In the second half of the last century, with the development of large-capacity generator sets and the development of computer technology, the control technology and control level of thermal power plants have been greatly improved. Especially in the last twenty or thirty years, the decentralized control system quickly occupied almost all positions. At the same time, a new control theory and control technology is also quietly interested in, and began to open up their own areas in the market, this is the intelligent control system (Intelligentcontrolsystem), and in a short period of time, the intelligent control system has been displayed Its own strong vitality and beautiful development prospects.
Intelligent control system can be said to have emerged. The reason is:
First, users need this technology. Modern thermal power generating units are developing in the direction of large-scale and super-large ones. The domestic 600 MW unit has become the main unit and the 900 MW unit is under construction. With the increase in the size of power grids and the development of electricity in industry, agriculture, commerce, and residential life, with the development of power generation equipment manufacturing technology, larger generator sets will also emerge. The large-capacity generator sets must adopt supercritical parameters, as well as the need for complicated environmental protection technologies. All these factors cause the control systems of thermal power plants to become very large and complex. Although the functions of computers have become more and more powerful, the development of application software has increased the workload by many times. System debugging has become more and more difficult in terms of workload or difficulty. The intelligent control system has its unparalleled advantages precisely in these aspects.
Second, the current power plant construction and planned economic management models are very different. We hope that shortening the construction period and pursuing rapid capital returns will be the common mode of operation of investors. Therefore, a new project, preparation work, feasibility study, tendering, calibration, etc., take a lot of time to reduce the risk of investment. At the same time, once construction begins, a large amount of funds will be invested. Investors hope that the shorter the construction period, the better, and that they can achieve the goal of recovering funds as soon as possible. In engineering odometers, the time for commissioning the control system is often very limited and cannot be guaranteed. For complex control systems, the control logic needs to be tested under various operating conditions and operating modes. If the required debugging time is longer, the debugging work is even more difficult, and even the system cannot be put in time. Can we debug while running? For conventional control systems based on precise mathematics, this is very difficult, but for intelligent control systems, it is its characteristics. It is also its advantages.
Third, the classical control system based on PID law requires an accurate mathematical model when designing, and requires accurate parameters when debugging. Obviously, there are difficulties for this complex system. Intelligent control systems are precisely designed to solve these problems.
Fourth, because of its advanced theory, the intelligent control system does not require the hardware of the computer compared to the classic control system. However, in the long run, it is easy to program, and the commissioning workload is reduced. It is very likely to reduce the investment in the control system. And shorten the construction period.
Fifth, the development of intelligent control systems requires the cooperation of computer technology development. Needless to say, the rapid development of computer technology has provided a vast world for intelligent control systems.
The intelligent control system came into being and has a broad prospect in the field of power plant control. So what is an intelligent control system? What is its current status? What strategy should be adopted in the intelligentization of power plant control systems? This is the issue discussed in this paper.
1. Intelligent Control System Introduction Intelligent control system as a frontier discipline in the field of automatic control, it is difficult to make a complete definition, because there are new theories and new systems appearing at almost any time here. In general, intelligent control is the use of artificial intelligence intuition reasoning methods for self-learning control systems, which can be summarized as a combination of automatic control and artificial intelligence. For example, in a power plant control system, the operational experience of the operating personnel is combined with the control system, and the level of intelligence of the system can be enriched through self-learning. In intelligent control systems, the higher the level of intelligent decision making, the more artificial artificial intelligence is used.
Intelligent control has the following basic features:
â—† The higher the level of control system, the higher the intelligence embodied.
â—†Open-loop and closed-loop control combined.
â—† The combination of qualitative control and quantitative control, that is, the combination of control logic and logic execution process (such as speed, etc.).
â—† Intelligent control has self-learning, self-adaptation and organization functions.
â—† Information processing methods include both mathematical and logical reasoning.
Intelligent control currently has approximately the following categories:
â—† Fuzzy control system based on fuzzy mathematics.
â—† Expert control system.
â—† neuron control system.
â—† Integrated intelligent control system combining multiple methods.
1.1 Fuzzy Control System (FuzzycOntr01sys-
The basic idea of ​​fuzzy control is to generalize the control strategy of the person on the controlled object as follows: IF (condition) THEN (action) forms the control rules of the form, through the fuzzy inference to obtain the control function set, acting on the controlled object or process. The fuzzy algorithms of the control process are:
Define fuzzy sets and establish fuzzy control rules;
â—† Transformation from basic discourse domain to fuzzy set discourse;
â—† fuzzy relation matrix operation;
â—† Fuzzy inference synthesis, find the control output fuzzy subset;
â—† Inverse fuzzy operation, discrimination, get precise control The general structure of fuzzy control is as follows.
In fuzzy control, unlike the traditional control theory, there is a continuous function between 0 and 1. For example, the “drum water level is highâ€, the traditional control theory believes that if the value is higher than the set value, it will be high and the regulator will be integrated in the negative direction. In the fuzzy control theory, it is considered that there may be different degrees of height, for example, the measured value is depicted as: "PB is positive, PM is in the middle, PS is small, Z0 is zero, NS is negative, and NM is negative, NB is negative." This is the process of obfuscation. According to different conditions of the measured values, different conclusions are obtained through reasoning, that is, different outputs. It can be seen that although the traditional theory defines the process parameter as "accurate", that is, it has a definite value, it cannot describe the complicated process. Although fuzzy theory uses some fuzzy concepts, its description is closer to the actual situation of the process. This is to say that a certain number of genera does not belong to a certain range, and conventional mathematics is not "0", that is, "1". Fuzzy mathematics means that there are many intermediate values ​​and can be described by the "subordination degree" function. It actually describes the status of the control process more accurately.
The "knowledge base" in fuzzy control refers to the knowledge of the known control laws in the field of control, which consists of data and fuzzy language control rules. For example, IF (drum water level high deviation large) AND (long duration), THEN (water supply regulation gate low). If there are several fuzzy rules, you can write:
RlIFisA1ANDECisB1THENUisC1
R2IFisA2ANDECisB2THENUisC2
R1IFisAnANDECisBnTHENUisCn
Here, R fuzzy rules, E, EC state variables, U control output variables.
There are dozens of algorithms for fuzzy reasoning, and the direct reasoning is usually adopted by considering the membership degree of the membership function as a true value.
Inverse defuzzification is to use the weighted average judgment method for the membership function of the control quantity U obtained from the result of fuzzy inference.
The basic steps of fuzzy control system design:
â—† Determine the input variables and output variables.
â—† Design fuzzy control rules.
â—† Determine the fuzzification and defuzzification methods.
â—† Select the parameters for input and output variables.
â—† Develop fuzzy control program.
â—† Select the system sampling period.
Compared with the traditional PID control system, the biggest difference of the fuzzy control system is that the gain of the system is optimized according to the operating conditions of the control system, such as the proportional band, the conventional system is always a constant, and the fuzzy control system is based on the system. Calculated by operating conditions, with self-tuning.
1.2 expert intelligent control system (ExpertIntel—
ligentControlSystem)
The expert intelligent control system imitates human intelligence behavior, adopts the inference mechanism of the home system, integrates human intelligence, experience and expert systems, automatic control, and fuzzy technology to solve complex system problems, and makes the control system get a high level.
1.2.1 Expert System The expert system is a smart computer system with expert knowledge and experience in a certain field. That is to say, the expert system makes inferences and judgments based on the knowledge and experience provided by human experts and simulates the decision process of human experts. Solve the control problem of complex systems. Therefore, the main function of the expert system depends on a large amount of knowledge and knowledge expression and application. The difference between the expert system and the general control system is that the problem to be solved by the expert system generally has no algorithmic solution, and the conclusion needs to be obtained based on incomplete and inaccurate information.
The structure of the expert system is shown below.
â—† Expert knowledge is stored in the knowledge base and expressed in a machine-readable way. Such as logical representation, semantic network method, production rule, characteristic representation, frame representation, and or diagram, process representation, blackboard structure, neural network, Petri net method and so on. Different representations have their own advantages and disadvantages, which can meet the needs of different systems.
â—† The database is used to store inference engineering information and intermediate results.
â—† Inference engine is to store reasoning methods and reasoning.
The interpreter is the man/machine interface of the expert system.
â—† Knowledge acquisition is the interface of expert systems and experts. Expert knowledge is stored in the knowledge base through this interface.
Starting from the intelligence of the experts, according to the information of the process, according to certain principles and methods for machine thinking, it is reasoning. Obviously, in the expert system, except for expert knowledge and its representation, reasonable reasoning is very important. The theory on this aspect has seen new developments in the development and practice of leaps and bounds.
â—† According to the representation of knowledge, it can be divided into a graph search method and a logical argument method. According to the graph search method, starting from the existing "knots" of knowledge, according to a certain graphical problem form (such as state space maps, and or maps, etc.) to reach the target state node search method. Logical argumentation is equivalent to the process of theorem proving using mathematical logic.
â—† According to the method used in the solution process, it is divided into heuristic reasoning and non-heuristic reasoning.
â—† According to reasoning results, it is divided into precise reasoning and inexact reasoning. The conclusion of accurate reasoning is unique. The rules of non-exact reasoning, such as the probability theory method, credibility method, fuzzy subset method, etc., are inaccurate.
â—† The use of special and general relations based on inferential reasoning methods, deductive reasoning and inductive reasoning.
â—† There are forward reasoning and reverse reasoning based on the inferential thinking methods;
1.2.2 Expert Intelligence Control System The combination of expert system and control system is an expert intelligent control system. Obviously, the expert intelligent control system should have the following characteristics compared with the general expert system:
â—† The control system must be time-sensitive, that is, the response characteristics of the system must be considered.
â—† The system must have a good man/machine interface and can set and modify control knowledge and logic online.
