起重機(jī)變幅液壓系統(tǒng)設(shè)計(jì)【帶CAD圖紙】,帶CAD圖紙,起重,機(jī)變,液壓,系統(tǒng),設(shè)計(jì),cad,圖紙 IMPLEMENTATION OF HYDRAULIC SERVO CONTROLLERS WITH ONLYPOSITION MEASURE Abstract Hydraulic actuators have nonlinear dynamics and are often used in environments (robotic, aerospace, underwater explo-rations/inspections, mining) where uncertain disturbances are present. Linear controllers designed using a linearized model of the hydraulic system are widespread. In alternative, nonlinear and ro-bust control techniques can be used to achieve better performances. Among these techniques, sliding mode control with dynamics inversion is a good choice, but it usually requires measurements of actuator’s velocity and hydraulic pressures in addition to actu-ator’s position. This paper presents the design and experimental evaluation of a position controller for an hydraulic actuator where the only available measure is the actuator’s position. A detailed description of servosystem components is firstly presented. Then a linear control law, whose design is based on a linearized model of the actuator, is designed and tested. Finally, a sliding mode control law is developed. Experimental results, carried out on a real case study, have shown the ectiveness of the proposed controllers even when only actuator’s position is available for feedback. Key Words :Hydraulic actuators nonlinear systems nonlinear control system sliding mode control. 1. Introduction Hydraulic actuators are used in many industrial appli-cations as they over the following advantages: ness; compactness; payload capability; high immunity to wear thanks to lubricant action of fluid; high speed of response, with fast starts, stops and speed reversals; ability to main-tain their loading capacity indefinitely, while this would usually cause excessive heat generation in electrical com-ponents [1, 2]. Furthermore, their high power-to-weight ratio allows their use in a direct-drive manner, as e.g., in industrial robots, so that wear-sensitive gear-boxes can be avoided. ? DIIIE, Universit`a degli Studi di Salerno, Via Ponte Don Melillo, 1-84084, Fisciano(SA), Italy; e-mail: {fbasile, ddel-grosso}@unisa.it Recommended by Prof. Zhihua Qu (paper no. 206-3128) One of their major drawbacks is their strongly non-linear behaviour. The main nonlinearities are: magnetic hysteresis in the armature of the servovalve driving the actuator, usually neglected [3, 4]; static nonlinear relation between the control input and the flow to the actuator; ori-fices tolerances (overlap/underlap), which generate dead-zones; pressure and temperature dependency of isothermal bulk modulus; asymmetry of the hydraulic cylinder [5]; friction force acting on the actuator [6]. Another ma-jor drawback is the dis culty of accurately estimating the model’s parameters, and their variations under dis erent operating conditions. Therefore, to design a servo posi-tion system for an hydraulic actuator nonlinear and robust control techniques have to be used to achieve good per-formances. Today more complex control laws, like inverse dynamics laws, are successfully used in many applications and have been recently applied also to hydraulic servosys-tems [4, 7–10]. It is important to point out that in all these works control laws are presented which rely on mea-surement of at least actuators’s position and velocity and hydraulic pressures. Figure 1. Control architecture’s