自动化 外文翻译 外文文献 英文文献 锅炉蒸汽温度模糊神经网络的广义预测控制

自动化 外文翻译 外文文献 英文文献 锅炉蒸汽温度模糊神经网络的广义预测控制 目 录 Part 1 PID type fuzzy controller and parameters adaptive method1 Part 2 Application of self adaptation fuzzy-PID control for main steam temperature control system in power station错误～未定义 书 关于书的成语关于读书的排比句社区图书漂流公约怎么写关于读书的小报汉书pdf 签。7 Part 3 Neuro-fuzzy generalized predictive control of boiler steam temperature .............................................................. ………13 Part 4 为Part3译文:锅炉蒸汽温度模糊神经网络的广义预测控制 22 Part 1 PID type fuzzy controller and Parameters adaptive method Wu zhi QIAO, Masaharu Mizumoto Abstract: The authors of this paper try to analyze the dynamic behavior of the product-sum crisp type fuzzy controller, revealing that this type of fuzzy controller behaves approximately like a PD controller that may yield steady-state error for the control system. By relating to the conventional PID control theory, we propose a new fuzzy controller structure, namely PID type fuzzy controller which retains the characteristics similar to the conventional PID controller. In order to improve further the performance of the fuzzy controller, we work out a method to tune the parameters of the PID type fuzzy controller on line, producing a parameter adaptive fuzzy controller. Simulation experiments are made to demonstrate the fine performance of these novel fuzzy controller structures. Keywords: Fuzzy controller; PID control; Adaptive control 1. Introduction Among various inference methods used in the fuzzy controller found in literatures , the most widely used ones in practice are the Mamdani method proposed by Mamdani and his associates who adopted the Min-max compositional rule of inference based on an interpretation of a control rule as a conjunction of the antecedent and consequent, and the product-sum method proposed by Mizumoto who suggested to introduce the product and arithmetic mean aggregation operators to replace the logical AND (minimum) and OR (maximum) calculations in the Min-max compositional rule of inference. In the algorithm of a fuzzy controller, the fuzzy function calculation is 1 also a complicated and time consuming task. Tagagi and Sugeno proposed a crisp type model in which the consequent parts of the fuzzy control rules are crisp functional representation or crisp real numbers in the simplified case instead of fuzzy sets . With this model of crisp real number output, the fuzzy set of the inference consequence will be a discrete fuzzy set with a finite number of points, this can greatly simplify the fuzzy function algorithm. Both the Min-max method and the product-sum method are often applied with the crisp output model in a mixed manner. Especially the mixed product-sum crisp model has a fine performance and the simplest algorithm that is very easy to be implemented in hardware system and converted into a fuzzy neural network model. In this paper, we will take account of the product-sum crisp type fuzzy controller. 2. PID type fuzzy controller structure As illustrated in previous sections, the PD function approximately behaves like a parameter time-varying PD controller. Since the mathematical models of most industrial process systems are of type, obviously there would exist an steady-state error if they are controlled by this kind of fuzzy controller. This characteristic has been stated in the brief review of the PID controller in the previous section. If we want to eliminate the steady-state error of the control system, we can imagine to substitute the input (the change rate of error or the derivative of error) of the fuzzy controller with the integration of error. This will result the fuzzy controller behaving like a parameter time-varying PI controller, thus the steady-state error is expelled by the integration action. However, a PI type fuzzy controller will have a slow rise time if the P