â—† Ability to process non-linear processes at random. With interrupt capability.
â—† Interface with other control systems to exchange information.
The expert intelligent control system is divided into two categories: a direct expert intelligent control system and an indirect expert intelligent control system.
The knowledge-based expert system directly affects the controlled object and becomes a direct expert intelligent control system.
Knowledge-based expert intelligent control systems indirectly affect the control system, such as monitoring and control system parameters only (such as proportional band, etc.), called indirect expert intelligent control system.
The direct expert control system structure is shown in Figure 3.
The design of the direct expert control system and the expression of knowledge are usually based on the characteristics of industrial control and the requirements of real-time control. The causality of the process is described by the production rule, and the control rule set is established by the fuzzy control rule with the adjustment factor. As follows:
U=f(E,K,I)
In addition, f is a smart operator and its basic form is:
IFEANDKTHEN(IF0THENU)
E is the input information set for the controller, K is the empirical data and facts of the knowledge base, o is the output set of the inference mechanism, and U is the output of the control system.
The indirect expert control system controller includes both algorithms and logic, as shown in Figure 4.
The basic controllers such as PID and Fuzzy are configured in the lowest level controller. The expert system is used to coordinate all the algorithms. Based on the situation on the site, expert rules in the knowledge base are used to determine which algorithm and parameters to use.
1.3 Humanoid Hierarchical Hierarchical Intelligent Control System Humanoid Hierarchical Hierarchical Intelligent Control System consists of three levels: organization level, coordination level, and execution level. From the top down, intelligent decrement, accuracy increases, according to the control system deviation and deviation The level of rate is controlled hierarchically, and human intelligent control is used to achieve automatic, stable and optimized operation of process control. The structure is shown in Figure 5.
â—† Use expert control EC when the system deviation changes greatly.
â—† When the system deviation is slightly larger, fuzzy control FC is used.
â—† When the system deviation is small, use self-tuning PID control and self-optimization control.
1.4 Neural Network Intelligent Control System The theoretical basis of the neural network intelligent control system is through the research of neurophysiology and neuroanatomy, imitating human neuron information processing and transmission, and establishing an intelligent control system. Therefore, neural networks have the following characteristics:
â—† Distribution storage and fault tolerance. Each neuron stores information in accordance with the division of labor, and at the same time stores a variety of related information. Loss of part of the information can be recovered, with fault tolerance and association.
â—† plasticity and self-organization, self-organization. The communication modes among various neurons are diverse, and the ability to connect is plastic, which satisfies the burst of information transmission capabilities. This allows the network to self-organize through learning and training and adapt to different information processing requirements.
â—† Parallel processing, the entire network of information processing can be massively parallel processing, speed up the information processing speed.
â—† hierarchical, all kinds of information can be carried out at different levels of neural networks step by step.
Neural network intelligent control system has the following advantages:
â—† Has the ability to approach any non-linear function.
â—† high speed, high fault tolerance.
â—† Easy to use in multivariable control systems.
â—† Self-learning and self-adaptive features.
Neural network intelligent control system is an emerging field of intelligent control system, and it is developing rapidly from theory to practice. And with other intelligent control systems, a neural network adaptive control system, a fuzzy neural network intelligent control system, a neural network optimal control system and the like are derived.
The following is a block diagram of several typical neural network intelligent control systems.
â—† neural network adaptive control system (NNM-RAC) (Figure 6)
The NNC corrects according to the output error e = ym - y so that the error e is equal to zero.
â—†Neural network and conventional control combined control system (NNC)
The NNC is a compound control system consisting of a feedforward and a conventional controller. When the NNC is finished learning, it is equal to u2. The conventional control does not work and is controlled by the NNC.
â—† Neural Network Internal Model Control System (1MC)
The NN state estimation is proposed by the neural network. The controller can be provided by the NNC or by a conventional controller.
Neural network non-linear predictive control system (NPC) (Figure 9)
This algorithm uses a predictive model, rolling optimization, and feedback corrections to provide the output value extremum.
â—† Neural Network Expert Control System (NPC)
The neural network expert control system combines the expert system with the neural network system. The structure of Figure 10 can have three modes of operation: expert system operation (EC), neural system operation (NNC), and NNC and EC coordination. Combining the advantages of the two systems can achieve better results.
Neural network fuzzy control system The neural network system cannot process structured knowledge. It requires a large amount of training data. At the same time, the fuzzy system can directly deal with structured knowledge, and the learning mechanism of the Bible network can be introduced into the fuzzy system to make the fuzzy system self-learning and self-organizing. Adaptability and completion of fuzzy reasoning with parallel processing structures form a neural network fuzzy control system. The structure is shown in the figure below.
1.5 In conclusion, the intelligent control system shows superiority over conventional classical control systems both theoretically and practically, and it is fully compatible with conventional control systems, such as PID regulation systems. At the same time, for a complex control system, a system with a large lag, a non-linear control system, and a system that can not accurately express the function, can find a control method through a special method. Obviously, these features are of great use in complex process control, such as the control systems of large thermal power generating units.
2. Application Status of Intelligent Control System 2.1 General Situation Although the intelligent control system is not fully mature in theory, it is in rapid development in practice. In the thermal power plant control system, many suppliers also began to provide corresponding products and systems.
â—† The United States Bentley company put forward an expert system on the fault diagnosis of turbines more than ten years ago. This system can provide diagnosis results for only seven common accidents of turbine generator sets.
â—† In the TR-2000 series provided by Riss Inc. (Rochester Instrument System.), an expert system for substation monitoring can be provided.
â—† The information provided by Siemens of Germany shows that their fuzzy controllers have been applied in environmental protection power plants. The basic structure adopted is as follows. In the above control scheme, the basic idea is to keep the conventional control methods as much as possible, and the fuzzy control speeds up the response characteristics of the system. Using conventional systems to eliminate system deviations.
As shown in the figure, the control system includes two loops in parallel, the fuzzy control output corrects the conventional PI control, and the input of the fuzzy control is the control deviation of the controlled system and the differential of this deviation. Fuzzy control includes expert system and fuzzy control.
This system has been applied to BCD boiler FGD furnace pressure closed loop control and chemical water treatment PH control system.
â—† Other manufacturers, such as ABB Bailey and MAX, have also introduced their own ideas and plans for intelligent control.
2.2 In the control of power plants in China, there are many pioneers who are doing pioneering work. Intelligent control strategies are adopted in local control systems of power stations, and some have achieved good results. Examples are as follows:
â—† A military unit adopts fuzzy control to improve the control of the drum water level.
This control system uses the P-Fuzzy-PI scheme.
Different control methods are used in different fields.
When the deviation is greater than or equal to a certain value, proportional control is used to increase the response speed of the system.
When the deviation is less than this value, fuzzy control is used to reduce system overshoot.
When close to the set value, PI adjustment is adopted to eliminate system deviation.
After this scheme is adopted, the system response characteristics and adjustment quality are improved, as shown in Fig. 13.
In the figure above, curve a is PID adjustment, curve b is Fuzzy adjustment, and curve c is P-Fuzzy-PI adjustment.
â—† In order to optimize the combustion control system, a power plant adopts a fuzzy control system, which increases the boiler efficiency by 2.8%.
The power plant is a chain furnace with a large lag and poor control characteristics. During the renovation, a humanoid intelligent fuzzy controller was used to realize the negative pressure control of the combustion and the furnace.
The basic design scheme adopts three stages of coarse adjustment, fine adjustment and fine adjustment, and 5 fuzzy controllers. The fuzzy controllers Ao and A are coarse adjustments. According to the thermal load and the calorific value of the coal, the thickness of the coal seam and the grate rotation speed are calculated. Then, according to the optimized deviation value, the grate change amount and the blower variation amount are output. The fuzzy controller B is fine-tuning, optimizes the coarse adjustment result, and makes the system have a memory function and a learning function. The fuzzy controller C finely adjusts the wind/coal ratio based on the above control. Fuzzy controller D completes the negative pressure control of the furnace.
â—† For the control of variable frequency motors that are increasingly widely used in thermal power plants, some supply companies use fuzzy intelligent control instead of PID traditional control systems to improve their control characteristics. Figure 14 shows the effects of the two control features A: Intelligent controller response characteristics.
B: PID control response characteristics.
◆ AI series instruments. The AI ​​series instruments developed by Xiamen Yuguang Electronic Technology Research Institute add fuzzy control algorithm rules to the PID function. When the deviation is large, fuzzy logic is used to determine the output, just as a skilled worker performs manual operations. When the deviation becomes smaller, the improved PID algorithm output is used. The system features no overshoot and high control accuracy. For non-linear complex adjustments, adaptive adjustment rules are provided to improve the quality of adjustment. A self-tuning expert system is set in the table to save debugging time and workload.
2.3 Many examples of development and application of intelligent control systems can be cited at home and abroad. However, only some of the above-mentioned examples can indicate that intelligent control has already started in the field of power plant control. The ability of intelligent control in solving complex systems is being confirmed. Intelligent control systems improve the quality of automated systems and are recognized by many people. The application areas of intelligent systems are advancing both in depth and breadth.
However, on the other hand, in the field of thermal power plant control, smart systems are just starting.
Although there are many examples of successful applications, it is still rare compared to the widespread adoption of classical control systems and it still does not constitute any climate.
Although many manufacturers introduced their intelligent system control strategy, no manufacturer has promised to adopt a smart algorithm DCS system on a large scale.
In the large-scale project construction of thermal power plants, few owners require intelligent control systems.
In a word, although the intelligent control system has already started in the field of control of thermal power plants, the road to go is still very long. There are still many problems to be solved.
