外文文獻原稿和譯文原 稿Introduction In the modern industrial control field, along with the rapid development of computer technology, the emergence of a new trend of intelligent control, namely to machine simulation human thinking mode, using reasoning, deduce and induction, so the means, the production control, this is artificial intelligence. One expert system, fuzzy logic and neural network is the artificial intelligence of several key research hot spot. Relative to the expert system, the fuzzy logic belongs to the category of computational mathematics and contain the genetic algorithm, the chaos theory and linear theory etc, it comprehensive of operators practice experience, has the design is simple and easy to use, strong anti-interference ability and reaction speed, easy to control and adaptive ability, etc. In recent years, in a process control, built to touch, estimation, identify, diagnosis, the stock market forecast, agricultural production and military sciences to a wide range of applications. To carry out in-depth research and application of fuzzy control technology, the paper introduces the basic theory of fuzzy control technology and development, and to some in the application of the power electronics are introduced. Fuzzy Logic and Fuzzy Control 1, fuzzy logic and fuzzy control concept In 1965, the university of California, Berkeley, computer experts Lofty Zadeh put forward "fuzzy logic" concept, the root lies in the areas logic or clear logic distribution, used to define the confused, unable to quantify or the problem of precision, for in a mans von based on "true-false" reasoning mechanism, and thus create a electronic circuit and integrated circuit of the Boolean algorithm, fuzzy logic to fill the gaps in special things in sampling and analysis of blank. On the basis of fuzzy logic fuzzy set theory, a particular things as the set of features membership, he can be in "is" and "no" within the scope of the take between any value. And fuzzy logic is reasonable quantitative mathematical theory, the mathematical basis for fundamental for is to deal with these the statistical uncertain imprecise information. Fuzzy control based on fuzzy logic is a process of description of the control algorithm. For parameters precisely known mathematical model, we can use Berd graph or chart to analysts the Nyquist process to obtain the accurate design parameters. And for some complex system, such as particle reaction, meteorological forecast equipment, establishing a reasonable and accurate mathematical model is very difficult, and for power transmission speed of vector control problems, although it can be measured by the model that, but for many variables and nonlinear variation, the accurate control is very difficult. And fuzzy control technology only on the basis of the practical experience and the operator and intuitive inference, also relies on design personnel and research and development personnel of experience and knowledge accumulation, it does not need to establish equipment model, so basically is adaptive, and have strong robustness. After many years development, there have been many successful application of the fuzzy control theory of the case, such as Rutherford, Carter and Ostergaard were applied and metallurgical furnace and heat exchangers control device. 2, the analysis method is discussed Industrial control stability of the system is discussed the premise of the problem, because of the nonlinear and not to the unity of the description, make a judgment, so the fuzzy control system analysis method of stability analysis has been a hot spot, comprehensive in recent years you of scholars paper published the system stability analysis has these several circumstances : 1）, LiPuYa panov method: direct method based on the discrete time (D-T) and continuous time fuzzy control stability analysis and design method, the stability condition of the relative comparison conservative. 