parameters are chosen small, and have a large overshoot if the P or I parameters are chosen large. So there may be the time when one wants to introduce not only the integration control but the derivative control to the fuzzy control system, because the derivative control can reduce the overshoot of the system's response so as to improve the 2 control performance. Of course this can be realized by designing a fuzzy controller with three inputs, error, the change rate of error and the integration of error. However, these methods will be hard to implement in practice because of the difficulty in constructing fuzzy control rules. Usually fuzzy control rules are constructed by summarizing the manual control experience of an operator who has been controlling the industrial process skillfully and successfully. The operator intuitively regulates the executor to control the process by watching the error and the change rate of the error between the system's output and the set-point value. It is not the practice for the operator to observe the integration of error. Moreover, adding one input variable will greatly increase the number of control rules, the constructing of fuzzy control rules are even more difficult task and it needs more computation efforts. Hence we may want to design a fuzzy controller that possesses the fine characteristics of the PID controller by using only the error and the change rate of error as its inputs. One way is to have an integrator serially connected to the output of the fuzzy controller as shown in Fig. 1. In Fig. 1,andare scaling factors for e KK12 and ~ respectively, and fl is the integral constant. In the proceeding text, for convenience, we did not consider the scaling factors. Here in Fig. 2, when we look at the neighborhood of NODE point in the e - ~ plane, it follows from (1) that the control input to the plant can be approximated by (1) 3 Hence the fuzzy controller becomes a parameter time-varying PI controller, its equivalent proportional control and integral control components are BK2D and ilK1 P respectively. We call this fuzzy controller as the PI type fuzzy controller (PI fc). We can hope that in a PI type fuzzy control system, the steady-state error becomes zero. To verify the property of the PI type fuzzy controller, we carry out some simulation experiments. Before presenting the simulation, we give a description of the simulation model. In the fuzzy control system shown in Fig. 3, the plant model is a second-order and type system with the following transfer function: K (2) G(s),(Ts，1)(Ts，1)12 Where K = 16, = 1, and= 0.5. In our simulation experiments, we use TT12 the discrete simulation method, the results would be slightly different from that of a continuous system, the sampling time of the system is set to be 0.1 s. For the fuzzy controller, the fuzzy subsets of e and d are defined as shown in Fig. 4. Their cores The fuzzy control rules are represented as Table 1. Fig. 5 demonstrates the simulation result of step response of the fuzzy control system with a Pl fc. We can see that the steady-state error of the control system becomes zero, but when the integration factor fl is small, the system's response is slow, and when it is too large, there is a high overshoot and serious oscillation. Therefore, we 4 may want to introduce the derivative control law into the fuzzy controller to overcome the overshoot and instability. We propose a controller structure that simply connects the PD type and the PI type fuzzy controller together in parallel. We have the equivalent structure of that by connecting a PI device with the basic fuzzy controller serially as shown in Fig.6. Where ~ is the weight on PD type fuzzy controller and fi is that on PI type fuzzy controller, the larger a/fi means more emphasis on the derivative control and less emphasis on the integration