3. Several problems with the use of intelligent controls in thermal power plants.
3.1 The use of intelligent controls in thermal power plant control systems will at least help solve the following problems:
â—† Help solve the problems in the control system.
Among the control systems of thermal power plants, some systems are difficult to put into automation due to a variety of reasons, and they become the most difficult problems. Such as steel ball mill load adjustment. The use of intelligent control systems can solve this problem by providing another way to properly handle this issue. Another example is the fact that combustion regulation is often difficult to invest, and in many domestic power plants, the use of smart regulation has set a precedent for solving this problem.
â—† Improve the control system.
One of the weaknesses of the traditional control system is that it can only be operated in accordance with the set control law and parameters. In fact, the operating conditions are very complicated. Even if many correction loops are added, the system cannot be guaranteed to operate in the best state. For example, boiler excess air coefficient. The control, the traditional control system is to refer to the boiler manufacturer data, empirical data, and in accordance with the thermal test results, to determine a given value. In order to adapt to the conditions under different loads, the above given value is added with "load" correction. In fact. It is a very important parameter affecting the economic operation of the boiler. It is related to whether the boiler is operating at a suitable wind/coal ratio. It is related to whether the combustion is complete and whether the air supply is excessive. That is to say, it is the most important issue that affects the economic operation of the boiler. The index q2 (exhaust loss), q3, (chemical incomplete combustion loss), q4 (mechanical incomplete combustion loss) values. At the same time, the reasonable value of α is far from being determined by theoretical calculations and thermal tests, and it is not only related to the load. For example, it is related to the type of coal, coal, calorific value, moisture, ash, etc. It is also related to boiler conditions such as air leakage, cleanliness of the heat transfer surface, and so on. Due to the existence of many complex factors, the conventional regulation system, even if the combustion regulation system can be put into automatic, does not mean that the system is operating in an ideal state.
The use of intelligent control systems can improve the functionality of this system.
Intelligent control strategy to solve this problem, not directly determine the value of α, but to determine the relationship between α and boiler efficiency, and affect the dynamic relationship between α factors, and through the self-learning adaptive system to achieve the desired state. Therefore, α is not a definite, fixed value here, but is a dynamic data that adapts to various complex situations and achieves the best numerical value of boiler efficiency.
Similar systems in power plant control systems also have, in particular, optimized PID parameters that allow them to operate at a truly reasonable value. The use of intelligent systems is an option.
â—† Improve the level of control.
The need to continue to improve the control level of thermal power plants is a consensus of all parties. Regardless of the size of large-scale units or small and medium-sized units, the number of operating personnel is decreasing. This requires that the control logic, especially the design of protection logic, must include as many complicated situations as possible. This also has difficulties for conventional control systems.
For example, for the unit-level self-starting logic design, due to the complexity of the situation, although many people have made unremitting efforts over the years, there are many achievements. However, with the conventional control system, it is still necessary to achieve the goal of achieving unit-level program completion from the completion. A small distance.
If the intelligent control system is used to implement the unit-level program startup, many uncertain factors can be taken into account in the logic of the control system; the impact of many sudden factors on the startup process can be handled; many operational experiences can be programmed into the control system logic; Through self-learning, self-adaptive and other functions, the system has been continuously improved through operation. With these advantages, there is a new way to solve this problem.
Another example is the increase in the level of control of thermal power plants and the requirement to reduce operating personnel. The most prominent issue in this regard is the automation of auxiliary workshops. Auxiliary workshops can be equipped with a program control system. However, changes in the way of operation of the workshop, handling of accidents, and switching of emergency operations are often inseparable from human intervention. Using intelligent control system, a large amount of logic analysis enters the control system. The judgment of accidents, the enhancement of logical reasoning function, and the adoption of humanoid intelligence provide possibilities for solving the unattended design of the auxiliary workshop.
Therefore, to improve the control level, there must be new ideas. The intelligent control system is an optional path.
â—† Develop new control features.
The first problem in this area is the equipment fault diagnosis.
We know that the fault diagnosis of turbogenerators, such as Bentley, can already provide expert system software. Similar software for boilers and power plant auxiliary systems is also under development. In the nuclear power plant control system, software similar to the intelligent system used for fault diagnosis and alarm of the equipment is being developed.
The intelligent system was used to diagnose the fault of the equipment, and the reasoning judgments of some operators were changed to the intelligent control system to speed up the time for fault diagnosis, and it was possible to judge the health status of the unit. This is of great significance for improving the availability of equipment.
In addition, the auxiliary engine operation control design of the power plant now basically stays in the scope of the protection interlock logic. As for the principle of the CCS auxiliary operation mode management, it is only designed for the unit's safe operation requirements. However, the mode of operation of auxiliary engines has a great bearing on the economic efficiency of power plants. For example, the operation of a coal mill, the operation of a feed pump, etc., and the operation of several units under what load, etc., are now mostly determined by humans. With the introduction of intelligent control systems, it is possible to calculate and judge the rationality and economic efficiency of the auxiliary machine operation mode, and improve the economic operation level of the power plant.
Secondly, the introduction of intelligent systems in the operation and management of power plants, such as technical management, overhaul management, and operation and management of units, will also increase the level of management.
3.2 Some issues to be taken into account when adopting intelligent systems In summary, the use of intelligent control systems in thermal power plants can accomplish a lot. Whether it is a new-build project or the technological transformation of existing power plants, the deployment of intelligent systems will bring benefits. .
However, for an enterprise, any new technology adoption must pursue economic benefits. Generally, it cannot allow failure, especially big failure. This is fundamentally different from scientific research. The intelligent control system is completely a new thing, so what issues should be noticed here? I think of the following points for reference.
First, the use of intelligent control systems requires special attention to software development.
For different objects, different systems, choose the appropriate type of intelligent system, algorithm and parameters in it. For this reason, designers, end users (operators) should be more involved in the system's software design process to reduce risk.
Second, use relatively mature systems and software as much as possible.
As the intelligent control system is growing, for the control system of the thermal power plant, information on the complete set of software for the entire power plant control system has not yet been seen. Some suppliers promise to provide some local software. In addition to what has been mentioned above, there are ABB-BAILEY companies that can provide some software. However, due to the rapid changes in the situation, the company concerned should be consulted.
The software provided by suppliers, especially those developed for power plants, is expensive, but it is less risky and practical.
It should be pointed out that for some systems that are not very complicated, it is entirely possible to develop application software on their own.
Third, we must consider combining with the original system.
The intelligent control system has advantages in logic judgment, control mode selection, control parameters, etc. However, no difference adjustment is realized, and traditional PI control is very convenient. The use of intelligent control systems should not lose these advantages.
Logical reasoning of intelligent control systems has its own characteristics and advantages, but its foundation cannot leave the knowledge of the process knowing. On the contrary, it is on this basis that the control system logic is established. Therefore, making full use of the original control system that includes known operating laws and knowledge and making an interface with the intelligent control system is an important condition for the success of the project.
Some vendors provide systems that are parallel to conventional systems and intelligent systems. In addition to division of labor, there are also redundant considerations. In the current situation, it is also a risk reduction measure.
Of course, there are still many problems involved in the design of intelligent control systems, such as the selection of hardware, the calculation of benefits, etc. I am also researching.
3.3 Newly built power plant intelligent control system configuration assumptions Here, the idea of ​​configuring intelligent control parts for the new thermal power plant control system is proposed. Of course, this idea is not mature and is only for the reference of colleagues in similar work. I think that if other conditions are equal, manufacturers that can provide intelligent control systems should have advantages because intelligent control systems can bring economic benefits to the company.
â—† CCS section Generally speaking, the PID operation part can be intelligent, even for a base-type adjustment system or an electric base transmitter, or an auxiliary workshop adjustment system, can be intelligent. Intelligent PID regulation can at least improve the quality of the adjustment, and it will not result in a large increase in investment for new projects.
For a conventional method, a difficult adjustment system, or a system with poor results, should be intelligent, such as a combustion regulation system, an air supply regulation system, a pulverizing system regulation, a main steam and a reheat steam regulation system, etc. Intelligentization will increase the automation investment rate, improve the safety and economic operation level, and bring considerable economic benefits to the company.
For the complete adjustment of complex systems, such as water supply regulation, including the switching logic of the steam pump, electric pump, and low-load control valve, should be intelligent. Especially for very large units, the control logic is complex and intelligent is an important way to solve the problem.
Of course, accurately determining the scope of intelligence needs to be determined according to the actual conditions in various aspects of project implementation.
â—† DAS The main problems existing in the current DAS function are the performance calculation and optimization of economic operation modes; the cause of the accident and the corresponding operation guidance. In theory, intelligent control systems can be introduced in both of these areas, and the necessary judgment systems can be established to provide necessary decision-making reference materials for enterprises. It is also possible to seek ways to completely solve these problems. Unfortunately, no information has yet been provided on providing mature software vendors in this area.
â—† Fault Diagnosis of Turbine Generator Turbogenerator fault diagnosis expert systems are available from suppliers. However, it is rarely applied to practical projects. Needless to say, it is not a simple matter for existing turbogenerator fault diagnosis software to make it play its due role. This includes the input of relevant materials; the configuration of professional and technical management personnel and so on. Especially the latter problem is not easy for a power plant. Because of the limited number of generator sets in a power plant, experts in fault diagnosis for specialized turbine generator sets usually have very few things to do, and experts who do not have the ability to use smart systems at the time of the key are not. To solve this problem, we should give full play to the advantages of computer networks and establish regional fault diagnosis centers so that we can give full play to the role of intelligent systems and experts.
