The basic structure of DCS and the differences from PLC

Thread Content

There are many forum users who are not very clear about the differences between DCS and PLC, and they have raised questions to varying degrees. For this reason, I previously wrote an article titled \"Differences between DCS Control Systems and PLC Control Systems\" at http://bbs.hcbbs.com/thread-431214-1-1.html. Today’s article can be considered a follow-up to that one; I hope it will be helpful to all forum users! The basic structure of DCS and the differences from PLC: DCS is the abbreviation for Distributed Control System (TOTAL DISTRIBUTED CONTROL SYSTEM). It refers to the control of hazard dispersion, management, and centralized display. In the late 1960s, programmable controllers for performing logical operations were developed. Abbreviated as PLC. It is mainly used in the automotive manufacturing industry. In the mid-1970s, DCS systems with analog control were introduced to the market, replacing analog instrument control based primarily on PID calculations. The concept of DCS was first proposed by the original manufacturers of measuring instruments, and it was initially used primarily in the chemical industry. Later, the computer industry got involved in the development of DCS. In the 00s, microcomputer technology was not yet mature, and computer technology was not sufficiently developed. The operation station, controllers, I/O boards, and network interface boards are all developed by the DCS manufacturers themselves; in other words, all components are dedicated. In the early 1970s, people used minicomputers such as the PDP/1124 to replace the previously used centrally installed analog instruments for control. There are many cables connected to the central control room. If a minicomputer is used as both the controller and the CRT connected to it as the display device (i.e., the human-machine interface). A minicomputer is required to receive signals from thousands of transmitters or other sensors, and to perform calculations for hundreds of circuits. Obviously, the risks are somewhat concentrated. There are as many cables connected to the analog instruments, and once the minicomputer fails, neither control nor display is available. Digital control did not achieve the intended purpose. Later, someone proposed separating control from display. One computer handles the control and calculation tasks, while another computer handles the display tasks. Furthermore, a process as a controlled object may have many points that need to be displayed and controlled; some of these require closed-loop control or logical operations. The various components of such a process possess relative independence and can be divided into several separate units. The input and output points that need to be monitored and controlled within these individual units can then be distributed across multiple computers. Thus, the calculation tasks that were previously handled by a single minicomputer can now be carried out by several or even dozens of computers (controllers). If one of the machines breaks down, it doesn’t affect the overall system. The so-called tactic of “using wolves to replace tigers” means dispersing the risk. It brings together management functions such as display, operation, and printing, and connects the control and display components via a network to form a single system. At that time, some people called this system a distribution system. To what extent should the risk be distributed to be considered appropriate? This is related to the level of development of computer technology at that time. In the mid-1970s, complete decentralization meant that a single controller handled the calculations for one circuit. At that time, since people were not very familiar with digital technology and were accustomed to analog instruments, loop controllers were popular in the late 1970s and 1980s. These controllers were designed to look almost identical to the original analog instruments, so as not to change users’ operating habits, while the PID calculations were carried out digitally inside. One instrument (a computer) performs the control tasks for one circuit. Its price is relatively high, but the risk is dispersed. Then, a communication network is used to connect the various controllers and the CRT-based human-machine interface into a single system. At this time, the network structure is usually star-shaped. The manufacturing cost of the controller for this circuit is too high, resulting in a poor price/performance ratio. Later, to reduce costs, two-channel and four-channel controllers were introduced, which offered a slightly better cost/performance ratio. For large and medium-sized systems, the price/performance ratio of DCS is better than that of systems composed of circuit controllers. In some special cases, circuit controllers are still needed. If too many circuits need to be processed, such as when a controller collects data from thousands of points and performs calculations for hundreds of circuits, the risk becomes extremely high. In this case, the risk must be dispersed. With the development of computer technology, the computing power, storage capacity, and reliability of computers have been continuously improving, allowing computers to handle more tasks. Completed tasks can also be consolidated. In addition, redundancy technologies such as controllers and networks have also been developed, allowing control operations to be centralized to some extent. Based on current DCS systems, a single controller is capable of performing calculations for dozens of loops and collecting data from hundreds of points, along with carrying out appropriate logical operations; in practical use, this approach yields good results. This raises the issue of controller upgrades. Sometimes the distance between the controller and the sensing elements is quite large, which has promoted the development of fieldbuses. Fieldbuses such as CAN, LOONWORKS, FF, etc., as well as HART protocol receiving boards, are all used in DCS systems. DCS is divided into three main parts: the controller with I/O boards, the communication network, and the human-machine interface (HMI). It is connected directly to the production process via an I/O board and a terminal block to read signals from sensors. There are several different types of I/O boards, and each type comes with