control, and vice versa. It follows from (7) that the output of the fuzzy controller is (3) 3. The parameter adaptive method Thus the fuzzy controller behaves like a time-varying PID controller, its equivalent proportional control, integral control and derivative control components are respectively. We call this new controller structure a PID type fuzzy controller (PID fc). Figs. 7 and 8 are the simulation results of the system's step response of such control system. The influence of ~ and fl to the system performance is illustrated. When ~ > 0 and/3 = 0, meaning that the fuzzy controller behaves like PD fc, there exist a steady-state error. When ~ = 0 and fl > 0, meaning that the fuzzy controller behaves like a PI fc, the steady-state error of the system is eliminated but there is a large overshoot and serious oscillation. 5 When ~ > 0 and 13 > 0 the fuzzy controller becomes a PID fc, the overshoot is substantially reduced. It is possible to get a comparatively good performance by carefully choosing the value of and. ,, 4. Conclusions We have studied the input-output behavior of the product-sum crisp type fuzzy controller, revealing that this type of fuzzy controller behaves approximately like a parameter time-varying PD controller. Therefore, the analysis and designing of a fuzzy control system can take advantage of the conventional PID control theory. According to the coventional PID control theory, we have been able to propose some improvement methods for the crisp type fuzzy controller. It has been illustrated that the PD type fuzzy controller yields a steady-state error for the type system, the PI type fuzzy controller can eliminate the steady-state error. We proposed a controller structure, that combines the features of both PD type and PI type fuzzy controller, obtaining a PID type fuzzy controller which allows the control system to have a fast rise and a small overshoot as well as a short settling time. To improve further the performance of the proposed PID type fuzzy controller, the authors designed a parameter adaptive fuzzy controller. The PID type fuzzy controller can be decomposed into the equivalent proportional 6 control, integral control and the derivative control components. The proposed parameter adaptive fuzzy controller decreases the equivalent integral control component of the fuzzy controller gradually with the system response process time, so as to increase the damping of the system when the system is about to settle down, meanwhile keeps the proportional control component unchanged so as to guarantee quick reaction against the system's error. With the parameter adaptive fuzzy controller, the oscillation of the system is strongly restrained and the settling time is shortened considerably. We have presented the simulation results to demonstrate the fine performance of the proposed PID type fuzzy controller and the parameter adaptive fuzzy controller structure. Part 2 Application of self adaptation fuzzy-PID control for main steam temperature control system in power station ZHI-BIN LI Abstract: In light of the large delay, strong inertia, and uncertainty characteristics of main steam temperature process, a self adaptation fuzzy-PID serial control system is presented, which not only contains the anti-disturbance performance of serial control, but also combines the good dynamic performance of fuzzy control. The simulation results show that this control system has more quickly response, better precision and stronger 7 anti-disturbance ability( Keywords:Main steam temperature;Self adaptation;Fuzzy control;Serial control 1. Introduction The boiler superheaters of modem thermal power station run under the condition of high temperature and high pressure, and the superheater’s temperature is highest in the steam channels(so it has important effect to the running of the whole thermal