overview. The hydraulic actuator considered in this work is com-posed of a flapper-nozzle type two-stage servovalve and an asymmetric hydraulic cylinder. A servo position system has to be devised. First, an accurate model of the actu-ator has been written obtained which takes into account the main phenomena influencing its behaviour. Dynamic parameters were partly available by the supplier. To vali-date these parameters and identify the others, it was not possible to procede with open-loop experiments but it was mandatory for safety requirements to move the actuator only if closed-loop controlled. To this purpose, a discrete-time linear controller has been first designed based on a linearized model of the actuator. Then, a sliding mode con-trol law has been designed in the continuous-time domain by using a Lyapunov-based approach and implemented in discrete-time without requiring direct measurements of ve-locity and pressures. As it is commonly accepted in prac-tice [11] for control engineers the sliding mode controller designed in continuous-time domain has been implemented in the discrete-time by selecting a ciently fast sampling rate. The model uncertainties, the fact that pressures are not directly measured and frictions are not compensated, and the time delays make the chattering reduction crit-ical [11]. To avoid this undesirable phenomenon [12] a boundary layer approach has been adopted.The main contribution of this paper is to prove that low-cost robust control of hydraulic actuators is possible even if not all the required direct measurements are avail-able. In fact, the proposedcontrollers requires only direct measurement of actuator’s position. 2. Experimental Setup 2.1 Control Architecture The hydraulic actuator taken into account in this work is a linear single-rod cylinder whose piston’s position is measured by a Linear Variable Dis erential Transformer (LVDT) sensor. The digital control architecture, shown in Fig. 1, has the following features: PowerPC CPU running at 500 MHz; The sampling rate is 500 Hz; 12 bit A/D and D/A converters; Voltage-to-current and current-to-voltage converters with a bandwidth of 10 KHz; LVDT Current Conditioningand Transmitter (LCCT) with a bandwidth of 250 Hz; A 100 Hz second-order Butterworth anti-aliasing filter; Hydraulic system (the actuator) including an LVDT sensor for position measuring. Notice that The only available measure is piston’s position; The digital implementation and the presence of limited bandwidth circuits and filters makes possible to neglect all components over 100 Hz; The presence of quantizers produces a measurement noise, in fact the measures resolution is 2.3 · 10?5 m; The actuator has to work in a wide and variable range of temperatures. 2.2 Implementation Issues Experimental tests on the real architecture have drawn the attention to some important aspects for the design of both the linear and the robust controller. 1. Time delays on feedback chain: A time delay on the feedback chain of about 6 ms has been measured which heavily influences the dynamics and must be taken into account in designing the controllers. 2. Noise in the position’s measure: The LVDT sensor’s non-ideal behaviour, joint with the presence of quan-tizers, produces a not-negligible measurement noise. Then, it is necessary to verify the filtering capacity of the controllers, to avoid vibrations of the actuator. 