Data model, the digital twin
Data Quality, Qualification and Semantics
The notion of data semantics is important. The information pyramid below presents in a pictorial way the types of data according to the richness of the information processed:
The layers of the information pyramid are :
Layer 1, raw data: this is the data that we find directly at the sensor output. For example, a temperature sensor will give us, for a temperature of 23 ° C, a raw data of 23.
Layer 2, information: the information corresponds to raw data enriched with metadata or attributes allowing it to be specified. In the previous example, we can enrich the data with their unit, their sensor number, their date ... Initiatives like Haystack aim to standardize attribute names in order to be able to more easily find the data in a table
Any table in an SQL database, object in programming, object in a BIM model ... can be considered as information. At this scale of representation, the links between objects do not exist. Linked data can be retrieved provided you know:
The nature of the attributes associated with the object
The table to which the attribute refers
In an SQL database for example, only information is stored. The DBMS (Database Management System) does not know a priori the links between objects, but it is possible for an operator knowing the structure of the database (nature of the attributes, target table, etc.) to write a specific request allowing to find the data (make joins). Without this knowledge of the structure of the database, it is not possible to find the relationships between data.
Layer 3, context : a context is a data graph in which each piece of information is a node and the nodes can be linked together by relationships. In a DBMS graph, nodes and relationships are base objects (both of which can contain metadata). Thus, the DBMS graph knows the nodes and the relationships and is able, in a single request, to return a complete graph while maintaining the consistency and consistency of all the data and the relationships between data.
BIM, Building graph, BRICKS… models are based on this typology of representation. A digital twin is also based on this typology of representation. The SpinalCore BOS is based on a DBMS graph.
Layer 4, wisdom: in an information system, applications, analysis system ... play on this layer. Applications use contextualized data to add value (decision-making, action, advice, etc.)
The objective of the BOS is to manage the digital twin of the building (also called the dynamic building graph). It is a software platform that works on layers 2 and 3 and which allows the management of data graphs to facilitate the development of applications on layer 4.
The SpinalCom BOS is based on the SpinalCore core which implements:
An in-memory, graph-oriented, distributed DBMS
An asynchronous & bi-directional graph replication protocol
A distributed event management system.
Digital twin, unique repository & Building Graph
The single building repository ensures the consistency and updates of all systems connected to this repository (like LDAP to manage people on a computer network).
The diagram below organizes this repository in the form of a spatial tree :
The notion of pivotal data is important to understand what is the core target, the role and the positioning of each system. These data provide important information about what this system makes sense of. This also makes it possible to distinguish between the different solutions:
A Field Information System (FIS) or BMS will focus on data from measurement points or equipment. This means that the other data (building, floor, room, etc.) are generally only attributes (metadata) associated with the points and not fully-fledged objects in the database. Haystack, Bricks and more generally GTB or GTB middleware initiatives work with this pivotal data.
A Building Information System (BIS) will focus on a detailed description of the building. This means that each floor, room, structural elements… is an object of its base. On the other hand, the address, the region or the city are attributes of the building. The BOS SpinalCore is based on this type of pivot data and uses the BIM model (s) to initialize the building data repository (or building graph).
An Information System (PIS) or BaaS portfolio will focus on a detailed description of the heritage and a “rough” view of each building. In a CMMS for example, the tree structure starts out looks like contracts / region / sites / buildings / floors / rooms / group of equipment. The details of the luminaires in a room and their positioning are rarely specified and the walls, floors and other equipment that do not directly fit into the maintenance contract are not described. They will be the subject of a note, an attribute in a ticket linked to a part. A GIS (Geographic Information System) also falls within this framework.
Data models managed by the BOS SpinalCore
In principle, it is possible to manage any type of data model in SpinalCore technology. The SDK allows you to create a new data format if it does not already exist.
Today, SpinalCom provides and supports in its digital twin, a set of open data models which can be summarized in the following diagram :