power station(If the temperature is too high, it will be probably burnt out. If the temperature is too low ,the efficiency will be reduced So the main steam temperature mast be strictly controlled near the given value(Fig l shows the boiler main steam temperature system structure. Fig.1 boiler main steam temperature system It can be concluded from Fig l that a good main steam temperature control system not only has adequately quickly response to flue disturbance and load fluctuation, but also has strong control ability to desuperheating water disturbance. The general control scheme is serial PID control or double loop control system with derivative. But when the work condition and external disturbance change large, the performance will become instable. This paper presents a self adaptation fuzzy-PID serial control system. which not only contains the anti-disturbance performance of serial control, but also combines the good dynamic character and quickly response of fuzzy control( 1. Design of Control System The general regulation adopts serial PID control system with load feed 8 forward(which assures that the main steam temperature is near the given value 540?in most condition(If parameter of PID control changeless and the work condition and external disturbance change large, the performance will become in stable(The fuzzy control is fit for controlling non-linear and uncertain process. The general fuzzy controller takes error E and error change ratio EC as input variables(actually it is a non-linear PD controller, so it has the good dynamic performance(But the steady error is still in existence. In linear system theory, integral can eliminate the steady error. So if fuzzy control is combined with PI control, not only contains the anti-disturbance performance of serial control, but also has the good dynamic performance and quickly response. In order to improve fuzzy control self adaptation ability, Prof(Long Sheng-Zhao and Wang Pei-zhuang take the located in bringing forward a new idea which can modify the control regulation online(This regulation is: U,,E，(1,,)EC,,,[0,1] This control regulation depends on only one parameter.Onceis ,, fixed(the weight of E and EC will be fixed and the self adaptation ability will be very small(It was improved by Prof. Li Dong-hui and the new regulation is as follow; ,,EECE，(1,),,000 ,,EECE，(1,),,,111U,{,,EECE，(1,),,,2 22 ,,EECE，(1,),,,333 ,,,,,,,,[0,1]0123 Because it is very difficult to find a self of optimum parameter, a new method is presented by Prof(Zhou Xian-Lan, the regulation is as follow: 2,,1,exp(,ke),(k,0) But this algorithm still can not eliminate the steady error(This paper combines this algorithm with PI control，the performance is improved( 2. Simulation of Control System 9 3.1 Dynamic character of controlled object Papers should be limited to 6 pages Papers longer than 6 pages will be subject to extra fees based on their length( Fig .2 main steam temperature control system structure Fig 2 shows the main steam temperature control system structure， are main controller and auxiliary controller,are W(s),W(s)W(s),W(s),1,2o1o2 characters of the leading and inertia sections,are measure unit. W(s),W(s)H1H2 3.2 Simulation of the general serial PID control system The simulation of the general serial PID control system is operated by MATLAB, the simulation modal is as Fig.3.Setp1 and Setp2 are the given value disturbance and superheating water disturb & rice .PID Controller1 and PID Controller2 are main controller and auxiliary controller( The parameter value which comes from references is as follow: W(s),k,25,2p2 1W(s),k，k，ks 1p1I1D1,s k,3.33,k,0.074,k,37.667p1I1D1 10 Fig.3. the general PID control system simulation modal 3.3 Simulation of self adaptation fuzzy-PID control system Spacing The simulation modal is as Fig 4.Auxiliary controller is:.Main controller