NOGUCHI Akira1 SAJIKI Jirou2 OHTANI Tetsuya3
As the market and social environment surrounding industrial production systems change, production systems for individual users must respond flexibly. We call the environment and fundamental technologies to support this flexible change "Field-ubiquitous Computing," and are developing key technologies in this area. This paper reports on our R&D in Field Ubiquitous Computing toward creating a flat system infrastructure, a system architecture capable of continual growth, and services providing essential information.
- Ubiquitous Field Computing Research Center, Corporate R&D Headquarters
- Network & Software Development Dept., Corporate R&D Headquarters
- Instrument & Control Research Center, Corporate R&D Headquarters
INTRODUCTION
The circumstances surrounding industrial production systems have been changing along with changes in the market and social environment. In addition to safe, efficient production which has always been the aim of manufacturing, systems must be increasingly flexible to meet growing issues such as optimizing all global activities and addressing environmental and energy needs. At manufacturing sites (hereinafter referred to as the "field"), systems must flexibly adapt to the changes for individuals involved in broad-ranging issues.
In this paper, we describe our system approach to handling those changes and diversification of the issues by introducing ubiquitous computing technology, such as Internet technology, to the field.
UBIQUITOUS SOCIETY AND UBIQUITY AT MANUFACTURING SITES
The Internet has unleashed new services and business models through its flat interconnection environment regardless of distance and organizational hierarchy, and by bringing together the diverse values of companies and individuals. Moreover, the popularization of personal information equipment, such as mobile phones, has created an environment where individuals can obtain services and information through networks. A ubiquitous society where information can be accessed at anytime, from anywhere, and by anybody is emerging. This ubiquitous society keeps evolving to meet tangible and potential expectations of diverse individuals; the system is not built according to a plan for certain purposes. This creates different characteristics for the development process compared with conventional systems.
As the Internet has spread, companies have started to expand beyond national borders to obtain advantages in global competition. They are working on optimizing their production systems linked with business information system and supply chain management. However, future issues include population aging and low birthrates in developed countries, and the worldwide issues of energy saving and global warming. The people involved in solving these issues will increase beyond the boundary of conventional production organization, and the information and functions required by them will have different scopes and life-cycles for each issue.
We call the environment and its technologies for supporting individuals' issues related to production systems "Field-ubiquitous computing." This means ubiquity in the field of production. Field-ubiquitous computing is an environment where devices such as sensors and actuators in the field are connected to the network and can be accessed from anywhere, and devices on the network capable of autonomous operation can process maintenance information and alarms for users in cooperation with each other before they know it. Furthermore, the environment enables system functions to be reconfigured by providing information with a "realistic feeling" according to each user's role and from the user's point of view, adapting to continually changing users' issues.1
ENABLING FIELD-UBIQUITOUS COMPUTING
We are researching and developing key technologies for Field-ubiquitous computing in three core areas: infrastructure, system architecture, and information provision.
Achieving a flat infrastructure
At higher-level layers than controllers in existing industrial production systems such as the Manufacturing Execution System (MES) layer, the introduction of information technology is advancing, and there is increasing cooperation among systems as well as global level integration. Meanwhile, in the controller layer and lower layers in the field, new applications such as device diagnosis and asset management have emerged as field information has gone digital. However, networks are based on various field buses, and the higher layers and the field are configured on different networks.
 |
| Figure 1 Paradigm shift to network-centric model |
To achieve Field-ubiquitous computing, we believe that a flat infrastructure, where the entire system runs on a unified network with transparent access to information, is necessary. This flat infrastructure allows field devices themselves to provide functions for solving the specific problems of each user. This means that a system different from the conventional hierarchical system for controlling production facilities is needed in the field. Of course, this system must have characteristics as a mission-critical system which is required for the control system.
Among the many types of conventional control systems, it is common to define the system by relating the sensors and actuators of production facilities with the controllers that handle the control functions. Thus, the system is based on a controller-centric model. Although this model is excellent for guaranteeing consistency among control functions, it is necessary to inform changes in the system configuration and addition of exchange information to the controller itself since the controller manages the addition and information of devices at sites. Usually, there is a trade-off relationship between the stability of a control system and the flexibility of changing connection information after system startup. To ensure both stability and flexibility, a model is required in which the controllers have only those functions necessary for control. Also, autonomous devices in the field are desired to transmit the information directly for the usage beyond the control function. To achieve this, as shown in Figure 1, a paradigm shift from the conventional physical network configuration with controllers and devices, to a network-centric model based on a flat network. Furthermore, the structure must allow controllers and sensors to be added and logically connected as needed.
Evolution of system architecture
 |
| Figure 2 Overlay type operating model |
With the Internet, systems have grown overlapping various virtual and independent sub-systems on the network by adding nodes by many unidentifiable participants freely. The flat network environment mentioned in the "Achieving a flat infrastructure" section will provide a mechanism for adding devices and expanding the functions of field devices to meet users' requests dynamically, and suitable services will be provided incrementally. However, because production systems are aimed mainly for production as a mission-critical system, it has been very difficult for non-production departments to introduce functions in the field that are not directly related to production. To solve this problem, we propose the overlay type operating model shown in Figure 2.