2）, sliding variable structure system analysis method 3）, round stability criterion methods: use sector bounded nonlinear concept, according to the stability criterion, led to the stability of the fuzzy control. 4）, POPOV criterion 5）, other methods such as relationship matrix analysis, exceed stable theory, phase-plane, matrix inequality or convex optimization method, fuzzy hole-hole mapping etc, detailed information and relevant literature many, in this one no longer etc. Set Design of Fuzzy Control The design of the fuzzy control is a very complicated process, in general, take the design steps and tools is more normative. The fuzzy controller general use of the special software and hardware, universal hardware chip in on the market at present is more, including main products are shown below. And special IC has developed very fast, it special IC and software controller integrates in together. In the process of design, the design of the general to take steps for: 1, considering whether the subject by fuzzy control system. That is considered the routine control mode of may. 2, from equipment operation personnel place to get as much information. 3 and selecting the mathematical model could, if use the conventional method design, estimate the equipment performance characteristics. 4, determine the fuzzy logic control object. 5, determine the input and output variables. 6, determine the variables as determined the belonging of the range. 7, confirm the variables of the corresponding rules. 8, determine the scale coefficients. 9, if have a ready-made, mathematical model of fuzzy controller with already certain of system simulation, observation equipment performance, and constantly adjust rules and scale coefficients until reaching satisfaction performance. Or to design fuzzy controller. 10, real-time operation controller, constantly adjust to the best performance. Fuzzy Control Application and Prospect As artificial intelligence of a new research field, the fuzzy control absorb lessons from the traditional design method and other new technologys essence, in many fields has made considerable progress. In the new type of power electronic and automatic control system, some experts in the linear adding the conditions of the power amplifier, the application of the fuzzy control based on the servo motor control, in the fuzzy control system with the PID and model reference adaptive control (MRAC) comparison proved the advantages of the method of fuzzy control. Fuzzy turn sent gain tuned controller views of the induction motor drive system vector control Fuzzy control as a is the development of new technology, now in most experts also to focus on application system research, and make considerable achievement, but in the theory research and system analysis or relative backward, so much so that some scholars have questioned its theoretical basis and effective. In view of this can be clear that the fuzzy control the combination of theory and practice is still needs to be further explored. The development prospects are very attractive, and in recent years, its theoretical study also made significant progress. In the past forty years of the development process, the fuzzy control also has some limitations: 1) control precision low, performance is not high, stability is poorer; 2) theory system is not complete. 