The boiler's fault diagnosis system is still under development, and no systematic product appears. This is because the complexity of the problem makes it difficult to describe complex processes using classical mathematical models. With the development of intelligent control, it is possible to adopt new methods to achieve breakthroughs. For example, in the determination logic of the MFT action condition, the first-out cause determination may be added. For other system protection actions, there should also be first-out reason determination. This will provide the operator with important operating instructions. It should be mentioned that in the nuclear power plant's control system, someone uses the principle of neuron system control to develop similar software. Its purpose is to assist operators to quickly determine the real cause of the accident.
â—† The SCS system SCS system is currently lack of optimized operation mode to improve the safety and economic level of auxiliary equipment operation. Intelligent control system can make a difference in this respect, but at present there is still a blank space.
◆辅助车间控制现在辅助车间多采用PLC控制，控制的范围一般包括程åºå¯åŠ¨ã€åœæ¢ï¼Œå±€éƒ¨å‚数调节，事故时åœæ¢æŠ¥è¦ï¼Œä»¥åŠå’ŒDCS通讯ç‰ã€‚没有包括设备的è¿è¡Œæ–¹å¼ä¼˜åŒ–，è¿è¡Œæ–¹å¼è‡ªåŠ¨çš„自动切æ¢ï¼Œç‰¹åˆ«åœ¨äº‹æ•…时自动切æ¢ã€‚如果就地ä¸å†é…ç½®è¿è¡Œäººå‘˜ï¼Œè¿è¡Œäººå‘˜åŽŸæ¥å¤„ç†çš„许多æ“作，特别事故处ç†çš„æ“作，完全采用精确数å¦æ¨¡åž‹å¤„ç†ï¼Œå°±ä¼šæœ‰è¾ƒå¤§çš„困难。而采用智能系统，就å¯ä»¥æ¯”ä¼ ç»Ÿçš„ç³»ç»Ÿå¤„ç†æ›´å¤šçš„问题。
4.åŠ å¿«ç«åŠ›å‘电厂控制系统智能化建议4．1应该把ç«åŠ›å‘电厂控制系统智能化工作æ到日程上æ¥ã€‚
应该如何对待ç«åŠ›å‘电厂控制系统智能化这一问题呢?温故而知新，在这里我们ä¸å¦¨ç®€å•å›žå¿†ä¸€ä¸‹æœ€è¿‘一个时期我国ç«åŠ›å‘电厂控制系统å‘展的历å²ã€‚
上个世纪七ã€å…«å年代，我们çƒè¡·äºŽå¼€å‘DDZâ…¡ã€â…¢é‡åž‹ä»ªè¡¨æ—¶ï¼Œè¿™ç±»ä»ªè¡¨åœ¨å›½å¤–æ£åœ¨èµ°å‘åŽæœŸï¼Œè®¡ç®—机控制系统已ç»å…´èµ·ã€‚当我们DDZ系列仪表刚刚趋å‘æˆç†Ÿï¼Œä¹Ÿåˆ°äº†æ·˜æ±°çš„边沿。DCS分散控制系统国内产å“，在市场还未æ¥å¾—åŠéœ²é¢ï¼Œå›½å¤–厂商已ç»è¿…速å 领了ä¸å›½å¸‚åœºã€‚ä¸ºä»€ä¹ˆä¼šè¿™æ ·?为æ¤å¯ä»¥è®²å‡ºè®¸è®¸å¤šå¤šçš„åŽŸå› ï¼Œä½†æ˜¯ä»Žæ ¹æœ¬ä¸Šè®²ï¼Œæ˜¯å¯¹æŽ§åˆ¶ç³»ç»ŸæŠ€æœ¯å‘展方å‘和速度估计ä¸è¶³ã€‚ç›®å‰ï¼ŒæŽ§åˆ¶ç³»ç»ŸæŠ€æœ¯çš„å‘展速度并没有å‡ç¼“，对å‘电厂控制系统å‘展è¦æ±‚ä¾ç„¶å˜åœ¨ï¼Œå› æ¤ï¼ŒæŠŠæ¡ä½ä¸»è¦æŠ€æœ¯å‘展方å‘ï¼ŒåŠ å¿«å‘展速度，在相当长一段时间内都是很é‡è¦çš„。——应该从这个角度æ¥çœ‹å¾…控制系统智能化的问题。
æŽ§åˆ¶ç³»ç»Ÿæ™ºèƒ½åŒ–ï¼Œæ— è®ºä»Žç†è®ºä¸Šå’Œå®žè·µä¸Šéƒ½å·²ç»å…·å¤‡æ¡ä»¶ã€‚但是对æ¤é—®é¢˜çš„认识，在行业内还很ä¸ä¸€è‡´ã€‚相当多的åŒè¡Œå¯¹å®ƒçš„实用价值，还æŒæœ‰æ€€ç–‘æ€åº¦ã€‚把这方é¢çš„情况，资料，åŠæ—¶çš„介ç»ç»™ä¸šå†…人士，共åŒåŠªåŠ›ï¼Œæ˜¯åˆ‡ä¸å®¹ç¼“的事情。
æ¤å¤„还应特别æ出，在智能控制ç†è®ºç ”究上，ä¸å›½å¹¶ä¸è½åŽï¼Œåº”用技术也有一定的基础，问题是统一认识，认真去作，完全有å¯èƒ½åœ¨æ–°ä¸€è½®æŠ€æœ¯ç«žäº‰å æ®æœ‰åˆ©åœ°ä½ã€‚
4．2åŠ å¼ºæŠ€æœ¯äº¤æµï¼Œä¿ƒè¿›æ™ºèƒ½æŽ§åˆ¶æŠ€æœ¯çš„å‘展。
作为一ç§å‰æ²¿å¦ç§‘，智能控制系统在å„æ–¹é¢éƒ½ååˆ†æ´»è·ƒã€‚ç§‘ç ”å•ä½ã€å¤§ä¸“é™¢æ ¡ã€è®¾è®¡é™¢ã€å»ºè®¾å•ä½å’Œç”Ÿäº§å•ä½ï¼Œéƒ½æœ‰äººåœ¨ä»Žäº‹ç†è®ºæˆ–实践工作。但是，目å‰é™¤åŽ»å›½å¤–供货厂商为了推销他们的产å“介ç»ç«åŠ›å‘电厂智能控制技术外，很少能è§åˆ°è¿™æ–¹é¢çš„技术交æµã€‚å»ºè®®åŠ å¼ºæ¤é¡¹å·¥ä½œï¼Œè‚¯å®šæœ‰åˆ©äºŽæ™ºèƒ½æŽ§åˆ¶æŠ€æœ¯çš„å‘展。
4．3充分利用国外æˆæžœç›®å‰æˆ‘国DCSä¾›è´§æ ¼å±€ä»ç„¶ä»¥å›½å¤–系统为主（å«åˆèµ„ä¼ä¸šï¼‰ï¼Œç‰¹åˆ«æ˜¯å¯¹äºŽIï¼O规模超过3000点以上的大型系统更是如æ¤ã€‚既然花钱买技术，就应该使其物有所值。ä¸å¦¨åœ¨æˆ‘ä»¬çš„æ‹›æ ‡æ–‡ä»¶ï¼ŒæŠ€æœ¯è§„èŒƒä¸æ˜Žç¡®æ出è¦æ±‚对方æä¾›æˆç†Ÿçš„æ™ºèƒ½æŽ§åˆ¶æŠ€æœ¯ã€‚è¿™æ ·ï¼Œæˆ‘ä»¬ä¸ç‰å¾…国外形æˆå®Œæ•´ç³»åˆ—产å“å‰ï¼Œåˆ©ç”¨å…¶å·²æˆç†Ÿçš„éƒ¨åˆ†ï¼ŒåŠ ä¸Šæˆ‘ä»¬çš„å¼€å‘工作，有å¯èƒ½åœ¨æ™ºèƒ½æŽ§åˆ¶ç³»ç»Ÿå¼€å‘ç«žèµ›ä¸å¤„于较有利ä½ç½®ã€‚
4．4制定å‘展规划，适时进行总结å¤äººæ—¥ï¼šâ€œå‡¡äº‹é¢„则立â€ï¼ŒæŒ‡åŠžä»»ä½•ä¸€ä»¶äº‹æƒ…åªæœ‰æœ‰è®¡åˆ’æ‰èƒ½æˆåŠŸã€‚å‘电厂控制系统智能化，虽然é¢ä¸´å¤§å¥½æœºé‡ï¼Œå¹¶ä¸ç‰äºŽè¿™ä»¶äº‹å°±ä¼šè‡ªç„¶è€Œç„¶é¡ºåˆ©å‘展。电力行业应和仪表行业è”åˆèµ·æ¥ï¼ŒæŠ“ä½æœºé‡ï¼Œå°½æ—©ç¡®ç«‹æˆ‘国在æ¤é¢†åŸŸå†…应å 有的ä½ç½®ã€‚
å‘电厂控制系统智能化å‘展规划应该包括在å„个方é¢æŠ€æœ¯è¦æ±‚；未æ¥å‘电厂控制系统智能化基本规范；å‚åŠ å•ä½ã€ä¸ªäººå¤§ä½“分工和æƒç›Šï¼›ä»¥åŠå‘展里程表ç‰ç‰ã€‚æœ‰äº†è¿™æ ·ä¸€ä¸ªè§„åˆ’ï¼Œå°±ä¾¿äºŽè°ƒåŠ¨å„æ–¹é¢çš„积æžæ€§ï¼Œå°±æœ‰å¯èƒ½ä½¿ä¸ªåˆ«å•ä½çš„努力汇åˆèµ·æ¥ï¼Œå½¢æˆæˆ‘们的å‘展优势。
当然，仅仅有规划是ä¸å¤Ÿçš„。还è¦åŠ 上定期的把å–å¾—çš„æˆæžœè¿›è¡Œæ€»ç»“，归纳，形æˆç³»åˆ—ã€è§„范，æ‰èƒ½å·©å›ºå’Œå‘å±•ç ”ç©¶æˆæžœã€‚
Intelligent control system can be said to have emerged. The reason is:
First, users need this technology. Modern thermal power generating units are developing in the direction of large-scale and super-large ones. The domestic 600 MW unit has become the main unit and the 900 MW unit is under construction. With the increase in the size of power grids and the development of electricity in industry, agriculture, commerce, and residential life, with the development of power generation equipment manufacturing technology, larger generator sets will also emerge. The large-capacity generator sets must adopt supercritical parameters, as well as the need for complicated environmental protection technologies. All these factors cause the control systems of thermal power plants to become very large and complex. Although the functions of computers have become more and more powerful, the development of application software has increased the workload by many times. System debugging has become more and more difficult in terms of workload or difficulty. The intelligent control system has its unparalleled advantages precisely in these aspects.
