SCADA and DCS Architectures

[clockcleaner]

Hello all, I have some questions regarding what architecture its better for SCADA and Distributed Control Systems (DCS) depending on the application. Nowadays the technology allows the creation of control central stations, gathering and storing the data, providing access to remote control centers and corporate networks. My confusion is focus in the data storage and in the system configuration. I hope you can help me to clarify doubts, and for this I would like to particularize in a couple of examples. For instance a company which has several solar panels stations distributed geographically. In order to control and monitor these stations, there is a control center with servers that store in real time the data received from the sensors. In this case, the architecture to use is a centralized system, which monitors and controls the stations. The configuration and the data storage are all done in the control room, which can be located in the company central offices. Other case: a chemical company which has several plants distributed geographically in different places (long distances). Each plant is independent of the others. Nevertheless, the data of each plant wants to be monitored in the corporate networking or central office. What architecture is better to use in this case? Centralized or distributed? Mixed? If a distributed control system (DCS) is used, the distributed scada would be installed in each plant in a control room of the plant and the configuration would be done there. Each chemistry plant would have his local data storage, which could be transferred to the central office for security. If a centralized SCADA architecture is used, there would be one big center control room that could be located close to the main offices of the company. The servers for historical and alarms would be installed there, so corporate computers would be accessing the data easily. The configuration of each plant would be done in this control room. That would save a lot of license costs. Questions: 1) If DCS is used, the users at the main central office could access the data by installing clients to each server in the plants. It seems that many client licenses would be needed to access the information, increasing the cost. Or the solution is to transfer all the databases to the central office and the client access the data there? What would be the solution for this? 2) If centralized SCADA is used, the data are stored in real time in the central control room which could be miles away from the plant. Therefore plant operators are watching information in scada clients that are taking information from the central office. Not sure of this, but then the information is travelling from plant to the control center and from the control center to the client scada in plant, double distance, when the operators are located few meters to the devices. What happened if the connection fails or it’s slow? And this is only if is supervising, what would happen if a control is needed, for example, stop a machine from the control center. 3) If centralized SCADA is used, the configuration of each plant would be done in the central control room, again miles away. It can happen that plant changes constantly, introduction of new sensors, removing control, etc. Does the plant have to call to the central control room to apply these changes? Would it be easier to change and test this inside the plant? Or the configuration could be realized in the plant, although the scaca is the central office? These questions arise as well in other type of industrial control, such as refineries, construction sites, …places where the central office (corporate network) is far from the plants. What is commonly used in these cases? DCS or SCADA? Mixed? Thank you very much in advance for clarifying these concepts.

[Ken Moore]

I have used a third option. The local plants have a DCS installed along with a data historian. The DCS configuration done locally and iis backed up to a central location on a routine basis. The data historian collects data at the local site and stores it locally and at the central location. The central location monitors the historian data, while this is not real time, it is good enough for monitoring, but not controlling.

[NoelWestover]

You have raised vital points..Were you able to find correct answers? Seems like each option has some minor issue.. Edited April 2, 2014 by NoelWestover