its own terminal board. Analog input: standard 4-20 milliampere signal board and a millivolt signal board for reading thermocouples ； 4-16 channels, varying amounts ； Analog output typically uses a standard signal of 4-20 milliamps; it usually has a limited number of channels, ranging from 4 to 8 channels ； l Digital input ； 16–32 channels: digital output; there are also boards for digital input and output with different voltage levels, such as 24 volts and 125 volts DC ； 220 volts or 115 volts AC, etc ； 8-16 channels, varying amounts ； l Pulse input, used for the signal of the acquisition rate ； 4-8 channels, varying ； l Quick interrupt input ； l HART protocol input board ； l Fieldbus I/O board ； Each I/O board is connected to the I/O bus. To ensure the safety and integrity of the signal, it must be processed before entering the I/O board. This processing includes checking upper and lower limits, performing temperature compensation, and applying filtering. These tasks can be carried out on the terminal board or separately; boards that are used for signal processing are now sometimes referred to as signal conditioning boards. The I/O bus is connected to the controller. In the 1980s, due to the limited computing power of the controllers in DCS systems, in order to increase the number of I/O points, the tasks of the controllers were divided; in fact, there were three types of controllers. That is: a controller that performs closed-loop operations, an analog data collector, and a logic operator. They each have their own I/O bus, and the I/O buses of different DCS systems vary from one another. If speed is required, a parallel bus is the best choice. Serial buses are generally used more often. Especially when there are many RS485 buses, the I/O count of analog data acquisition devices and logic operators can be higher. Closed-loop controllers, analog data collectors, and logic operators can be directly connected to the human-machine interface over a communication network, with each individual controller on the network acting as an independent node. Each node performs a different function. They should all have network interfaces. Some DCS systems, in order to save on network interfaces, connect all the devices used for process control – namely closed-loop controllers, analog data acquisition units, and logic operators – to the control bus in advance, which are referred to as process control stations. This can increase the number of I/O points that the process control station can receive, while also saving on interfaces. It is then connected to the network via an interface, and linked to the human-machine interface. With the development of computer technology, the computing power of controllers has been continuously improving. For example, a controller based on a PC possesses strong capabilities, being able to handle both analog calculations and digital logic operations. A controller becomes a node on the network. Connected to a human-machine interface via the network. The controller is the core component of a DCS; it is equivalent to a PC. In some DCS systems, the controllers themselves are PCs. It mainly consists of chips such as a CPU, RAM, E2PROM, and ROM, as well as two interfaces: one that receives signals from the I/O bus, and another that sends signals to the network to connect with the human-machine interface. ROM is used to store the control algorithms that carry out various computational functions (in some DCS systems, this is referred to as a function block library). Store functional blocks in the library, such as control algorithms like PID, PID with deadband, integral-separate PID; arithmetic operations including addition, subtraction, multiplication, division, squaring, square root; function operations such as first-order filtering, sine, cosine, X-Y function generators; as well as lead-lag functions ； More advanced algorithms include Smith prediction, C language interface, matrix addition, and matrix multiplication ； Logical operations include logical AND, logical OR, logical NOT, and logical NAND, among others. Typically, station function blocks not only combine analog and digital signals but also connect people. The more function blocks there are, the easier it is for users to write applications (i.e., configure them). The configuration involves connecting functional blocks according to the process requirements to form a control scheme. Store the control scheme in E2PROM. Since E2PROM can be erased and written, and the configuration needs to change as the manufacturing process changes, the configuration is stored in E2PROM. Different users have different configurations. During configuration, the user selects the desired function blocks from the function block library, enters the parameters, and connects the function blocks together. The formation control scheme is stored in the E2PROM. At this time, the controller is in configuration mode; once it is put into operation, it switches to operation mode. The controller is equipped with an operating system, function block configuration software, and communication software. For the safe operation of the system, the closed-loop controllers must operate in a redundant manner, with one in use and one as a backup, and this backup is hot-swappable. To ensure successful redundancy, the following points should be noted: the hardware and software versions of the two controllers must be identical ； Check whether the send-receive chip is intact ； Whether the redundant chip is intact. Check whether the settings of the two modules are the same, as well as whether a manual operation station is available, etc. The communication network connects the process stations and the human-machine interface into a system. Communication networks have several different structural formats. Such as bus, ring, and star topologies (see figure). In logic, a bus topology is also circular. Star shapes are only suitable for small systems. Whether it is ring or bus topology, broadcast is generally used. Some other protocol methods are used less frequently. The speed of communication networks is around 10M and 100M. The human-machine interface has four different types of nodes: the operation station, the engineer workstation, the historical trend station, and the dynamic data server. The u operation station is equipped with an operating system, monitoring software, and