Application case 23: The application of intelligent optimization control system in float glass production line
1 Overview
Intelligent control is suitable for the occasion which is unable to be solved by the conventional technology, conventional control, or effects is not good. From the present situation of Chinese automation applications, the application of conventional automation system such as PLC, DCS, and IPC has been very popular. However, these systems are the products based on classical control, when solves the complex object existence in the industrial site of nonlinear, time-varying, large time lag, multi variable, strong coupling and unable to modeling, appears to be inadequate, especially is incapable of action in process optimization . However, this kind of complex process can meet the eye everywhere, and not good control, production is not stable, energy consumption do not down, quality do not go on. Intelligent control method, because of its highly intelligent characteristics (knowledge model), is more advanced and effective than the traditional control method (analytical model) in the description and processing problems of complex control and management. However, due to the lack of complete engineering intelligent control system, the intelligent control is only the patent of experts and professors; and general engineer did not know how to use.
2 The introduction of the general intelligent optimization control system of CIMAC
The control system of CIMAC is a complete set of intelligent optimization control system which is based on the accumulation of automation engineering experience of years to absorb advanced technology at home and abroad and then creatively closely combines artificial intelligence with conventional automation, so that the intelligent optimization control technology to engineering and practicality, including both hardware and software. It can be used independently to control problems of complex process industry.
3．Application of systems in flat glass production
Glass furnace is a complex control object of nonlinear, multivariable coupling, establishing a dynamic mathematic model, and its system cannot solve the control problem using the conventional control method. This is also a problem for many years that plagued the glass industry for long time. If the temperature cannot be good controlled, a lot of fuel will be wasted. To solve this problem, it is the only way at present that using the recent emerging intelligent optimization control method. In addition, artificial intelligence method also plays a very prominent effect on quality prediction, electric heating control, and air-fuel ratio optimization and so on.
The characteristics of the system are as following:
(1) operation mode use a combination of individual and group, namely in the raw materials, ingredients, tin bath, annealing kiln, the cold end can be local control, also focus on the control room to control (selected by the selection switch with lock), and easy to operate, monitor and control and a reduction in staff positions. The detection and control use BCS and networking, to process and display device, data can be shared, and the display screen can call each other.
(2) Unitary taking into account, but also can be implemented step by step, for example firstly the pivotal kiln system can be installed, to receive good results.
(3) three-level automation system structure, including the based automation of the detection driver stage L1 of detection meter actuator, all kinds of electrical equipment, motor control center, control devices, industrial TV etc, and the equipment control level L2 composed of BCS, as well as process automation of control system L3.
(4) The integration of management and control. The system may include some management functions, such as the quality analysis, statistics and planning, as well as monitoring in the room and control room etc.
(6) Integrating industrial TV network with EICC four power (transmission, instrument, computer, communication), to form new system of typical multimedia technology. CCD industrial TV will enter the EICC system by BCS.
(7) Using a large capacity industry PC (1.7G hard disk, memory above 16M) to instead of the traditional process computer not only saves the investment, but also increases the flexibility.
(8) Using the IE integration equipment, namely the PLC system, not only the cost is lower than that of DCS, but also reduce spare parts because of the integration of electric and instrument equipment and.
(9) Using artificial intelligence technology (mathematical model of control or expert system) to control the operation of pivotal equipment melting kiln, to achieve energy saving and prolonging the lifespan of kiln, so that the inexperienced operators also can achieve proficiency level of operators. Using neural network technology to predict Glass quality, so that we can timely correct ingredients and take other measures, but also predict index of the production technology after changing the process parameters, in order to make the decision and strengthening production. In addition, the traditional control mode is PID, but it is difficult to adapt to the numerous disturbances and load changes in practice, especially with the production process, the adjusting valve often become not flexible, valve size is too large, the control quality can't meet, oscillation and so on, and then make it cannot be used, and finally change into manual control. At domestic and aboard the mathematical model of control is used to solve, it is mathematical model and optimization by manual operation rule, it not only makes the automation without replacing any equipment to the possible (because normally, the valve is difficult to replace, because of that not only the price is expensive, and we must stop production to install, and even cut production to meet the pipeline pipe valve size requirements, and it is difficult to achieve). And the regulation quality is better than PID, such as regulating time, attenuation, wave number, and the adaption of the quality control to interference are more superior, it is particularly suitable for the situation of control kiln temperature by adjusting the hot gas gate, because the hot gas gate with tar is not flexible, in addition, also can make combustion optimal by neural network, namely air-fuel ratio is optimal and the maximum flame temperature, to solve the problem that the air fuel ratio control cannot be realized, because the thermal gas flow could not be measured.
