东CIO大会www.ecio.com.cn
What are the unsolved mysteries of enterprise that we haven’t discovered today? We can’t IT for the purpose of the above system, online, but should think about what key factors in the enterprise are not yet solved. In fact, the more complex the system, the better, but to solve human problems, the use of top-down integration, horizontal and vertical simple system, and for our manufacturing industry to start from research and development design to break through the whole digital chain.
Five levels of intelligent manufacturing
State perception
For example, many people can make cakes. Cakes need a lot of ingredients, ingredients, mixing, starch, chocolate. The ingredients are first mixed, then baked, and then the chocolate sauce is added. The process is simple, but it can have all sorts of quality problems, it can be too dry to chew and so on.
Timely analysis
For example, there are all kinds of quality problems in the process of making cakes. There is a lot of information in all kinds of systems, but there is little useful information. It is difficult for us to obtain these values from massive information. When we talk about AI, we see fact analysis, but systems need cause analysis. So we need to understand the mixing system, to analyze it as a whole, why is it the way it is? Sugar first or egg first. How much sugar was put in to cause these problems, and what was the batch?
Autonomous decision making
In fact, what we want is an expert system, that is, we can know in advance what problems occur, what process to use, and all aspects of the production process, perceive the existence of problems from the production process of a cake, and then make a decision independently.
Precise execution
Real intelligent manufacturing AI should be able to detect problems from the perspective of quality, accurately track products forward and backward, view the entire supply chain in the entire timeline, track serial numbers according to purchase orders, and analyze equipment status, sensor parameters, operators, etc.
Deep learning华东CIO大会www.ecio.com.cn
Now there are many platforms and technical means that have provided these things for us to learn, and we can continuously improve ourselves. Maybe we only get 40 points at the beginning, and then gradually get 60 points and 80 points, and constantly learn and improve ourselves.
Storage system and intelligent manufacturing
01
What we want to do today is actually to start with EDI. Because the whole thing starts from the chain of parts, from EDI to order, to the whole plan of the supply chain, to the whole procurement demand, to manufacturing, to the whole Internet of Things, the whole unmanned Fuyao can now be done. EDI begins with the creation of orders without humanization, which leads to the creation of production work order and purchase order, then to procurement, and finally to manufacturing.
02
Site as the most critical piece, in fact, should have a relatively special is the real-time perception of the site, there is a cost control.
03
Storage system is a very important part of intelligent manufacturing, if the warehouse management is not good, efficiency is very low. All other things don’t think about. Now the advanced enterprise warehouse management system has a variety of unmanned robots, mechanical arms, forklifts, and a variety of balers. There are also technologies like electronic tags, industrial cameras, shuttle cars and so on. Underneath it is a system called RCS or RMS. These things are doing well now.
For example, Hikvision’s good warehouse management system on this floor is called WES, which includes handling, sorting, etc., all of which are unmanned. The last thing on is a WMS system. Two years ago when I looked at the WMS system, I used erp manufacturers and so on, anyway, Fuyao warehouse management system used a dozen, at that time said that they were doing well. Later, when you look at other systems, professional wms systems, you will know what warehouse management is. I am talking about intelligent manufacturing. You must use the top three leaders in the industry segment.
04
Today’s warehouse management is a large number of applications of artificial intelligence algorithms, intelligent tally, commodity dynamic adjustment, and combined with mobile robots to maximize the efficiency of operation. So tens of thousands of square meters of warehouse as long as a person can manage.华东CIO大会www.ecio.com.cn
05
Warehousing intelligent module includes warehousing, distribution and warehousing.
Intelligent storage. It means that the whole thing can be sensed by the object person, automatic unloading, automatic scanning code. And finally the object can be automatically processed to figure out what the object is and its volume.
Intelligent distribution: refers to how the distribution link lean pull the whole process according to the order. According to the production site, this is from goods to people, is flexible handling.
Intelligent delivery: in fact, before running a set of systems, or a new design of a process . We must pass a simulation algorithm. This is what I said, the original investment may be that intelligent manufacturing may be a big investment. Now investment may be two to three years can be completely recovered.
发表回复