is Fuzzy-PI structure, and the PI controller W(s),k,25,2p2 is: 1W(s),k，k1p1I1, s k,3.33,k,0.074p1I1 Fuzzy controller is realized by S-function, and the code is as fig.5. Fig.4. the fuzzy PID control system simulation modal 11 Fig 5 the S-function code of fuzzy control 3.4 Comparison of the simulation Given the same given value disturbance and the superheating water disturbance，we compare the response of fuzzy-PID control system with PID serial control system. The simulation results are as fig.6-7. From Fig6-7,we can conclude that the self adaptation fuzzy-PID control system has the more quickly response, smaller excess and stronger anti-disturbance( 4. Conclusion (1)Because it combines the advantage of PID controller and fuzzy controller, the self adaptation fuzzy-PID control system has better performance than the general PID serial control system. (2)The parameter can self adjust according to the error E value. so this kind of controller can harmonize quickly response with system stability( 12 Part 3 Neuro-fuzzy generalized predictive control of boiler steam temperature Xiangjie LIU, Jizhen LIU, Ping GUAN 13 Abstract: Power plants are nonlinear and uncertain complex systems. Reliable control of superheated steam temperature is necessary to ensure high efficiency and high load-following capability in the operation of modern power plant. A nonlinear generalized predictive controller based on neuro-fuzzy network (NFGPC) is proposed in this paper. The proposed nonlinear controller is applied to control the superheated steam temperature of a 200MW power plant. From the experiments on the plant and the simulation of the plant, much better performance than the traditional controller is obtained. Keywords: Neuro-fuzzy networks; Generalized predictive control; Superheated steam temperature 1. Introduction Continuous process in power plant and power station are complex systems characterized by nonlinearity, uncertainty and load disturbance. The superheater is an important part of the steam generation process in the boiler-turbine system, where steam is superheated before entering the turbine that drives the generator. Controlling superheated steam temperature is not only technically challenging, but also economically important. From Fig.1,the steam generated from the boiler drum passes through the low-temperature superheater before it enters the radiant-type platen superheater. Water is sprayed onto the steam to control the superheated steam temperature in both the low and high temperature superheaters. Proper control of the superheated steam temperature is extremely important to ensure the overall efficiency and safety of the power plant. It is undesirable that the steam temperature is too high, as it can damage the superheater and the high pressure turbine, or too low, as it will lower the efficiency of the power plant. It is also important to reduce the temperature fluctuations inside the superheater, as it helps to minimize mechanical stress that causes micro-cracks in the unit, in order to prolong the life of the unit and to reduce 14 maintenance costs. As the GPC is derived by minimizing these fluctuations, it is amongst the controllers that are most suitable for achieving this goal. The multivariable multi-step adaptive regulator has been applied to control the superheated steam temperature in a 150 t/h boiler, and generalized predictive control was proposed to control the steam temperature. A nonlinear long-range predictive controller based on neural networks is developed into control the main steam temperature and pressure, and the reheated steam temperature at several operating levels. The control of the main steam pressure and temperature based on a nonlinear model that consists of nonlinear static constants and linear dynamics is presented in that. Fig.1 The boiler and superheater steam