(10) Using the new scheme and new advanced control method. Using optimization mathematical model with the objective of the minimum cost to calculate ingredient in order to reduce costs, and increase market competition ability.
3.1 Process automation
Mainly used to monitor the whole production process and its main function are:
(1) Data acquisition. Including the time series data of process parameters in whole processes, such as crushing, proportioning, melting, forming, cutting and packing, parameters the manually inputted, to live collect.
(2) The setting of production process parameter. For example, the toughened glass production needs to set parameters (pressure, toughened glass and air nozzle distance, power), motion parameters (oscillation time, speed, acceleration), and to adjust the parameters (the heating temperature and time of each * * section) for lower-computer, it is to choose according to the glass thickness, specifications etc., the toughened glass production line imported from Switzerland of a factory in China has 31 kinds of specifications, namely, 31 sets of data, which are stored in the host computer, for later use.
(3) Model operation. Such as calculating model of ingredient, optimization model of ingredient (the lowest cost).
(4) The advanced control and artificial intelligence applications. including furnace temperature mathematical model control, furnace pressure mathematical model control, ingredient mathematical model control, it solves the lag, the valve diameter is too large, is not flexible, nonlinear object and the oscillations of the system, and control quality bad etc, which are difficult to be solved for the conventional PID control. Another purpose of artificial intelligence is to find the optimal fuel-air ratio to make flame temperature highest, it uses neural network.
(5) Quality prediction. It uses statistical modeling or neural network methods, only using the process parameters of ingredient can predict the quality of flat glass, solve the problem that knowing whether the ingredients are correct or not until the finished product is out later in the past and not timely, in addition, changing the operating parameters, strengthening production, can also be predicted by this model, in order to making decision.
(6) Data display. Including the time series data, the instantaneous data, the process flow chart of process parameters and key dynamic data, alarm data, historical data query, report forms, special display, model display, setting data, storage process parameters, curve, quality standards and technical regulations query etc..
(7) Data recording. Including prints, dates and graphs and other displayed hardcopies about class newspaper, daily newspaper, monthly newspaper, and alarming report.
(8) Data communication, including the data communications which is used to communicate with the computer of factory management, lower computer, computer of director and control room, or terminal.
(9) Other management functions. Such as the equipment diagnosis, planning, the process of remote monitoring production by the director room and control room and so on.
3.2 The basic automation
The base automation functions of furnace include (1) data collection: the kiln temperature, kiln pressure, pumping force of the flue, glass level, gas or heavy-oil and air pressure etc; (2) the automatic control model of kiln pressure; (3) the mathematical model of expert system for controlling the temperature of each section; (4) to realize optimization control of air-fuel ratio by the neuron network; (5) automatic changeover of flame; (6) the automatic control of glass surface; (7) temperature liquid, gas or heavy oil and air pressure alarm.
The basic automation functions of tin groove and annealing kiln include: (1) data acquisition; (2) the temperature intelligent control; (3) neurons electric heating control system; (4) the temperature alarm.
4 The main contents of the glass intelligent optimization control
4.1 automatic calculation and optimization of ingredients
A. batch feeding automatically, according to the requirements of kiln, storage bin and material level on the front of kiln, realize the full automatic feeding.
B. automatic calculation model is automatically calculate the weights of 6 kinds of raw materials (dry) according to the composition of target glass and process requirements and ingredients, and then according to the moisture of target glass, the amount of mixture broken glass and the actual water content of each raw material, to calculate the actual wet weight of raw materials, at last print a single ingredient. This calculation can be completed in 1 second, while the artificial need half a day.
C. the optimization burden model, it is the hybrid model using the minimum cost, the 7 compositions of objective glass and other process requirements as objective function. It matched with the conventional calculation method can reduce 5-10% of the cost.
4.2 The system of glass quality prediction and fault diagnosis expert, can predict the common 4 categories of 39 kinds of glass defects
A. includes 7 kinds of chemical composition, physical performance prediction.
B. Using statistical modeling or neural network mode，mathematical model of glass quality parameter is automatically established, online real-time indicating the glass quality under the conditions of some process operation.
