After analyzing what the Industrial Internet of Things is connected to, we get the content of the connected items. Next we need to decide which attributes we should digitize for each/category of connected objects. Here we encounter a unique obstacle to the Industrial Internet of Things. The connectivity problem of the objects that need to be connected to the industrial Internet of Things, especially when the devices are interconnected, the connectivity performance is particularly prominent. For example, some devices have open communication protocols and available communication interfaces, and some devices are not open. Agreements, etc., then connectivity is a big test for solution providers.
Our experience is that there are four options to choose from:
1. Use an open protocol for the device;
2. Use the sensor that comes with the device;
3. Add a new sensor;
4. Change the viewing side and dimensions, using a new acquisition mode;
The fourth one, changing the side and dimension of the observation, uses the new connection method is to use the first principle, avoiding the obstacles of the device not opening the agreement or the interface, avoiding the direction that the equipment supplier takes the nose, from the essence Get the data on. For example, to obtain the state of use of the device through energy-efficiency detection, to analyze the failure of the component of the device through vibration sensing, or even the rotational speed, etc., as long as the first-principle principle is used to start with the information you need, instead of passively starting from the data that the device can provide. To provide an approach to the IoT solution. Directly target the information we need, and observe how we can get the information we need, in addition to directly connecting the device, because only the data we get can be “isomorphic” with the data provided by the device. For example, we can install a vibration sensor on our IoT device. From the data obtained by the sensor, we get the device to boot, get the start work, and get the speed of the device. If you don't use the first principle, but you have to connect to the device, you need to collect at least three data, and the device may not be able to give you. This is the typical case of edge calculation. The calculation rules of edge calculation must have the ability to customize. It can be said that edge calculation must be a knowledge container, which can easily integrate the knowledge of customers, manufacturers and even third parties into the container. The script-enabled devices already have preliminary edge computing capabilities, and we need to continue to increase support in this area.
Therefore, by analyzing the connectivity of the enterprise value and the material, we can clearly define the level of connectivity that needs to be connected, and also the connection depth of the connected objects;
An important concept in the hierarchy of things connected to the Internet of Things is the management granularity. For manufacturing, the granularity of management of connected objects is roughly divided into the following levels:
1. Sensing level;
2, equipment level;
3. Production line level;
4, workshop level;
5. Enterprise level;
In other words, we need to define the granularity of data acquisition on the premise of economic feasibility. In theory, the fine-grained size must be better and more valuable than the coarse-grained, but when added to the cost analysis, it may not be as fine as possible, and a balance point needs to be found according to various constraints.