generation process Fuzzy logic is capable of incorporating human experiences via the fuzzy rules. Nevertheless, the design of fuzzy logic controllers is somehow time consuming, as the fuzzy rules are often obtained by trials and errors. In contrast, neural networks not only have the ability to approximate non-linear functions with arbitrary accuracy, they can also be trained from experimental data. The neuro-fuzzy networks developed recently have the advantages of model transparency of fuzzy logic and learning capability of neural networks. The NFN is have been used to develop self-tuning control, and is therefore a useful tool for developing nonlinear predictive control. Since NFN is can be considered as a network that consists of several local re-gions, each of which contains a local linear model, nonlinear predictive control based on NFN can be devised with the network incorporating all the local generalized predictive 15 controllers (GPC) designed using the respective local linear models. Following this approach, the nonlinear generalized predictive controllers based on the NFN, or simply, the neuro-fuzzy generalized predictive controllers (NFG-PCs)are derived here. The proposed controller is then applied to control the superheated steam temperature of the 200MW power unit. Experimental data obtained from the plant are used to train the NFN model, and from which local GPC that form part of the NFGPC is then designed. The proposed controller is tested first on the simulation of the process, before applying it to control the power plant. 2. Neuro-fuzzy network modelling Consider the following general single-input single-output nonlinear dynamic system: ''y(t),f[y(t,1),...,y(t,n),u(t,d),...,u(t,d,n，1), yu 'e(t,1),...,e(t,n)]，e(t)/, (1) e where f[.]is a smooth nonlinear function such that a Taylor series expansion exists, e(t)is a zero mean white noise andΔis the differencing '''n,n,noperator,and d are respectively the known orders and time delay of yue the system. Let the local linear model of the nonlinear system (1) at the operating pointbe given by the following Controlled Auto-Regressive o(t) Integrated Moving Average (CARIMA) model: ,1,d,1,1A(z)y(t),zB(z),u(t)，C(z)e(t) (2) ,1,1,1,1,1zWhereare polynomials in, the backward A(z),,A(z),B(z)andC(z) shift operator. Note that the coefficients of these polynomials are a function of o(t)the operating point.The nonlinear system (1) is partitioned into several operating regions, such that each region can be approximated by a local linear model. Since NFN is a class of associative memory networks with knowledge stored locally, they can be applied to model this class of nonlinear systems. A 16 schematic diagram of the NFN is shown in Fig.2.B-spline functions are used as the membership functions in the NFN for the following reasons. First, B-spline functions can be readily specified by the order of the basis function and the number of inner knots. Second, they are defined on a bounded support, and the output of the basis function is always positive, jji.e.,and.Third, the basis functions ,(x),0,x,[,,,],(x),0,x,[,,,]kj,kjkj,kj form a partition of unity, i.e., j (3) ,(x),1,x,[xx].,kmammin,j And fourth, the output of the basis functions can be obtained by a recurrence equation. Fig. 2 neuro-fuzzy network The membership functions of the fuzzy variables derived from the fuzzy rules can be obtained by the tensor product of the univariate basis functions. As an example, consider the NFN shown in Fig.2, which consists of the following fuzzy rules: xx IF operating condition i (is positive small, ... , andis negative large), n1 THEN the output is given by the local CARIMA model i: ˆˆˆ y(t),ay(t,1)，...，ay(t,n)，b,u(t,d)，... ii1iiniai0ia ，b,u(t,d,n)，e(t)，...