C. In all the processes of flat glass production including the ingredients, melting furnace, tin furnace, annealing furnace, cutting and packaging etc, when quality defects appear -- there are 4 major categories of 39 kinds of defects: air bubble defect, vitreous inclusion defects, non transparent inclusions defects and other defects (such as blowhole, wave, Linzi, gang Zi, non transparent inclusions, tin, iridescent, edge crack, substandard ingredients, refraction unqualified) can indicate the reason by means of expert system at any time, then process these problems including the reasons from which process, and is provided with an internal database, with the national standard and plant standard of internal produced glass, the diagnostic system based on automatic acquisition or manual input data, automatically compared with the standard database, if there is a defect, it will automatically reason out the quality defects and processing method according to the rules of the system knowledge.
4.3 New-style flame reversing system optimization model
New-style Flame reversing system optimization model, to reduce the temperature fluctuation of flame reversing, from the original 40℃temperature drop down to 10℃, including:
A. New-style commutation chart -- meeting gas air flame, so that idle time is the shortest, furnace temperature minimum, to save fuel and reduce the turnaround time.
B. Optimizing commutation interval time -- according to the optimization model and the temperature of regenerator to calculate the optimum commutation interval time, and setting an expert system to automatically correct according to the production process, process parameter.
C. Using the mathematical model system, to make the fluctuation of furnace pressure minimum, and the flame most stable, when commutation.
4.4 The temperature mathematical model control of glass melting kiln
The formation of glass needs suitable viscosity, and the viscosity depends on the glass composition and temperature, therefore the temperature of melting furnace must be stable. Automatic temperature control of furnace uses zone control method, including the indirect control method and direct control method. When using mathematical model control method, the temperature fluctuation is reduced from conventional 5℃to 1℃.
The temperature of melting furnace is very high (about 1500℃); molten pool is longer, the temperature distribution needs to meet the specific temperature curve. In addition, in the glass melting process, the changeover of flame is one time each about 20 minutes, the positions of foam line and hot spot should be maintained at a certain position, flame length, brightness and direction have certain requirements, the temperature measuring point should be the temperature control point which is not only representative, but also not affected by the above disturbance.
The direct control of melting furnace temperature has a variety of ways, one of which is "using gas to adjust oil", the formation of glass in melting furnace can be divided into three stages such as material, melting, clarifying, and then is divided into three areas to control. It is by adjusting the amount of atomizing steam by the kiln temperature to change the fuel injection quantity (glass furnace uses high-pressure nozzle, changing the amount of atomizing steam can change the fuel injection quantity in a certain range) to keep the temperature set value. The system temperature fluctuation is ± 3℃, and the manual is about ± 12 ~ 30℃.
Another method of direct control of the melting furnace temperature is directly adjusting oil, to make steam and oil proportional control according to the pressure. The system temperature fluctuation is ± 2 ~ 3℃, the manual is about ±20~30℃.
For the melting furnace using coal gas as fuel, the direct control of the furnace temperature has different system according to gas conditions:
A. when using the gas passing by the cleaned gas furnace, because the gas is clean and can be directly measuring the gas flow, the control system will be as shown in Figure 3, it is a series parallel cross limiting system, i.e. if the temperature deviation from the specified value, change the preheated gas and air flow to make the temperature back to the specified value. K is mainly used to regulating excess air coefficient.
B. when using the hot coal gas passing by the no-cleaned gas furnace, because the gas contains a large amount of tar and cannot directly measuring the gas flow, there are two options, one is to modulate the gas gate opening before getting into the gas regenerator in order to keep the temperature in a given value (Figure 4); the second is to adjust the incoming air volume of gas furnace, because the gas quantity of gas furnace and the volume of blast is in proportion, but this only is suitable for that condition, that the gas furnace was only used by the melting furnace.
Figure 4 The principle of the automatic control system of the temperature of glass melting furnace which combusts hot gas passing by the no-cleaned gas furnace
The indirect control of temperature of melting furnace has “small furnace measurement and control". Its starting point is according to the process to supply the stable fuel flow, and then can make the temperature of the kiln stability and meet the requirements of temperature distribution curve. For the melting furnace using coal gas as fuel, gas consumption is stable supplied. Because the online measurement of temperature is difficult to be accurate, so the indirect method of controlling temperature is more commonly used.
