Abstract : This paper analyzes the architecture of existing graphical systems in general monitoring systems, pointing out the shortcomings of the simple mapping relationship between graphical objects and domain entity data. It proposes a new mapping mode between graphical objects and domain entities based on meta-graphics and basic domain entities, aiming to establish a general monitoring graphical system architecture based on this mode to meet the needs of developing ubiquitous monitoring and control systems for the Internet of Things.
Keywords : Internet of Things, general monitoring graphics system architecture, meta-graphics, basic domain entities
I. Introduction
There are many general-purpose graphics systems, but they are unsuitable for monitoring systems because, although monitoring systems display data as graphics, they represent data in a graphical form. While the graphics we see also contain data—for example, a rectangle has length and width—this data is the parameter set of the graphic. The set of parameters is equivalent to the graphic itself. On a specific display platform, we can consider: Graphic = Parameter Set, but these parameters are not yet what we call data. In a monitoring system, the data we refer to are the characteristic values that represent the operating status of the monitored object (such as power equipment). The four remote sensing functions of power system equipment generally represent this type of data. The task of a graphical monitoring system is to use graphics to display the operating status of the monitored object and control its behavior. Therefore, graphics processing becomes one of the key subsystems of a monitoring system. Graphics processing has two main tasks: to reflect the characteristic values of the monitored object graphically; and to define the graphics graphically before monitoring (generally called graphics configuration). The characteristic values are processed by other subsystems and then used by the graphics subsystem.
Due to the unique nature of graphical monitoring systems, many different styles of monitoring systems have emerged. From simple Excel spreadsheets displaying coal mine gas concentrations to GE's DCS configuration system, these systems, while sharing the same purpose, differ in their monitoring scale and level of graphical representation. In recent years, with the development of the Internet of Things (IoT), the use of graphical monitoring systems will increase significantly, as IoT end-users will undoubtedly prefer graphical interfaces to spreadsheet interfaces—a fact already evident when Windows replaced DOS as the operating system.
The architecture of a typical monitoring system is as follows:
Communication layer: Responsible for data acquisition from the monitored object. It uses media such as RS233, RS485, and RJ45 to acquire measurement and control data reflecting the operating characteristics of the monitored object through communication protocols such as MODEBUS and TCP/IP, and then hands them over to the data layer for further processing.
Data Layer: Responsible for receiving data from the communication layer and performing prescribed processing based on the meaning of the data. New secondary data, such as data exceeding limits and false remote communication, will be generated. It also provides shared channels (data forwarding, service interface publishing, etc.).
Presentation Layer: This is the graphics processing subsystem discussed in this article. It is responsible for displaying the data from the data layer in a graphical format. From a data flow perspective, it transforms the data from the data layer into a set of attribute parameters for graphical objects.
However, a presentation layer can use different data layers, and different presentation layers can use the same data layer. This also exists in the communication layer, forming a mesh structure as follows:
Since all presentation layers, data layers, and communication layers are isomorphic, for ease of analysis, they are generally represented in the following simplified form:
If the data layer does not perform data processing, and the presentation layer uses Excel spreadsheets instead of graphical objects, it becomes a very simple monitoring system, as follows:
II. Common Architectures of Current Graphics Systems
Due to the rapid development of the Internet of Things (IoT), the data layer is also evolving towards distributed and cloud computing models. Data as a Service (DaaS) research focuses on delivering the right data to the right location at the right time. The corresponding architecture of the graphical system is as follows:
Data Object Access: Achieving unified access to data objects through different data access methods such as WCF and OPC. The data objects are consistent with the application domain, such as IEC 61970 in the power industry.
Data Mapping (Rule of DataTranslate): The transformation rules for capturing domain data and converting it into a set of graphical object parameters determine how the graphics system uses the data.
ShapePresentation: This uses graphical objects to represent domain objects. Various presentation methods exist, including SVG, WPF, Silverlight (SVL), and HTML, some suitable for web applications and others for desktop applications. Here, the domain objects are the core, and the presented graphical objects are the tables.
As we can see, data acquisition, data mapping, and graphical display all become meaningful only when supported by domain entities. So why are domain entities so important?
III. Analysis
First, data reflects the attributes of domain entities. The current flowing through a power circuit breaker, the water pressure and flow rate in a water pipe, the smoke concentration in a fire alarm, and the temperature at a measurement point are all attributes reflecting the relevant monitored objects, even though they can be measured by multiple devices. Without domain entities, this data has no actual physical meaning. Logically, this data is organized according to domain entities. D = m(E), where D is the data set, E is the set of domain entities, and m is the organizational relationship.
On the other hand, graphics also reflect domain entities. Graphics such as power busbars, water pipe pressure reducing valves, and rooms at fire alarm levels also reflect the attributes of the relevant monitored objects. The same power busbar can vary in thickness and color to represent different voltage levels. Without domain entities, these graphics also lack actual physical meaning. Therefore, logically, these graphics are organized according to domain entities. G = g(E), where G is the set of graphic objects, E is the set of domain entities, and g is the graphic data operation.