driver software for the controllers. Displays system labels, dynamic flowcharts, and alarm information. The engineer workstation is used for configuring controllers (CAD), as well as for configuring operation stations (to create dynamic flowcharts). If the monitoring software has strong graphing capabilities, the graphing task can be completed independently by it. Another function of the engineer station is to read the controller’s configuration, which is used for controller upgrades and fault detection. We call it the reverse engineering station. The historical trend station is used to store historical data, typically using a disk array (referred to as RAID technology). The u dynamic data server serves as an interface between DCS and MIS systems, as well as a device that isolates DCS from the Web. There are significant differences in the design principles of DCS and PLCs. PLCs were developed by imitating the control principles of traditional relays; PLCs in the 1970s had only digital logic control capabilities, and they were first applied in the automotive manufacturing industry. It stores instructions for performing logical operations, sequential control, timing, counting, and arithmetic operations ； And various machines or production processes are controlled through digital input and output operations. The control programs developed by the user reflect the process requirements of the production process, and are stored in the user program memory of the PLC in advance. During operation, it executes each instruction according to the content of the stored program in order to carry out the operations required by the process flow. The CPU of a PLC contains a program counter that indicates the storage address of the current program step. As the program runs, this counter increases by 1 with each step executed; the program proceeds sequentially from the starting step (with a step number of zero) to the final step (usually the END instruction), after which it returns to the starting step to repeat the cycle. The time required for a PLC to complete one cycle of operation is called a scan period. For different models of PLCs, the cycle scanning time ranges from 1 microsecond to several dozen microseconds. Loop operations such as the program counter are something that DCS does not have. This is also why the redundancy of PLCs is inferior to that of DCS. DCS was developed based on operational amplifiers. All functions and the relationships between various process variables are represented as function blocks (referred to as expansion blocks in some DCS systems). DCS in the mid-1970s had only analog control. In a system like the TDC2000, one controller can perform calculations for 8 PID loops per second. The first industry to apply it is the chemical industry. The main difference between DCS and PLC lies in the logical processing of digital signals and the calculation of analog values; even though there has been some overlap between the two over time, differences still exist. After the 1980s, in addition to logical operations, PLCs also had algorithms for use in control circuits; however, performing some complex calculations remained difficult. PLCs are programmed using ladder diagrams, and performing arithmetic operations on analog values is not very intuitive in such programming, making it rather cumbersome. However, in terms of computation logic, it boasts fast performance; it takes less than 1 millisecond to process a 1K logic program, on the order of microseconds. It treats all inputs as digital signals, with 16 bits (or 32 bits) constituting one analog signal. DCS treats all inputs as analog values, with 1 bit representing a digital value. Solving a logic problem takes anywhere from a few hundred microseconds to several milliseconds. It takes the PLC a few dozen milliseconds to perform a PID calculation, which is on par with the processing time of a DCS. Large PLCs use another CPU to perform analog calculations. Send the calculation results to the PLC controller. Different models of DCS take varying amounts of time to calculate PID values, but all are in the range of several dozen milliseconds. Like the early TDC2000 system, it could complete control calculations for 8 circuits in 1 second. With the advancement of chip technology, the time required to compute an algorithm is decreasing. The time required to solve an algorithm depends on the arrangement and configuration of the functional blocks. Regarding ground resistance, the requirements may not be high for PLCs, but for DCS it must be below a few ohms (usually below 4 ohms). Analog isolation is also very important. In areas with explosion hazards, intrinsically safe barriers should be installed. For systems with the same number of I/O points, using a PLC is cheaper than using a DCS (saving approximately 40%). PLCs do not have dedicated operation stations; the software and hardware they use are standard, which means their maintenance costs are much lower than those of DCS. A PLC controller can handle thousands of I/O points (up to over 8,000 I/Os). The controllers of DCS can only handle a few hundred I/O points (no more than 500 I/Os). If the controlled objects mainly involve equipment interlocks with few circuits, a PLC is more suitable. If it is mainly analog control with many functional operations, a DCS is the best choice. In terms of redundancy in controllers, I/O cards, communication networks, etc., as well as for certain advanced calculations and industry-specific requirements, DCS is far superior to PLC. Due to the use of general monitoring software, PLCs make it easier to design management information systems for enterprises. It is particularly important to note that a dedicated operation station for DCS is not an inherent necessity. It was formed for historical reasons. If DCS manufacturers fail to make the operation stations connected to the factory’s management information system, certain DCS systems risk disappearing from the market. With the emergence of new technologies, negative effects follow as well. With the introduction of new operation stations, viruses and hackers can more easily invade the system. When designing, set a password at the operation station, and implement additional isolation and firewalls for the system. Minimize the negative impacts. This post was last edited by Mobei Yihai on 2009-4-6 at 14:04.]