4.5 The optimal control system of air - fuel ratio
Using optimal control system of air - fuel ratio, about 5% of the fuel can be saved. Regardless of fuel for heavy oil or gas, its principle is similar, is a series parallel system, when the temperature changes, changing the secondary air and the setting value of control loop of fuel flow so as to make the temperature back to the specified value, and the high-low value selector to ensure system in dynamic adjustment to achieve that, when low temperature, firstly increasing air, and then increasing the gas, or, firstly decreasing oil, and then decreasing wind, in order to obtain sufficient combustion, no black smoke from the chimney. The system also has a self optimizing process, according to the maximum flame temperature, i.e. the optimal firstly- secondary air ratio of fuel.
4.6 The neuron electric heating control system
Using neuron electric heating control system, 10-20% of electric can be saved. Float glass production line in addition to consume the more fuel (oil / coal), also consume the electrical energy. Glass production has three thermal equipments, namely the glass melting furnace, tin bath and the annealing kiln, in glass melting furnace burning by the small batch system supply of fuel, in the other two thermal equipment -- tin slot and the annealing kiln temperature control mainly by the electric heating and cooling fan, the electric heating control system is very important in the production of glass. Using the intelligent optimization control system of CIMAC as the basis, and combining with the specific process requirements, to complete the neuron electric heating control system. The detailed structure of control system is as following:
In the dynamic optimization algorithm of electric heating control system, we adopt the structure of dynamic optimization algorithm learning control, also called the supervision and control of dynamic optimization algorithm. On the basis of the conventional PID controller, adding a dynamic optimization algorithm controller. At this time, the dynamic optimization algorithm controller actually is a feed forward controller, so it is established that the inverse model of the controlled object. From the graph we can see that, dynamic optimization algorithm controller is learning through the output to traditional controller, online adjust their selves, the goal makes feedback error E (T) or U1 (T) approach to zero, so that they gradually occupy the leading position in the control effect, so as to eventually cancel the feedback controller function. Once the interference of system appears, the feedback controller is still play a role. It can ensure the robustness and stability of the system, but also can effectively improve the control accuracy.
5 The enterprise benefits
Using the intelligent optimization control system of CIMAC to automatically control the production process of flat glass, can save 10%-20% of energy, improve at least 5% of the yield and yield rate, in addition, also save manpower and reduce staff etc, to obtain significant economic benefits.
Next, we use the 400t/d float line as an example. After the 400t/d float line is put into use by company, due to the glass quality is improved, and the cost is reduced, so 5000000 Yuan of the annual profit can be increased. Another increase is 1 year of furnace service life, amounting to 7000000 Yuan.
(1) Improving glass quality, the company added profit about 1200000 Yuan per year.
Through correction of the water and weight of ingredients, quality prediction and the temperature control of three thermal equipments and so on, to make the rate of finished products of glass be increased, the annual production of glass increase 10000 weight boxes, and the annual benefit reaches 600000 Yuan, and make the high quality rate increase, namely the high quality glass can be produced, and the economic benefit is 600000 Yuan.
(2) Reducing the cost of the product, the company added profit of 3800000 Yuan per year.
A. Through the Glass Batching optimization, the company can save 2% of raw materials costs, the materials costs is 27600000 Yuan per year, so it can save 560000 Yuan.
B. Through saving energy, the company can reduce the cost about 3260000 Yuan.
a. Reducing fuel consumption about 15%, saving 8100 tons of coal, saving expenditure about 2260000 Yuan. The specific ways:
•Water supplemented makes the furnace burden easy to melt, reducing fuel consumption about 5%.
• The air - fuel ratio optimization, reducing fuel consumption 10%
• The optimal flame reversing
• The kiln temperature mathematical model control
b. Reducing the tin consumption
Reducing the original 7g/ weight boxes of glass liquid to 3-4g/ weight boxes of glass liquid, tin can be saved about 6 tons per year, the value is 300000 Yuan.
c. Saving electricity
The annual power consumption of the tin groove and the annealing kiln is 36000000 degrees, if we use neuron electric heating control system, the mathematical model of the temperature control and variable frequency device, can save electricity about 10%, the annual saving is 3600000 kwh, reducing electric charge about 1800000 Yuan.
(3) Improve the kiln body life, the average annual savings fund is 7000000 Yuan.
From the original 4 years up to 5-6 years, using the improved life of 1 year as an example, the increased annual production of glass is 1700000 weight boxes, can save more than 7000000 Yuan for the enterprise.
(4) Reducing the generation of black smoke, and protecting environment. Basically the black smoke cannot be seen.