Second, the current power plant construction and planned economic management models are very different. We hope that shortening the construction period and pursuing rapid capital returns will be the common mode of operation of investors. Therefore, a new project, preparation work, feasibility study, tendering, calibration, etc., take a lot of time to reduce the risk of investment. At the same time, once construction begins, a large amount of funds will be invested. Investors hope that the shorter the construction period, the better, and that they can achieve the goal of recovering funds as soon as possible. In engineering odometers, the time for commissioning the control system is often very limited and cannot be guaranteed. For complex control systems, the control logic needs to be tested under various operating conditions and operating modes. If the required debugging time is longer, the debugging work is even more difficult, and even the system cannot be put in time. Can we debug while running? For conventional control systems based on precise mathematics, this is very difficult, but for intelligent control systems, it is its characteristics. It is also its advantages.
Third, the classical control system based on PID law requires an accurate mathematical model when designing, and requires accurate parameters when debugging. Obviously, there are difficulties for this complex system. Intelligent control systems are precisely designed to solve these problems.
Fourth, because of its advanced theory, the intelligent control system does not require the hardware of the computer compared to the classic control system. However, in the long run, it is easy to program, and the commissioning workload is reduced. It is very likely to reduce the investment in the control system. And shorten the construction period.
Fifth, the development of intelligent control systems requires the cooperation of computer technology development. Needless to say, the rapid development of computer technology has provided a vast world for intelligent control systems.
The intelligent control system came into being and has a broad prospect in the field of power plant control. So what is an intelligent control system? What is its current status? What strategy should be adopted in the intelligentization of power plant control systems? This is the issue discussed in this paper.
1. Intelligent Control System Introduction Intelligent control system as a frontier discipline in the field of automatic control, it is difficult to make a complete definition, because there are new theories and new systems appearing at almost any time here. In general, intelligent control is the use of artificial intelligence intuition reasoning methods for self-learning control systems, which can be summarized as a combination of automatic control and artificial intelligence. For example, in a power plant control system, the operational experience of the operating personnel is combined with the control system, and the level of intelligence of the system can be enriched through self-learning. In intelligent control systems, the higher the level of intelligent decision making, the more artificial artificial intelligence is used.
Intelligent control has the following basic features:
â—† The higher the level of control system, the higher the intelligence embodied.
â—†Open-loop and closed-loop control combined.
â—† The combination of qualitative control and quantitative control, that is, the combination of control logic and logic execution process (such as speed, etc.).
â—† Intelligent control has self-learning, self-adaptation and organization functions.
â—† Information processing methods include both mathematical and logical reasoning.
Intelligent control currently has approximately the following categories:
â—† Fuzzy control system based on fuzzy mathematics.
â—† Expert control system.
â—† neuron control system.
â—† Integrated intelligent control system combining multiple methods.
1.1 Fuzzy Control System (FuzzycOntr01sys-
The basic idea of ​​fuzzy control is to generalize the control strategy of the person on the controlled object as follows: IF (condition) THEN (action) forms the control rules of the form, through the fuzzy inference to obtain the control function set, acting on the controlled object or process. The fuzzy algorithms of the control process are:
Define fuzzy sets and establish fuzzy control rules;
â—† Transformation from basic discourse domain to fuzzy set discourse;
â—† fuzzy relation matrix operation;
â—† Fuzzy inference synthesis, find the control output fuzzy subset;
â—† Inverse fuzzy operation, discrimination, get precise control The general structure of fuzzy control is as follows.
In fuzzy control, unlike the traditional control theory, there is a continuous function between 0 and 1. For example, the “drum water level is highâ€, the traditional control theory believes that if the value is higher than the set value, it will be high and the regulator will be integrated in the negative direction. In the fuzzy control theory, it is considered that there may be different degrees of height, for example, the measured value is depicted as: "PB is positive, PM is in the middle, PS is small, Z0 is zero, NS is negative, and NM is negative, NB is negative." This is the process of obfuscation. According to different conditions of the measured values, different conclusions are obtained through reasoning, that is, different outputs. It can be seen that although the traditional theory defines the process parameter as "accurate", that is, it has a definite value, it cannot describe the complicated process. Although fuzzy theory uses some fuzzy concepts, its description is closer to the actual situation of the process. This is to say that a certain number of genera does not belong to a certain range, and conventional mathematics is not "0", that is, "1". Fuzzy mathematics means that there are many intermediate values ​​and can be described by the "subordination degree" function. It actually describes the status of the control process more accurately.
The "knowledge base" in fuzzy control refers to the knowledge of the known control laws in the field of control, which consists of data and fuzzy language control rules. For example, IF (drum water level high deviation large) AND (long duration), THEN (water supply regulation gate low). If there are several fuzzy rules, you can write:
RlIFisA1ANDECisB1THENUisC1
R2IFisA2ANDECisB2THENUisC2
R1IFisAnANDECisBnTHENUisCn
Here, R fuzzy rules, E, EC state variables, U control output variables.
There are dozens of algorithms for fuzzy reasoning, and the direct reasoning is usually adopted by considering the membership degree of the membership function as a true value.
Inverse defuzzification is to use the weighted average judgment method for the membership function of the control quantity U obtained from the result of fuzzy inference.
The basic steps of fuzzy control system design:
â—† Determine the input variables and output variables.
â—† Design fuzzy control rules.
â—† Determine the fuzzification and defuzzification methods.
â—† Select the parameters for input and output variables.
â—† Develop fuzzy control program.
â—† Select the system sampling period.
Compared with the traditional PID control system, the biggest difference of the fuzzy control system is that the gain of the system is optimized according to the operating conditions of the control system, such as the proportional band, the conventional system is always a constant, and the fuzzy control system is based on the system. Calculated by operating conditions, with self-tuning.
1.2 expert intelligent control system (ExpertIntel—
ligentControlSystem)
The expert intelligent control system imitates human intelligence behavior, adopts the inference mechanism of the home system, integrates human intelligence, experience and expert systems, automatic control, and fuzzy technology to solve complex system problems, and makes the control system get a high level.
1.2.1 Expert System The expert system is a smart computer system with expert knowledge and experience in a certain field. That is to say, the expert system makes inferences and judgments based on the knowledge and experience provided by human experts and simulates the decision process of human experts. Solve the control problem of complex systems. Therefore, the main function of the expert system depends on a large amount of knowledge and knowledge expression and application. The difference between the expert system and the general control system is that the problem to be solved by the expert system generally has no algorithmic solution, and the conclusion needs to be obtained based on incomplete and inaccurate information.
The structure of the expert system is shown below.
â—† Expert knowledge is stored in the knowledge base and expressed in a machine-readable way. Such as logical representation, semantic network method, production rule, characteristic representation, frame representation, and or diagram, process representation, blackboard structure, neural network, Petri net method and so on. Different representations have their own advantages and disadvantages, which can meet the needs of different systems.
â—† The database is used to store inference engineering information and intermediate results.
â—† Inference engine is to store reasoning methods and reasoning.
The interpreter is the man/machine interface of the expert system.
â—† Knowledge acquisition is the interface of expert systems and experts. Expert knowledge is stored in the knowledge base through this interface.
Starting from the intelligence of the experts, according to the information of the process, according to certain principles and methods for machine thinking, it is reasoning. Obviously, in the expert system, except for expert knowledge and its representation, reasonable reasoning is very important. The theory on this aspect has seen new developments in the development and practice of leaps and bounds.
â—† According to the representation of knowledge, it can be divided into a graph search method and a logical argument method. According to the graph search method, starting from the existing "knots" of knowledge, according to a certain graphical problem form (such as state space maps, and or maps, etc.) to reach the target state node search method. Logical argumentation is equivalent to the process of theorem proving using mathematical logic.
â—† According to the method used in the solution process, it is divided into heuristic reasoning and non-heuristic reasoning.
â—† According to reasoning results, it is divided into precise reasoning and inexact reasoning. The conclusion of accurate reasoning is unique. The rules of non-exact reasoning, such as the probability theory method, credibility method, fuzzy subset method, etc., are inaccurate.
â—† The use of special and general relations based on inferential reasoning methods, deductive reasoning and inductive reasoning.
â—† There are forward reasoning and reverse reasoning based on the inferential thinking methods;
1.2.2 Expert Intelligence Control System The combination of expert system and control system is an expert intelligent control system. Obviously, the expert intelligent control system should have the following characteristics compared with the general expert system:
â—† The control system must be time-sensitive, that is, the response characteristics of the system must be considered.
â—† The system must have a good man/machine interface and can set and modify control knowledge and logic online.
â—† Ability to process non-linear processes at random. With interrupt capability.
â—† Interface with other control systems to exchange information.
The expert intelligent control system is divided into two categories: a direct expert intelligent control system and an indirect expert intelligent control system.
The knowledge-based expert system directly affects the controlled object and becomes a direct expert intelligent control system.