，ce(t,n) (4) inibiinicbc ,,d,,111ˆor A(z)y(t),z,B(z)u(t)，C(z)e(t) iiiiii (5) 17 ,1,1,1Whereare polynomials in the backward shift A(z),B(z)andC(z)iii ,1u(t)operatorz, and d is the dead time of the plant,is the control, and i 2e(t)is a zero mean independent random variable with a variance of . The ,i multivariate basis functionis obtained by the tensor products of the a(x)ik univariate basis functions, ni a,,A(x),i,1,2,...,p,ikk,1k (6) where n is the dimension of the input vector x, and p, the total number of weights in the NFN, is given by, n p,(R，k) (7) ,ii,1k Where kand R are the order of the basis function and the number of ii inner knots respectively. The properties of the univariate B-spline basis functions described previously also apply to the multivariate basis function, which is defined on the hyper-rectangles. The output of the NFN is, p ˆya,iip,1iˆˆ (8) y,,ya,iip,1ia,i,1i 3. Neuro-fuzzy modelling and predictive control of superheated steam temperature ,Letbe the superheated steam temperature, and, the flow of spray water to ,, the high temperature superheater. The response ofcan be approximated by a , second order model: 18 K,p,,s (9) (),,Gse,()(，1)(，1)sTsTs12, The linear models, however, only a local model for the selected operating point. Since load is the unique antecedent variable, it is used to select the division between the local regions in the NFN. Based on this approach, the load is divided into five regions as shown in Fig.3,using also the experience of the operators, who regard a load of 200MW as high,180MW as medium high,160MW as medium,140MW as medium low and 120MW as low. For a ,1sampling interval of 30s, the estimated linear local models used in the A(z) NFN are shown in Table 1. Fig. 3 Membership function for local models Table 1 Local CARIMA models in neuro-fuzzy model Cascade control scheme is widely used to control the superheated steam temperature. Feed forward control, with the steam flow and the gas temperature as inputs, can be applied to provide a faster response to large variations in these two variables. In practice, the feed forward paths are 19 activated only when there are significant changes in these variables. The control scheme also prevents the faster dynamics of the plant, i.e., the spray water valve and the water/steam mixing, from affecting the slower dynamics of the plant, i.e., the high temperature superheater. With the global nonlinear NFN model in Table 1, the proposed NFGPC scheme is shown in Fig.4. Fig. 4 NFGPC control of superheated steam temperature with feed-for-ward control. As a further illustration, the power plant is simulated using the NFN model given in Table 1,and is controlled respectively by the NFGPC, the conventional linear GPC controller, and the cascaded PI controller while the load changes from 160MW to 200MW.The conventional linear GPC controller is the local controller designed for the“medium”operating region. The results are shown in Fig.5,showing that, as expected, the best performance is obtained from the NFGPC as it is designed based on a more accurate process model. This is followed by the conventional linear GPC controller. The performance of the conventional cascade PI controller is the worst, indicating that it is unable to control satisfactory the superheated steam temperature under large load changes. This may be the reason for controlling the power plant manually when there are large load changes. 20 Fig.5 comparison of the NFGPC, conventional linear GPC, and cascade PI controller. 4. Conclusions The modeling and control of a 200 MW power plant using the neuro-fuzzy approach is presented in this paper. The NFN consists of five local CARIMA models. The out-put of the network is the interpolation of the local models using memberships given by the B-spline basis functions. The proposed NFGPC is similarly constructed, which is designed from the CARIMA models in the NFN. The NFGPC is most suitable for processes with smooth nonlinearity, such that its full operating range can be partitioned into several local linear operating regions. The proposed NFGPC therefore provides a useful alternative for controlling this class of nonlinear power plants, which are formerly difficult to be controlled using traditional methods. 