3) the adaptive ability low. For these weaknesses, the fuzzy control and some other new technology, such as neural network (NN), genetic algorithm, and the combination of to a higher level of application development expand the huge space. Summary Fuzzy control as a comprehensive application example, in the global information the push of wave, in the next few decades, to the rapid development of economy will inject new vitality, the expert thinks, the next generation of industrial control is the basis of fuzzy control and neural network, and chaos theory as the pillar of the artificial intelligence. With the fuzzy control theory research and further more perfect of, the scope of application of the growing and supporting the development and manufacture of IC, the fuzzy control will be open to the field of industrial automation development of light application prospect, but also to the various areas of the researchers suggest more important task. 譯 文引言在現(xiàn)代工業(yè)控制領(lǐng)域，伴隨著計算機技術(shù)的突飛猛進，出現(xiàn)了智能控制的新趨勢，即以機器模擬人類思維模式，采用推理、演繹和歸納等手段，進行生產(chǎn)控制，這就是人工智能。其中專家系統(tǒng)邏輯和神經(jīng)網(wǎng)絡(luò)是人工智能的幾個重點研究熱點。相對于專家系統(tǒng)，模糊邏輯屬于計算數(shù)、模糊學(xué)的范疇，包含遺傳算法，混沌理論及線性理論等內(nèi)容，它綜合了操作人員的實踐經(jīng)驗，具有設(shè)計簡單，易于應(yīng)用、抗干擾能力強、反應(yīng)速度快、便于控制和自適應(yīng)能力強等優(yōu)點。近年來，在過程控制、建摸、估計、辯識、診斷、股市預(yù)測、農(nóng)業(yè)生產(chǎn)和軍事科學(xué)等領(lǐng)域得到了廣泛應(yīng)用。為深入開展模糊控制技術(shù)的研究應(yīng)用，本文綜合介紹了模糊控制技術(shù)的基本理論和發(fā)展?fàn)顩r，并對一些在電力電子領(lǐng)域的應(yīng)用作了簡單介紹。模糊邏輯與模糊控制1.模糊邏輯與模糊控制的概念1965年，加州大學(xué)伯克利分校的計算機專家Lofty Zadeh提出“模糊邏輯”的概念，其根本在于區(qū)分布爾邏輯或清晰邏輯，用來定義那些含混不清，無法量化或精確化的問題，對于馮諾依曼開創(chuàng)的基于“真假”推理機制，以及因此開創(chuàng)的電子電路和集成電路的布爾算法，模糊邏輯填補了特殊事物在取樣分析方面的空白。在模糊邏輯為基礎(chǔ)的模糊集合理論中，某特定事物具有特色集的隸屬度，他可以在“是”和“非”之間的范圍內(nèi)取任何值。而模糊邏輯是合理的量化數(shù)學(xué)理論，是以數(shù)學(xué)基礎(chǔ)為為根本去處理這些非統(tǒng)計不確定的不精確信息。模糊控制是基于模糊邏輯描述的一個過程的控制算法。對于參數(shù)精確已知的數(shù)學(xué)模型，我們可以用Berd圖或者Nyquist圖來分析家其過程以獲得精確的設(shè)計參數(shù)。而對一些復(fù)雜系統(tǒng)，如粒子反應(yīng)，氣象預(yù)報等設(shè)備，建立一個合理而精確的數(shù)學(xué)模型是非常困難的，對于電力傳動中的變速矢量控制問題，盡管可以通過測量得知其模型，但對于多變量的且非線性變化，起精確控制也是非常困難的。而模糊控制技術(shù)僅依據(jù)與操作者的實踐經(jīng)驗和直觀推斷，也依靠設(shè)計人員和研發(fā)人員的經(jīng)驗和知識積累，它不需要建立設(shè)備模型，因此基本上是自適應(yīng)的，具有很強的魯棒性。歷經(jīng)多年發(fā)展，已有許多成功應(yīng)用模糊控制理論的案例，如Rutherford，Carter 和Ostergaard分別應(yīng)用與冶金爐和熱交換器的控制裝置。2.分析方法探討工業(yè)控制系統(tǒng)的穩(wěn)定性是探討問題的前提，由于難以對非線性和不統(tǒng)一的描述，做出判斷，因此模糊控制系統(tǒng)的分析方法的穩(wěn)定性分析一直是一個熱點，綜合近年來各位學(xué)者的發(fā)表的論文,目前系統(tǒng)穩(wěn)定性分析有以下集中:1), 李普亞諾夫法:基于直接法的離散時間(D-T)和連續(xù)時間模糊控制的穩(wěn)定性分析和設(shè)計方法,相對而言起穩(wěn)定條件比價保守。2),滑動變結(jié)構(gòu)系統(tǒng)分析法3),圓穩(wěn)定性判據(jù)方法:利用扇區(qū)有界非線性概念,根據(jù)穩(wěn)定判據(jù)可推導(dǎo)模糊控制的穩(wěn)定性.4),POPOV判據(jù)5),其他方法如關(guān)系矩陣分析法，超穩(wěn)定理論，相平面法，矩陣不等式或凸優(yōu)化法，模糊穴穴映射等，詳細資料及有關(guān)文獻很多，在這里不再一一闡述。模糊控制的設(shè)置設(shè)計模糊控制的設(shè)計是一個非常復(fù)雜的過程，一般而言，采取的設(shè)計步驟和工具比較規(guī)范。其中模糊控制器一般采用專用軟硬件，通用型的硬件芯片在目前市場上比較多，其中主流產(chǎn)品如下表所示。而專用IC發(fā)展也很迅速，它把專用IC和軟件控制器集成在一起。設(shè)計過程中，一般采取的設(shè)計步驟為：1，綜合考慮該課題能否采用模糊控制系統(tǒng)。即考慮采用常規(guī)控制方式的可能。2,從設(shè)備操作人員處獲取盡可能多的信息。3，選取可能的數(shù)學(xué)模型，如果用常規(guī)方法設(shè)計，估計設(shè)備的性能特點。4，確定模糊邏輯的控制對象。5，確定輸入輸出變量。6，確定所確定的各個變量的歸屬范圍。7，確定各變量的對應(yīng)規(guī)則。8，確定比例系數(shù)。9，如果有現(xiàn)成的數(shù)學(xué)模型，用已確定的模糊控制器對系統(tǒng)仿真，觀測設(shè)備性能，并不斷調(diào)整規(guī)則和比例系數(shù)直到達到滿意性能。否則重新設(shè)計模糊控制器。10，實時運行控制器，不斷調(diào)整以達到最佳性能。模糊控制應(yīng)用與前景展望作為人工智能的一種新研究領(lǐng)域，模糊控制吸收借鑒了傳統(tǒng)設(shè)計方法和其他新技術(shù)的精華，在諸多領(lǐng)域取得了長足的進展。在新型的電力電子和自動控制系統(tǒng)中，有些專家在線性功放的加設(shè)條件下，把模糊控制應(yīng)用于為基礎(chǔ)的伺服電機控制中，在把模糊控制系統(tǒng)與PID及模型參考自適應(yīng)控制（MRAC）進行比較后證明了模糊控制方法的優(yōu)越性。模糊控制作為一項正在發(fā)展的新技術(shù)，目前在大多數(shù)專家還把主要精力放在應(yīng)用系統(tǒng)研究上，并取得了相當(dāng)?shù)某晒?，但在理論研究和系統(tǒng)分析上還是相對落后的，以至于一些學(xué)者質(zhì)疑其理論依據(jù)和有效性。鑒于此可以明確得知：模糊控制理論和實踐的結(jié)合仍有待于進一步探索。其發(fā)展前景是十分誘人的，而且在近年來，其理論研究也取得了顯著進展。在近四十年的發(fā)展進程中，模糊控制也有一些局限性：1、控制精度低，性能不高，穩(wěn)定性較差；2、理論體系不完整；3、自適應(yīng)能力低。對于這些弱點，模糊控制與一些其他新技術(shù)，比如神經(jīng)網(wǎng)絡(luò)（NN），遺傳算法相結(jié)合，向更高層次的應(yīng)用發(fā)展拓展了巨大的空間?？偨Y(jié)模糊控制作為一門綜合應(yīng)用范例，在全球信息化浪潮的推動下，在未來的幾十年中，必將對經(jīng)濟的迅猛發(fā)展注入新的活力，有專家認(rèn)為，下一代工控的基礎(chǔ)是模糊控制，神經(jīng)網(wǎng)絡(luò)，混沌理論為支柱的人工智能。隨著模糊控制理論研究的日益完善和深入，應(yīng)用范圍的日益擴大和配套IC的研發(fā)制造，模糊控制將給工控領(lǐng)域的發(fā)展開辟光明的應(yīng)用前景，同時也給各領(lǐng)域的研究人員提出了更重大的任務(wù)。7