Knowledge-based expert intelligent control systems indirectly affect the control system, such as monitoring and control system parameters only (such as proportional band, etc.), called indirect expert intelligent control system.
The direct expert control system structure is shown in Figure 3.
The design of the direct expert control system and the expression of knowledge are usually based on the characteristics of industrial control and the requirements of real-time control. The causality of the process is described by the production rule, and the control rule set is established by the fuzzy control rule with the adjustment factor. As follows:
U=f(E,K,I)
In addition, f is a smart operator and its basic form is:
IFEANDKTHEN(IF0THENU)
E is the input information set for the controller, K is the empirical data and facts of the knowledge base, o is the output set of the inference mechanism, and U is the output of the control system.
The indirect expert control system controller includes both algorithms and logic, as shown in Figure 4.
The basic controllers such as PID and Fuzzy are configured in the lowest level controller. The expert system is used to coordinate all the algorithms. Based on the situation on the site, expert rules in the knowledge base are used to determine which algorithm and parameters to use.
1.3 Humanoid Hierarchical Hierarchical Intelligent Control System Humanoid Hierarchical Hierarchical Intelligent Control System consists of three levels: organization level, coordination level, and execution level. From the top down, intelligent decrement, accuracy increases, according to the control system deviation and deviation The level of rate is controlled hierarchically, and human intelligent control is used to achieve automatic, stable and optimized operation of process control. The structure is shown in Figure 5.
â—† Use expert control EC when the system deviation changes greatly.
â—† When the system deviation is slightly larger, fuzzy control FC is used.
â—† When the system deviation is small, use self-tuning PID control and self-optimization control.
1.4 Neural Network Intelligent Control System The theoretical basis of the neural network intelligent control system is through the research of neurophysiology and neuroanatomy, imitating human neuron information processing and transmission, and establishing an intelligent control system. Therefore, neural networks have the following characteristics:
â—† Distribution storage and fault tolerance. Each neuron stores information in accordance with the division of labor, and at the same time stores a variety of related information. Loss of part of the information can be recovered, with fault tolerance and association.
â—† plasticity and self-organization, self-organization. The communication modes among various neurons are diverse, and the ability to connect is plastic, which satisfies the burst of information transmission capabilities. This allows the network to self-organize through learning and training and adapt to different information processing requirements.
â—† Parallel processing, the entire network of information processing can be massively parallel processing, speed up the information processing speed.
â—† hierarchical, all kinds of information can be carried out at different levels of neural networks step by step.
Neural network intelligent control system has the following advantages:
â—† Has the ability to approach any non-linear function.
â—† high speed, high fault tolerance.
â—† Easy to use in multivariable control systems.
â—† Self-learning and self-adaptive features.
Neural network intelligent control system is an emerging field of intelligent control system, and it is developing rapidly from theory to practice. And with other intelligent control systems, a neural network adaptive control system, a fuzzy neural network intelligent control system, a neural network optimal control system and the like are derived.
The following is a block diagram of several typical neural network intelligent control systems.
â—† neural network adaptive control system (NNM-RAC) (Figure 6)
The NNC corrects according to the output error e = ym - y so that the error e is equal to zero.
â—†Neural network and conventional control combined control system (NNC)
The NNC is a compound control system consisting of a feedforward and a conventional controller. When the NNC is finished learning, it is equal to u2. The conventional control does not work and is controlled by the NNC.
â—† Neural Network Internal Model Control System (1MC)
The NN state estimation is proposed by the neural network. The controller can be provided by the NNC or by a conventional controller.
Neural network non-linear predictive control system (NPC) (Figure 9)
This algorithm uses a predictive model, rolling optimization, and feedback corrections to provide the output value extremum.
â—† Neural Network Expert Control System (NPC)
The neural network expert control system combines the expert system with the neural network system. The structure of Figure 10 can have three modes of operation: expert system operation (EC), neural system operation (NNC), and NNC and EC coordination. Combining the advantages of the two systems can achieve better results.
Neural network fuzzy control system The neural network system cannot process structured knowledge. It requires a large amount of training data. At the same time, the fuzzy system can directly deal with structured knowledge, and the learning mechanism of the Bible network can be introduced into the fuzzy system to make the fuzzy system self-learning and self-organizing. Adaptability and completion of fuzzy reasoning with parallel processing structures form a neural network fuzzy control system. The structure is shown in the figure below.
1.5 In conclusion, the intelligent control system shows superiority over conventional classical control systems both theoretically and practically, and it is fully compatible with conventional control systems, such as PID regulation systems. At the same time, for a complex control system, a system with a large lag, a non-linear control system, and a system that can not accurately express the function, can find a control method through a special method. Obviously, these features are of great use in complex process control, such as the control systems of large thermal power generating units.
2. Application Status of Intelligent Control System 2.1 General Situation Although the intelligent control system is not fully mature in theory, it is in rapid development in practice. In the thermal power plant control system, many suppliers also began to provide corresponding products and systems.
â—† The United States Bentley company put forward an expert system on the fault diagnosis of turbines more than ten years ago. This system can provide diagnosis results for only seven common accidents of turbine generator sets.
â—† In the TR-2000 series provided by Riss Inc. (Rochester Instrument System.), an expert system for substation monitoring can be provided.
â—† The information provided by Siemens of Germany shows that their fuzzy controllers have been applied in environmental protection power plants. The basic structure adopted is as follows. In the above control scheme, the basic idea is to keep the conventional control methods as much as possible, and the fuzzy control speeds up the response characteristics of the system. Using conventional systems to eliminate system deviations.
As shown in the figure, the control system includes two loops in parallel, the fuzzy control output corrects the conventional PI control, and the input of the fuzzy control is the control deviation of the controlled system and the differential of this deviation. Fuzzy control includes expert system and fuzzy control.
This system has been applied to BCD boiler FGD furnace pressure closed loop control and chemical water treatment PH control system.
â—† Other manufacturers, such as ABB Bailey and MAX, have also introduced their own ideas and plans for intelligent control.
2.2 In the control of power plants in China, there are many pioneers who are doing pioneering work. Intelligent control strategies are adopted in local control systems of power stations, and some have achieved good results. Examples are as follows:
â—† A military unit adopts fuzzy control to improve the control of the drum water level.
This control system uses the P-Fuzzy-PI scheme.
Different control methods are used in different fields.
When the deviation is greater than or equal to a certain value, proportional control is used to increase the response speed of the system.
When the deviation is less than this value, fuzzy control is used to reduce system overshoot.
When close to the set value, PI adjustment is adopted to eliminate system deviation.
After this scheme is adopted, the system response characteristics and adjustment quality are improved, as shown in Fig. 13.
In the figure above, curve a is PID adjustment, curve b is Fuzzy adjustment, and curve c is P-Fuzzy-PI adjustment.
â—† In order to optimize the combustion control system, a power plant adopts a fuzzy control system, which increases the boiler efficiency by 2.8%.
The power plant is a chain furnace with a large lag and poor control characteristics. During the renovation, a humanoid intelligent fuzzy controller was used to realize the negative pressure control of the combustion and the furnace.
The basic design scheme adopts three stages of coarse adjustment, fine adjustment and fine adjustment, and 5 fuzzy controllers. The fuzzy controllers Ao and A are coarse adjustments. According to the thermal load and the calorific value of the coal, the thickness of the coal seam and the grate rotation speed are calculated. Then, according to the optimized deviation value, the grate change amount and the blower variation amount are output. The fuzzy controller B is fine-tuning, optimizes the coarse adjustment result, and makes the system have a memory function and a learning function. The fuzzy controller C finely adjusts the wind/coal ratio based on the above control. Fuzzy controller D completes the negative pressure control of the furnace.
â—† For the control of variable frequency motors that are increasingly widely used in thermal power plants, some supply companies use fuzzy intelligent control instead of PID traditional control systems to improve their control characteristics. Figure 14 shows the effects of the two control features A: Intelligent controller response characteristics.
B: PID control response characteristics.
◆ AI series instruments. The AI ​​series instruments developed by Xiamen Yuguang Electronic Technology Research Institute add fuzzy control algorithm rules to the PID function. When the deviation is large, fuzzy logic is used to determine the output, just as a skilled worker performs manual operations. When the deviation becomes smaller, the improved PID algorithm output is used. The system features no overshoot and high control accuracy. For non-linear complex adjustments, adaptive adjustment rules are provided to improve the quality of adjustment. A self-tuning expert system is set in the table to save debugging time and workload.
2.3 Many examples of development and application of intelligent control systems can be cited at home and abroad. However, only some of the above-mentioned examples can indicate that intelligent control has already started in the field of power plant control. The ability of intelligent control in solving complex systems is being confirmed. Intelligent control systems improve the quality of automated systems and are recognized by many people. The application areas of intelligent systems are advancing both in depth and breadth.
However, on the other hand, in the field of thermal power plant control, smart systems are just starting.
Although there are many examples of successful applications, it is still rare compared to the widespread adoption of classical control systems and it still does not constitute any climate.
Although many manufacturers introduced their intelligent system control strategy, no manufacturer has promised to adopt a smart algorithm DCS system on a large scale.
In the large-scale project construction of thermal power plants, few owners require intelligent control systems.
In a word, although the intelligent control system has already started in the field of control of thermal power plants, the road to go is still very long. There are still many problems to be solved.
3. Several problems with the use of intelligent controls in thermal power plants.
3.1 The use of intelligent controls in thermal power plant control systems will at least help solve the following problems:
â—† Help solve the problems in the control system.
Among the control systems of thermal power plants, some systems are difficult to put into automation due to a variety of reasons, and they become the most difficult problems. Such as steel ball mill load adjustment. The use of intelligent control systems can solve this problem by providing another way to properly handle this issue. Another example is the fact that combustion regulation is often difficult to invest, and in many domestic power plants, the use of smart regulation has set a precedent for solving this problem.