21 Part 4 为Part3译文: 锅炉蒸汽温度模糊神经网络的广义预测控制 Xiangjie LIU, Jizhen LIU, Ping GUAN 摘要:发电厂是非线性和不确定性的复杂系统。现代电厂的运行中，为确保电厂的高效率和高负荷的能力，准确的控制过热蒸汽温度是必要的。本文提出了一类在非线性广义预测控制器的基础上的模糊神经网络。所提出的非线性控制器适用于控制一台200 MW电厂的过热蒸汽温度。从实验 方案 气瓶 现场处置方案 .pdf气瓶 现场处置方案 .doc见习基地管理方案.doc关于群访事件的化解方案建筑工地扬尘治理专项方案下载 的仿真结果中可以看出，此方案的控制品质优于传统的控制方案。 关键词:模糊神经网络;广义预测控制;过热蒸汽温度 1. 引言 电厂过热汽温控制系统的特点是非线性、不确定性和负载扰动。蒸汽发电的过程中锅炉-汽轮机温度过热是一个重要的问题，蒸汽加热后，进入涡轮驱动发电机，控制过热蒸汽温度不仅是在技术上具有挑战性，在经济上的意义也是十分重要的。 22 图1 锅炉过热器和蒸汽生成过程 从图1可以看出，产生的蒸汽从锅炉汽包通过低温过热器后进入辐射型屏。水变成喷涂的蒸汽，以控制过热蒸汽的温度。适当的控制电厂过热蒸汽温度是极其重要的，可以确保整体效率和安全性。蒸汽温度太高对系统是非常不利的，因为过热蒸汽可以损害高压力汽轮机，太低也不行，因为它会降低电厂热效率。减少温度波动也是非常重要的，因为它有助于减少在单位内机械应力造成的微裂纹，延长单位秩序寿命，并减少维修成本。作为GPC的推导应该尽量减少这些波动，它是众多的控制器中最适合实现这一目标的。 多变量多步自适应调节已适用于控制过热蒸汽温度在150的锅炉，提出t/h了广义预测控制以控制蒸汽温度，基于神经网络发展的非线性预测控制器是以控制主蒸汽温度和压力。控制主蒸汽温度和压力的基础上，得到非线性模型的构成是由非线性静力常数和非线性动力学组成。 模糊逻辑是把人类的经验透过模糊规则 表 关于同志近三年现实表现材料材料类招标技术评分表图表与交易pdf视力表打印pdf用图表说话 pdf 现出来。然而，设计模糊逻辑控制器是非常消费时间的，由于模糊规则的不确定，往往得到的试验是错误的。在此相反，神经网络不仅有近似的非线性职能与任意精度，他们也可以有经过试验的实验数据。该模糊神经网络的开发优势是模型的透明度，模糊逻辑的准确度和具有学习能力的神经网络。该模糊神经网络已被用来发展自适应控制，因此，一个有用的工具，可以发展出非线性预测控制。从模糊神经网络可以考虑到作为一个网络构成的若干项，其中每一项包含一个局部线性模型，在非线性预测控制的基础上模糊神经网络可以设计和使用各自的地方线性模型把当地所有的广义预测控制器。按照这一办法，在非线性广义预测控制器的基础上，模糊神经网络简单地说是由该模糊神经网络的广义预测控制器推导出来的。建立控制器，然后应用于控制过热蒸汽温度的200MW机组。实验所得的数据，用来试验模糊神经网络 23 模型方案。 2. 模糊神经网络的建模 考虑以下的一般单输入单输出的非线性动态系统: '' y(t),f[y(t,1),...,y(t,n),u(t,d),...,u(t,d,n，1),yu ' (1) e(t,1),...,e(t,n)]，e(t)/,e 其中f[.]是一个光滑的非线性函数，将其按泰勒级数展开，其中是零均值、e(t) '''Δ是差分算子。和D分别是该系统已知的命令和延迟的时间。非线性系统n,n,nyue (1)在操作点的局部线性模型是由以下控制自动回归综合移动平均线o(t) (CARIMA)模型给出的: ,1,d,1,1 (2) A(z)y(t),zB(z),u(t)，C(z)e(t) ,1,1,1,1,1z其中是多项式在落后的移位算子。注意，A(z),,A(z),B(z)和C(z) 该系数多项式函数的转向点为o(t)。非线性系统(1)分割成为几个作业区域，如每个区域可以近似当成线性模型。自模糊神经网络是一类在本地的记忆网络的知识存储，他们可以应用到这一类非线性系统的模型中。由示意图可知该模糊神经网络的结果，图2中B样条函数作为隶属函数在模糊神经网络是由于以下几个原因:B样条功能可随时在指定的秩序的基础功能和数目内，第二，他们是界定 j在一个范围内的支持和输出的基础上功能始终是积极的，,(x),0,x,[,,,]kj,kj j和。第三，在职能的基础上形成一个分割的集体， ,(x),0,x,[,,,]kj,kj j (3) ,(x),1,x,[xx].,kmammin,j 第四，输出功能的基础上可以得到一个递归方程。 24 图2 模糊神经网络 从模糊规则所产生的模糊变量的隶属函数可以得到衍生的模糊规则。例如，考虑在图2所示的模糊神经网络，构成以下模糊规则: 如果操作条件是( 是正小……是负大) ，那么输出是给予的区域xxin1 CARIMA模型: ˆˆˆ y(t),ay(t,1)，...，ay(t,n)，b,u(t,d)，...ii1iiniai0ia (4) ，b,u(t,d,n)，e(t)，...，ce(t,n)inibiinicbc ,,d,,111ˆ或是 (5) A(z)y(t),z,B(z)u(t)，C(z)e(t)iiiiii ,1,1,1,1z其中是在落后的移位运算符的多项式，D是死去A(z),B(z)和C(z)iii 2时间，u(t)是控制量，e(t)是零均值与方差独立的随机变量。多变量的基础,ii 功能a(x)是获得由单变量函数的张量积的基础上得到的 ik ni a,,A(x),i,1,2,...,p (6) K,,ikkp,,s(),,Gse,1k,()(，1)(，1)sTsTs12, 其中n是输入向量x，和p的维数，在模糊神经网络中的比重，由下式给出， n p,(R，k) (7) ,ii,1k kR其中和分别是基函数秩序和内部结数。在单变量B样条基函数的性质前ii 面描述也适用于多元的基础功能。该NFN输出为 p ˆya,iipi,1ˆˆ y,,ya (8) ,iip,1ia,i,1i 3. 过热蒸汽温度的神经模糊塑造和有预测性的控制 ,设,是过热蒸汽温度，为水流喷射到高温过热器后的温度。对,响应可以, 近似为二阶模型 (9) 25 式(9)是线性模型，然而，只有选定的工作点局部模型。 由于负载的独特前件变量，它被用于选择在NFN的地方之间。 基于这种方法，如图3所示，使用操作员的经验， 被划分成五个区域，200MW作为一个高度， 180MW作为中等上流，160MW作为中等下流的方式、140MW和120MW为低和更低的方式。 对于一个30s的采样间隔，估计线性模型的一个区域使用的NFN如表1所示。 图3 模型的隶属函数 表1 神经模糊控制的局域CARIMA模型 串级控制方案被广泛用于控制过热蒸汽温度。前馈控制的蒸汽流量和气体温度为输入时，在这两个变量的变化时，可用于提供更迅速的反应。在实践中，只有是在这些变量有显着变化时，前馈路径才被激活。控制系统也防止了喷水阀门和水/蒸汽混合而影响设计的方案。 全球性非线性NFN模型在表1提出的NFGPC 计划 项目进度计划表范例计划下载计划下载计划下载课程教学计划下载 如下图4所示。 26 图4 NFGPC过热蒸汽温度控制与前馈控制 作为进一步的说明，该电厂是利用模拟模型的NFN由表1中给出，并受控于NFGPC，传统的线性GPC的控制器、PI控制器的级联分别为160MW到200MW的负荷变化。传统的线性GPC的区域控制器是为“中等”经营工作范围。设计的结果在图7所示，显示出预期最好的表现是从NFGPC获得的，因为它的目的是基于一个更准确过程的模型。这是由传统的线性级联GPC的常规PI控制器表现的。其次是表现最差的，表明它是不能在大负载变化的情况下令人满意的控制过热蒸汽温度的控制方案。这可能是手动控制的电厂有大时负荷变化的原因。 图5 控制策略仿真曲线比较 27 4. 结论 本文通过对一个200MW的电厂模型的建立和控制分析提出了模糊神经网络控制策略。该NFN由五个地方CARIMA模型组成。在之前提出的网络是利用了B样条基函数的局部模型给其赋值的。NFGPC同样被建立，它是由在NFN的CARIMA模型所设计的。该NFGPC是光滑非线性的，这样，它的整个工作范围可按局部线性合理的工作区域来划分。因此，为控制这种非线性发电厂而提出了NFGPC，用传统的控制方法是无法达到这样的控制 要求 对教师党员的评价套管和固井爆破片与爆破装置仓库管理基本要求三甲医院都需要复审吗 的。 28