In this model, system resources are allocated to each user or application according to its importance, and each user or application can flexibly establish a system within the resources. For this model, it is very important to be able to share computational resources and information resources on field devices, and also network bandwidth under appropriate management without disorderly use of limited system resources caused by each user's free addition of functions.
Services providing essential information for production
The flat network environment mentioned in the "Achieving a flat infrastructure" section allows the system to gather field information from all over the world via the Internet. Since it is practically impossible for each user to access each data considering the physical configuration of each plant, the systems must provide essential information based on the knowledge and viewpoint of each user. The final goal for users is receiving information on the time scale and level of detail they require by converting the data measured by field devices. The user should also be presented with essential information for production, modeling the plant behavior.
SPECIFIC ACTIVITIES FOR FIELD-UBIQUITOUS COMPUTING
We believe that field-ubiquitous computing must offer interoperability between the various devices utilizing the technical infrastructure based on international standards. Yokogawa has joined the Widely Integrated Distributed Environment (WIDE) project since its early stage, and has been contributing to the development and diffusion of Internet Protocol version 6 (IPv6). Also, regarding network security technology, we have made proposals and had discussions in the Internet Engineering Task Force (IETF). In the wireless field, we have participated in the Instrumentation, Systems, and Automation Society (ISA) SP100, and have been conducting technical development based on international standards.
This section introduces each of the activities mentioned above.
IP instrumentation
Internet Protocol (IP) will be increasingly used in control systems. Introduction of IP will start with the backbone network in the field, and then the relationship between controller and field devices will change from physical to logical. This will allow free addition of and access to field devices without going through controllers.
As this relationship progresses, field devices will be expected to transmit information by themselves. As shown in Figure 3, an IP-enabled field device will be able to transmit information from the field to users directly, and the field and information systems will be connected directly. To introduce IP to field-level networks, issues unique to the field relating to devices and wiring need to be resolved.
Figure 3 Network evolution of field systems
The introduction of IP to the field network involves more issues than introducing information systems, such as ensuring a real-time response, implementation on devices with scarce computational resources, and making the system explosion- proof. There are also network security issues when migrating from the conventional closed environment to an open environment.
Yokogawa considers the latest Internet protocol, IPv6, as a next-generation protocol. We are researching the required real- time response, wiring technology, and security architecture on a light security framework achievable in field devices. For more details, refer to "Capability of IPv6 for control network and the activities toward its application" and "Application of IPv6 to Field Instruments level network and Virtual Wiring Technology" in this Yokogawa Technical Report.
Overlay model application environment
We are developing the Field Overlay Architecture (FOA), a system architecture which configures the system specifically for a user on top of the existing system, by applying an overlay model to the controllers and devices in the field and the field network which connects them. By deploying a common virtual machine in the controllers and devices in the field, FOA allows device functions to be enhanced remotely. FOA also provides an enhancement mechanism with minimal impact on existing functions. The aim of FOA is to dynamically secure the necessary resources for each application and provide a virtual environment for each application; this is achieved by controlling the resources and operating conditions of virtual machines on multiple controllers and devices in the field. This operating model makes it possible not only to enhance functions, but also to configure the system newly required after system startup and to run it independently from the existing system. For the details of FOA, refer to "System resource management for achieving Field Overlay Architecture" in this Yokogawa Technical Report.
Plant modeling
Conventional control systems provide predefined information to users, such as measured values of the plant, controller information and alarm information, and then each user uses the information selected from that provided by the control system. However, in future, control systems must generate and provide information from the user's perspective2. To do this, virtual information obtained through real-time simulation by modeling internal behavior of the plant will be important.
Yokogawa has developed the world's first tracking simulator which makes a model of internal plant behavior and simulates it in real time. An example of the plant model in real time usage and its future development is shown in Figure 4. This plant model makes it possible to visualize the inside of the plant, which was invisible before, and to predict the plant's behavior upon an operation. Since various plant behaviors can be analyzed, the system will provide information that meets individual users' expectations, such as quality, energy efficiency and safety allowance. Moreover, optimal operation will be possible by providing quality indicators and energy efficiency as operational indices. Refer to "Utilization of Tracking Simulator and its application to the future plant operation" in this Yokogawa Technical Report for details.
Figure 4 Utilization of plant model in real time
CONCLUSION
In this paper, we have reported on the future of the Field- ubiquitous environment and how we are working to achieve it. To respond to individual users' needs in industrial production systems, we are flattening the field network environment and responding to changes and diversification of systems as well as aiming to provide information tailored to each user. Yokogawa is creating technologies to deliver individualized production systems that can quickly respond to the needs of each user.
REFERENCES
- Akira Nagashima, "Technologies for Achieving Field Ubiquitous Computing," SICE Journal of Control, Measurement, and System Integration, Vol. 1, No. 3, May, 2008, pp. 192-197
- Tatenobu Seki, Gentarou Fukano, et al., "Innovative Plant Operation by Using Tracking Simulators," Yokogawa Technical Report Vol. 52, No. 1, 2008, pp. 35-38 in Japanese