â—† Improve the control system.
One of the weaknesses of the traditional control system is that it can only be operated in accordance with the set control law and parameters. In fact, the operating conditions are very complicated. Even if many correction loops are added, the system cannot be guaranteed to operate in the best state. For example, boiler excess air coefficient. The control, the traditional control system is to refer to the boiler manufacturer data, empirical data, and in accordance with the thermal test results, to determine a given value. In order to adapt to the conditions under different loads, the above given value is added with "load" correction. In fact. It is a very important parameter affecting the economic operation of the boiler. It is related to whether the boiler is operating at a suitable wind/coal ratio. It is related to whether the combustion is complete and whether the air supply is excessive. That is to say, it is the most important issue that affects the economic operation of the boiler. The index q2 (exhaust loss), q3, (chemical incomplete combustion loss), q4 (mechanical incomplete combustion loss) values. At the same time, the reasonable value of α is far from being determined by theoretical calculations and thermal tests, and it is not only related to the load. For example, it is related to the type of coal, coal, calorific value, moisture, ash, etc. It is also related to boiler conditions such as air leakage, cleanliness of the heat transfer surface, and so on. Due to the existence of many complex factors, the conventional regulation system, even if the combustion regulation system can be put into automatic, does not mean that the system is operating in an ideal state.
The use of intelligent control systems can improve the functionality of this system.
Intelligent control strategy to solve this problem, not directly determine the value of α, but to determine the relationship between α and boiler efficiency, and affect the dynamic relationship between α factors, and through the self-learning adaptive system to achieve the desired state. Therefore, α is not a definite, fixed value here, but is a dynamic data that adapts to various complex situations and achieves the best numerical value of boiler efficiency.
Similar systems in power plant control systems also have, in particular, optimized PID parameters that allow them to operate at a truly reasonable value. The use of intelligent systems is an option.
â—† Improve the level of control.
The need to continue to improve the control level of thermal power plants is a consensus of all parties. Regardless of the size of large-scale units or small and medium-sized units, the number of operating personnel is decreasing. This requires that the control logic, especially the design of protection logic, must include as many complicated situations as possible. This also has difficulties for conventional control systems.
For example, for the unit-level self-starting logic design, due to the complexity of the situation, although many people have made unremitting efforts over the years, there are many achievements. However, with the conventional control system, it is still necessary to achieve the goal of achieving unit-level program completion from the completion. A small distance.
If the intelligent control system is used to implement the unit-level program startup, many uncertain factors can be taken into account in the logic of the control system; the impact of many sudden factors on the startup process can be handled; many operational experiences can be programmed into the control system logic; Through self-learning, self-adaptive and other functions, the system has been continuously improved through operation. With these advantages, there is a new way to solve this problem.
Another example is the increase in the level of control of thermal power plants and the requirement to reduce operating personnel. The most prominent issue in this regard is the automation of auxiliary workshops. Auxiliary workshops can be equipped with a program control system. However, changes in the way of operation of the workshop, handling of accidents, and switching of emergency operations are often inseparable from human intervention. Using intelligent control system, a large amount of logic analysis enters the control system. The judgment of accidents, the enhancement of logical reasoning function, and the adoption of humanoid intelligence provide possibilities for solving the unattended design of the auxiliary workshop.
Therefore, to improve the control level, there must be new ideas. The intelligent control system is an optional path.
â—† Develop new control features.
The first problem in this area is the equipment fault diagnosis.
We know that the fault diagnosis of turbogenerators, such as Bentley, can already provide expert system software. Similar software for boilers and power plant auxiliary systems is also under development. In the nuclear power plant control system, software similar to the intelligent system used for fault diagnosis and alarm of the equipment is being developed.
The intelligent system was used to diagnose the fault of the equipment, and the reasoning judgments of some operators were changed to the intelligent control system to speed up the time for fault diagnosis, and it was possible to judge the health status of the unit. This is of great significance for improving the availability of equipment.
In addition, the auxiliary engine operation control design of the power plant now basically stays in the scope of the protection interlock logic. As for the principle of the CCS auxiliary operation mode management, it is only designed for the unit's safe operation requirements. However, the mode of operation of auxiliary engines has a great bearing on the economic efficiency of power plants. For example, the operation of a coal mill, the operation of a feed pump, etc., and the operation of several units under what load, etc., are now mostly determined by humans. With the introduction of intelligent control systems, it is possible to calculate and judge the rationality and economic efficiency of the auxiliary machine operation mode, and improve the economic operation level of the power plant.
Secondly, the introduction of intelligent systems in the operation and management of power plants, such as technical management, overhaul management, and operation and management of units, will also increase the level of management.
3.2 Some issues to be taken into account when adopting intelligent systems In summary, the use of intelligent control systems in thermal power plants can accomplish a lot. Whether it is a new-build project or the technological transformation of existing power plants, the deployment of intelligent systems will bring benefits. .
However, for an enterprise, any new technology adoption must pursue economic benefits. Generally, it cannot allow failure, especially big failure. This is fundamentally different from scientific research. The intelligent control system is completely a new thing, so what issues should be noticed here? I think of the following points for reference.
First, the use of intelligent control systems requires special attention to software development.
For different objects, different systems, choose the appropriate type of intelligent system, algorithm and parameters in it. For this reason, designers, end users (operators) should be more involved in the system's software design process to reduce risk.
Second, use relatively mature systems and software as much as possible.
As the intelligent control system is growing, for the control system of the thermal power plant, information on the complete set of software for the entire power plant control system has not yet been seen. Some suppliers promise to provide some local software. In addition to what has been mentioned above, there are ABB-BAILEY companies that can provide some software. However, due to the rapid changes in the situation, the company concerned should be consulted.
The software provided by suppliers, especially those developed for power plants, is expensive, but it is less risky and practical.
It should be pointed out that for some systems that are not very complicated, it is entirely possible to develop application software on their own.
Third, we must consider combining with the original system.
The intelligent control system has advantages in logic judgment, control mode selection, control parameters, etc. However, no difference adjustment is realized, and traditional PI control is very convenient. The use of intelligent control systems should not lose these advantages.
Logical reasoning of intelligent control systems has its own characteristics and advantages, but its foundation cannot leave the knowledge of the process knowing. On the contrary, it is on this basis that the control system logic is established. Therefore, making full use of the original control system that includes known operating laws and knowledge and making an interface with the intelligent control system is an important condition for the success of the project.
Some vendors provide systems that are parallel to conventional systems and intelligent systems. In addition to division of labor, there are also redundant considerations. In the current situation, it is also a risk reduction measure.
Of course, there are still many problems involved in the design of intelligent control systems, such as the selection of hardware, the calculation of benefits, etc. I am also researching.
3.3 Newly built power plant intelligent control system configuration assumptions Here, the idea of ​​configuring intelligent control parts for the new thermal power plant control system is proposed. Of course, this idea is not mature and is only for the reference of colleagues in similar work. I think that if other conditions are equal, manufacturers that can provide intelligent control systems should have advantages because intelligent control systems can bring economic benefits to the company.
â—† CCS section Generally speaking, the PID operation part can be intelligent, even for a base-type adjustment system or an electric base transmitter, or an auxiliary workshop adjustment system, can be intelligent. Intelligent PID regulation can at least improve the quality of the adjustment, and it will not result in a large increase in investment for new projects.
For a conventional method, a difficult adjustment system, or a system with poor results, should be intelligent, such as a combustion regulation system, an air supply regulation system, a pulverizing system regulation, a main steam and a reheat steam regulation system, etc. Intelligentization will increase the automation investment rate, improve the safety and economic operation level, and bring considerable economic benefits to the company.
For the complete adjustment of complex systems, such as water supply regulation, including the switching logic of the steam pump, electric pump, and low-load control valve, should be intelligent. Especially for very large units, the control logic is complex and intelligent is an important way to solve the problem.
Of course, accurately determining the scope of intelligence needs to be determined according to the actual conditions in various aspects of project implementation.
â—† DAS The main problems existing in the current DAS function are the performance calculation and optimization of economic operation modes; the cause of the accident and the corresponding operation guidance. In theory, intelligent control systems can be introduced in both of these areas, and the necessary judgment systems can be established to provide necessary decision-making reference materials for enterprises. It is also possible to seek ways to completely solve these problems. Unfortunately, no information has yet been provided on providing mature software vendors in this area.
â—† Fault Diagnosis of Turbine Generator Turbogenerator fault diagnosis expert systems are available from suppliers. However, it is rarely applied to practical projects. Needless to say, it is not a simple matter for existing turbogenerator fault diagnosis software to make it play its due role. This includes the input of relevant materials; the configuration of professional and technical management personnel and so on. Especially the latter problem is not easy for a power plant. Because of the limited number of generator sets in a power plant, experts in fault diagnosis for specialized turbine generator sets usually have very few things to do, and experts who do not have the ability to use smart systems at the time of the key are not. To solve this problem, we should give full play to the advantages of computer networks and establish regional fault diagnosis centers so that we can give full play to the role of intelligent systems and experts.
The boiler's fault diagnosis system is still under development, and no systematic product appears. This is because the complexity of the problem makes it difficult to describe complex processes using classical mathematical models. With the development of intelligent control, it is possible to adopt new methods to achieve breakthroughs. For example, in the determination logic of the MFT action condition, the first-out cause determination may be added. For other system protection actions, there should also be first-out reason determination. This will provide the operator with important operating instructions. It should be mentioned that in the nuclear power plant's control system, someone uses the principle of neuron system control to develop similar software. Its purpose is to assist operators to quickly determine the real cause of the accident.