Therefore, data and graphics can be considered two aspects of domain entities, like two sides of a coin. Of course, the domain entityization of data has already been completed outside the graphics processing subsystem; the graphics processing subsystem needs to focus on the domain entityization of graphics. Its essence is the correspondence between graphic objects and domain entities. For example, how does a graphic of a power transformer correspond to the power transformer equipment in a power system? Mathematically, this is reflected in the relationship: G = g(m-1(E)). Since domain entities are finite at a certain stage of development, the relationship between m and g is finite, and naturally g · m-1 is finite, making the set G a finite set. Here, domain entities are a theoretical abstraction, and their abstract models differ depending on the focus of the domain (IEC has abstracted a large power domain entity model to accommodate more power applications). The key point is how graphic objects correspond to domain entities given a defined domain model.
Domain entities have a characteristic: they embody one or more features of the domain. For example, a three-winding power transformer will exhibit features such as current and voltage in its primary, secondary 1, and secondary 2 windings. A three-winding transformer contains three power windings, each of which may represent a single power entity, including features such as voltage. Each winding is connected to a voltage level, and each voltage level may also be connected to circuit breakers, grounding switches, and other equipment. These devices together constitute the actual power system. Although this system is still an abstract system when reflected in a graphical system, the set of entities displayed graphically is less than or equal to the set of entities represented by data, making the graphical representation complete. Therefore, the domain model will correspond one-to-one with the set of graphical representations, allowing the graphics to fully encompass the domain model.
It is precisely under this line of thinking that standards for domain graphic symbols emerged (such as the standard for power graphic symbols). Moreover, in these standards, the graphic symbol for a domain entity may be represented by multiple sub-symbols. Clearly, compared to a general graphics system that uses points, lines, and surfaces to represent everything, the concept of domain meta-graphics is closer to reality. Its correspondence with domain entities is also simpler. This leads to a new graphics system architecture as follows:
Metagraphs embody a specific graphical characteristic of a domain entity, and a domain graph can be considered to be composed of multiple domain metagraphs. Since the development of domain business is constrained by overall technological advancements, the increase in metagraphs signifies significant development in the domain business, and their increase is relatively slow. Of course, the main reason is that the categories of metagraphs are generally governed by rules (standards), and most importantly, the number of metagraphs is small, yet the number of domain entities formed by their combinations is enormous, resulting in rich expressiveness.
IV. Advantages of the New Architecture
Compared to the previous architecture, the use of meta-graphs makes the mapping from entity attributes to meta-graphs much easier and more complete. Furthermore, since the graphical display is a combination of meta-graphs, graph categories can be expanded based on the graphs without affecting other modules. In particular, placing the graph definition in a configuration file allows for flexible graph expansion. This means that graph symbols are replaceable, and that the graphical display is a display of the combined meta-graphs; as long as the description of the meta-graphs is standard, the display will be standard. Of course, the implementation methods differ for different display technology platforms. This is something to be aware of. See below:
As we can see, the definitions of rectangles and circles are logical, because circles don't exist in Silverlight. However, when displaying graphics in the Silverlight interface, a corresponding translation is made, and this relationship is fixed. Thus, the definition of meta-graphics is no longer constrained by the technology used for graphic display.
This is not the most important advantage. The meta-graph corresponds to the basic domain entity. If you want to change the domain, you need to change the meta-graph, data mapping, and data acquisition. In this way, the graph system becomes a universal monitoring graph system!
Replacing the meta-graphics is simple; for systems using configuration files, it's just changing the file, and for fixed programs, it's changing the module. Changing data acquisition is also straightforward; for different data layers, simply changing the interface is sufficient. As long as the replacement is done properly, matching the graphics and data acquisition is easy. However, changing the data mapping is not so simple. In the previous system architecture, entity attribute data was directly mapped to and converted to graphic object attributes. In the new architecture, entity attribute data is not directly mapped to and converted to graphic object data. Instead, it must be standardized through constraints imposed by the underlying entity domain entities and the mapping rules to graphic categories. This is far more complex than the previous architecture.
Fortunately, there are already many research findings and tools available in this field. Expert systems, inference engines, and other artificial intelligence (AI) tools are among the better choices.
V. Conclusion
Therefore, it can be seen that due to the application of artificial intelligence, monitoring graphics systems can be customized by changing graphics systems, meta-graphics, data mapping, and data acquisition to achieve a universal monitoring graphics system. Currently, there are readily available technologies and tools for data acquisition, data mapping, meta-graphics, and graphic display involved in the system. The biggest problem is the lack of an integrated system or tool, while there is a real need for such a universal monitoring graphics system. The solution is to integrate existing technologies such as SVG, meta-graphics, transformation rule descriptions, and data entity descriptions to form a domain-specific graphics system that can be freely changed. Such a graphics system will undoubtedly find widespread application in the Internet of Things (IoT) based on cloud services, and may even enter millions of households.
VI. Outlook
A general-purpose monitoring graphics system integrating various technologies can provide a practical end-user system, but it is insufficient to offer a simple and easy-to-use system. If we take a goal-oriented approach, end-users can more readily understand and accept the monitoring graphics system's objectives by describing them in a natural language-like manner. For example, displaying "Smoke Alarm #1" in a "natural light" format at the "100 , 100" position on "Monitoring Page 5," updating data as quickly as possible, and using a "factory preset" alarm method. Such a system is more intelligent and easier for the general public to use. With the development of the Internet of Things, this system will undoubtedly become widely adopted.
If data server push is added [6], data collection at the data processing layer can be performed when the proxy server is idle, which can maximize the improvement of the update speed of monitoring graphics.
Author information : Zeng Mingchang, current main research areas: SOAP system architecture, graphics processing for power monitoring systems.
