Mes should be cautious about big data

Suddenly, "Industry 4.0", "Made in China 2025", "Big Data", "intelligent manufacturing", "smart factory" and other words become popular, as if not to talk about new words become outdated, as if the traditional MES has not adapted to the new era.

But is that really the case? Is there really a few people who can tell these new words clearly?

In my personal years of experience in MES projects, I think China manufacturing is still in Industry 2.0, at most in industry 2.5, most of the enterprises are far from the realization of automation, not to speak of intelligence.

To go straight to industry 4.0, to achieve the curve overtaking, this is a very plump ideal, but it is the infrastructure, cost, talent, management system and many other factors, it is difficult to land.

For example, in the case of automation, the current mainstream technical framework is difficult to interface with big data.

We know that MES is located between the planning layer and the control layer, MES on-site manufacturing data mainly from the client and the field equipment, and a considerable number of equipment is controlled by the PLC.

In China's mainstream Siemens 300/400 PLC, for example, the first thing the PLC to ensure is the stability of the business, followed by data collection and processing, and PLC in the formation of the network by many conditions constraints.

Firstly, when the PLC communicates with IT system through OPC, it is limited by the capacity of transmitting data. The OPC is set to single-sided, asynchronous communication, each communication cannot exceed 200 bytes.

Secondly, the communication stability is guaranteed when the PLC is PN/PN coupler Network, but the capacity of the switching zone is only 256 bytes.

Third, the PLC power-down capacity is very limited, 300 series up to only 700KB, so when the PLC for data cache processing, can maintain the data capacity is restricted.

The PLC working memory is also very limited, generally in a few m bytes, it is difficult to do a large number of data configuration on the PLC.

Many of the articles we see about manufacturing big data often introduce data that is common to many business-critical processes, so the impact on the business is limited in spite of the number of data collected.

On the control site only, first of all to ensure that the validity of the data, consistency, and then consider collecting as much data as possible.

At present, the PLC is still the most effective field control, but the limitations of PLC in data processing has not been a good solution in the short term, it is difficult from the technical aspects of both consistency and big data processing.

Instead of talking big data, it is better to step by step the automation, and then to realize intelligent and big data.

Mes should be cautious about big data