â—† The SCS system SCS system is currently lack of optimized operation mode to improve the safety and economic level of auxiliary equipment operation. Intelligent control system can make a difference in this respect, but at present there is still a blank space.
◆辅助车间控制现在辅助车间多采用PLC控制，控制的范围一般包括程åºå¯åŠ¨ã€åœæ¢ï¼Œå±€éƒ¨å‚数调节，事故时åœæ¢æŠ¥è¦ï¼Œä»¥åŠå’ŒDCS通讯ç‰ã€‚没有包括设备的è¿è¡Œæ–¹å¼ä¼˜åŒ–，è¿è¡Œæ–¹å¼è‡ªåŠ¨çš„自动切æ¢ï¼Œç‰¹åˆ«åœ¨äº‹æ•…时自动切æ¢ã€‚如果就地ä¸å†é…ç½®è¿è¡Œäººå‘˜ï¼Œè¿è¡Œäººå‘˜åŽŸæ¥å¤„ç†çš„许多æ“作，特别事故处ç†çš„æ“作，完全采用精确数å¦æ¨¡åž‹å¤„ç†ï¼Œå°±ä¼šæœ‰è¾ƒå¤§çš„困难。而采用智能系统，就å¯ä»¥æ¯”ä¼ ç»Ÿçš„ç³»ç»Ÿå¤„ç†æ›´å¤šçš„问题。
4.åŠ å¿«ç«åŠ›å‘电厂控制系统智能化建议4．1应该把ç«åŠ›å‘电厂控制系统智能化工作æ到日程上æ¥ã€‚
应该如何对待ç«åŠ›å‘电厂控制系统智能化这一问题呢?温故而知新，在这里我们ä¸å¦¨ç®€å•å›žå¿†ä¸€ä¸‹æœ€è¿‘一个时期我国ç«åŠ›å‘电厂控制系统å‘展的历å²ã€‚
上个世纪七ã€å…«å年代，我们çƒè¡·äºŽå¼€å‘DDZâ…¡ã€â…¢é‡åž‹ä»ªè¡¨æ—¶ï¼Œè¿™ç±»ä»ªè¡¨åœ¨å›½å¤–æ£åœ¨èµ°å‘åŽæœŸï¼Œè®¡ç®—机控制系统已ç»å…´èµ·ã€‚当我们DDZ系列仪表刚刚趋å‘æˆç†Ÿï¼Œä¹Ÿåˆ°äº†æ·˜æ±°çš„边沿。DCS分散控制系统国内产å“，在市场还未æ¥å¾—åŠéœ²é¢ï¼Œå›½å¤–厂商已ç»è¿…速å 领了ä¸å›½å¸‚åœºã€‚ä¸ºä»€ä¹ˆä¼šè¿™æ ·?为æ¤å¯ä»¥è®²å‡ºè®¸è®¸å¤šå¤šçš„åŽŸå› ï¼Œä½†æ˜¯ä»Žæ ¹æœ¬ä¸Šè®²ï¼Œæ˜¯å¯¹æŽ§åˆ¶ç³»ç»ŸæŠ€æœ¯å‘展方å‘和速度估计ä¸è¶³ã€‚ç›®å‰ï¼ŒæŽ§åˆ¶ç³»ç»ŸæŠ€æœ¯çš„å‘展速度并没有å‡ç¼“，对å‘电厂控制系统å‘展è¦æ±‚ä¾ç„¶å˜åœ¨ï¼Œå› æ¤ï¼ŒæŠŠæ¡ä½ä¸»è¦æŠ€æœ¯å‘展方å‘ï¼ŒåŠ å¿«å‘展速度，在相当长一段时间内都是很é‡è¦çš„。——应该从这个角度æ¥çœ‹å¾…控制系统智能化的问题。
æŽ§åˆ¶ç³»ç»Ÿæ™ºèƒ½åŒ–ï¼Œæ— è®ºä»Žç†è®ºä¸Šå’Œå®žè·µä¸Šéƒ½å·²ç»å…·å¤‡æ¡ä»¶ã€‚但是对æ¤é—®é¢˜çš„认识，在行业内还很ä¸ä¸€è‡´ã€‚相当多的åŒè¡Œå¯¹å®ƒçš„实用价值，还æŒæœ‰æ€€ç–‘æ€åº¦ã€‚把这方é¢çš„情况，资料，åŠæ—¶çš„介ç»ç»™ä¸šå†…人士，共åŒåŠªåŠ›ï¼Œæ˜¯åˆ‡ä¸å®¹ç¼“的事情。
æ¤å¤„还应特别æ出，在智能控制ç†è®ºç ”究上，ä¸å›½å¹¶ä¸è½åŽï¼Œåº”用技术也有一定的基础，问题是统一认识，认真去作，完全有å¯èƒ½åœ¨æ–°ä¸€è½®æŠ€æœ¯ç«žäº‰å æ®æœ‰åˆ©åœ°ä½ã€‚
4．2åŠ å¼ºæŠ€æœ¯äº¤æµï¼Œä¿ƒè¿›æ™ºèƒ½æŽ§åˆ¶æŠ€æœ¯çš„å‘展。
作为一ç§å‰æ²¿å¦ç§‘，智能控制系统在å„æ–¹é¢éƒ½ååˆ†æ´»è·ƒã€‚ç§‘ç ”å•ä½ã€å¤§ä¸“é™¢æ ¡ã€è®¾è®¡é™¢ã€å»ºè®¾å•ä½å’Œç”Ÿäº§å•ä½ï¼Œéƒ½æœ‰äººåœ¨ä»Žäº‹ç†è®ºæˆ–实践工作。但是，目å‰é™¤åŽ»å›½å¤–供货厂商为了推销他们的产å“介ç»ç«åŠ›å‘电厂智能控制技术外，很少能è§åˆ°è¿™æ–¹é¢çš„技术交æµã€‚å»ºè®®åŠ å¼ºæ¤é¡¹å·¥ä½œï¼Œè‚¯å®šæœ‰åˆ©äºŽæ™ºèƒ½æŽ§åˆ¶æŠ€æœ¯çš„å‘展。
4．3充分利用国外æˆæžœç›®å‰æˆ‘国DCSä¾›è´§æ ¼å±€ä»ç„¶ä»¥å›½å¤–系统为主（å«åˆèµ„ä¼ä¸šï¼‰ï¼Œç‰¹åˆ«æ˜¯å¯¹äºŽIï¼O规模超过3000点以上的大型系统更是如æ¤ã€‚既然花钱买技术，就应该使其物有所值。ä¸å¦¨åœ¨æˆ‘ä»¬çš„æ‹›æ ‡æ–‡ä»¶ï¼ŒæŠ€æœ¯è§„èŒƒä¸æ˜Žç¡®æ出è¦æ±‚对方æä¾›æˆç†Ÿçš„æ™ºèƒ½æŽ§åˆ¶æŠ€æœ¯ã€‚è¿™æ ·ï¼Œæˆ‘ä»¬ä¸ç‰å¾…国外形æˆå®Œæ•´ç³»åˆ—产å“å‰ï¼Œåˆ©ç”¨å…¶å·²æˆç†Ÿçš„éƒ¨åˆ†ï¼ŒåŠ ä¸Šæˆ‘ä»¬çš„å¼€å‘工作，有å¯èƒ½åœ¨æ™ºèƒ½æŽ§åˆ¶ç³»ç»Ÿå¼€å‘ç«žèµ›ä¸å¤„于较有利ä½ç½®ã€‚
4．4制定å‘展规划，适时进行总结å¤äººæ—¥ï¼šâ€œå‡¡äº‹é¢„则立â€ï¼ŒæŒ‡åŠžä»»ä½•ä¸€ä»¶äº‹æƒ…åªæœ‰æœ‰è®¡åˆ’æ‰èƒ½æˆåŠŸã€‚å‘电厂控制系统智能化，虽然é¢ä¸´å¤§å¥½æœºé‡ï¼Œå¹¶ä¸ç‰äºŽè¿™ä»¶äº‹å°±ä¼šè‡ªç„¶è€Œç„¶é¡ºåˆ©å‘展。电力行业应和仪表行业è”åˆèµ·æ¥ï¼ŒæŠ“ä½æœºé‡ï¼Œå°½æ—©ç¡®ç«‹æˆ‘国在æ¤é¢†åŸŸå†…应å 有的ä½ç½®ã€‚
å‘电厂控制系统智能化å‘展规划应该包括在å„个方é¢æŠ€æœ¯è¦æ±‚；未æ¥å‘电厂控制系统智能化基本规范；å‚åŠ å•ä½ã€ä¸ªäººå¤§ä½“分工和æƒç›Šï¼›ä»¥åŠå‘展里程表ç‰ç‰ã€‚æœ‰äº†è¿™æ ·ä¸€ä¸ªè§„åˆ’ï¼Œå°±ä¾¿äºŽè°ƒåŠ¨å„æ–¹é¢çš„积æžæ€§ï¼Œå°±æœ‰å¯èƒ½ä½¿ä¸ªåˆ«å•ä½çš„努力汇åˆèµ·æ¥ï¼Œå½¢æˆæˆ‘们的å‘展优势。
当然，仅仅有规划是ä¸å¤Ÿçš„。还è¦åŠ 上定期的把å–å¾—çš„æˆæžœè¿›è¡Œæ€»ç»“，归纳，形æˆç³»åˆ—ã€è§„范，æ‰èƒ½å·©å›ºå’Œå‘å±•ç ”ç©¶æˆæžœã€‚
Cosmetic Paper Box,Cosmetic Bottle Paper Box,Cosmetic Folding Paper Box,Cosmetic Packaging Paper Box
Shenzhen Xianglong Paper Product &Packaging Co.,Ltd , https://www.xlpaperbox.com