3. Unavailable measures: The only available measure is the actuator’s position. Thus, the proposed sliding mode controller works without the chambers’ pressure measures as commonly required. Actuator’s velocity and acceleration have instead been obtained from the position measure by using a derivative filter (see Fig. 2) for which the considered transfer function is: Filter’s parameters have been tuned to realize a deriva-tive behaviour up to 100 Hz without amplifying high-frequency components. Figure 2. Re-constructor filter. 3. Model Description The structure of the electro-hydraulic servoactuator, com-posed of a nozzle-flapper flow control servovalve and an asymmetric hydraulic actuator, is shown in Fig. 3 where the scale factors for the valve and the actuator are erent for sake of clarity. All the symbols reported in Fig. 3 are defined in Table 1 and will be used in the model description below. Figure 3. Hydraulic servoactuator layout. 3.1 Servovalve Model 3.1.1 Servovalve Nonlinear Model The nonlinear model of the nozzle-flapper valve [1, 2] can be represented by the scheme in Fig. 4 where its main elements are shown. The input is the electrical current of the torque motor I which is transformed in a torque at the armature: As the rotation of the armature is very small, a simplified linear relation is often considered: The equation describing the armature flapper’s dy-namics relating armature torque to flapper displacement (in case of small rotation) is: The equations describing nozzle’s flows relating flapper displacement to flows on spool’s sides are: 液壓伺服控制器的執(zhí)行位置測(cè)量 摘 要 液壓執(zhí)行器有非線(xiàn)性動(dòng)力特性，并且經(jīng)常在機(jī)械，航空航天，水下檢驗(yàn)環(huán)境中使用，檢測(cè)其中的不確定干擾存在。線(xiàn)性控制器的設(shè)計(jì)普遍采用線(xiàn)性模型的液壓系統(tǒng)。如果使用非線(xiàn)性和魯棒控制技術(shù)作為替代可以實(shí)現(xiàn)更好的性能。在這些技術(shù)中，采用動(dòng)態(tài)逆滑模控制是一個(gè)很好的選擇，但它通常需要除了實(shí)際位置的測(cè)量還需要執(zhí)行器的速度和液壓壓力測(cè)量。本文提出的是一個(gè)液壓制動(dòng)器的位置控制器設(shè)計(jì)的實(shí)驗(yàn)。一個(gè)詳細(xì)說(shuō)明的伺服系統(tǒng)組件首先被提出。然后，是基于線(xiàn)性模型的執(zhí)行機(jī)構(gòu)的線(xiàn)性控制法的設(shè)計(jì)和測(cè)試。最后，研制了一種滑?？刂坡?。實(shí)驗(yàn)結(jié)果，在一個(gè)真實(shí)的案例研究，證明所提出控制器即使當(dāng)只有致作動(dòng)器的位置是可用于反饋的時(shí)候依然有效。 關(guān)鍵詞 ： 液壓作動(dòng)器 非線(xiàn)性系統(tǒng) 非線(xiàn)性控制系統(tǒng) 滑?？刂? 第一章 介紹 液壓執(zhí)行器有許多工業(yè)應(yīng)用因?yàn)橐韵聝?yōu)點(diǎn)：剛性，緊湊性，載荷能力，由于液體流動(dòng)的高抗磨損性；響應(yīng)速度快，具有快速啟動(dòng)，停止和速度逆轉(zhuǎn)；保持無(wú)限容量裝載的能力，但這通常會(huì)導(dǎo)致電氣部件過(guò)熱。此外，他們的高功率重量比允許他們使用直接驅(qū)動(dòng)的方式，例如，在工業(yè)機(jī)器人中，這樣可以避免磨損敏感的齒輪箱。 其主要缺點(diǎn)之一是其強(qiáng)烈的非線(xiàn)性行為。主要的非線(xiàn)性是：伺服閥驅(qū)動(dòng)致動(dòng)器電樞中的磁滯，這經(jīng)常被忽視；在輸入控制和執(zhí)行器流動(dòng)中的靜態(tài)非線(xiàn)性關(guān)系；產(chǎn)生死區(qū)的孔公差；壓力和等溫體積彈性模量的溫度依賴(lài)性；液壓缸的不對(duì)稱(chēng)；作用在致動(dòng)器摩擦力。另一個(gè)主要的缺點(diǎn)是難以準(zhǔn)確估計(jì)模型的參數(shù)和不同試驗(yàn)條件下的變化的參數(shù)。 因此，為了為液壓致動(dòng)器技術(shù)的非線(xiàn)性設(shè)計(jì)一個(gè)伺服位置系統(tǒng)，需要應(yīng)用魯棒控制技術(shù)以獲得更好的性能。今天，更復(fù)雜的控制規(guī)律，如逆動(dòng)力學(xué)規(guī)律，成功地應(yīng)用在許多應(yīng)用程序并且最近也適用于液壓伺服系統(tǒng)。需要指出的是，在所有這些工作的控制律的提出依靠執(zhí)行器位置和液壓壓力的測(cè)量。 這項(xiàng)工作中的液壓執(zhí)行機(jī)構(gòu)是由噴嘴擋板式二級(jí)電液伺服閥和非對(duì)稱(chēng)液壓缸組成。一個(gè)伺服位置系統(tǒng)必須設(shè)計(jì)。 首先，已經(jīng)提到執(zhí)行器的精確模型需要考慮到那些影響其性能的主要現(xiàn)象。一部分動(dòng)態(tài)參數(shù)可從供應(yīng)商獲得。為了驗(yàn)證這些參數(shù)并確定其他參數(shù)，進(jìn)行開(kāi)環(huán)試驗(yàn)是不可能的，安全需求當(dāng)閉環(huán)控制去移動(dòng)執(zhí)行器是強(qiáng)制性的。為了這個(gè)目的，基于線(xiàn)性模型的執(zhí)行器首先設(shè)計(jì)出了一個(gè)離散時(shí)間線(xiàn)性控制器。然后，一個(gè)滑動(dòng)模式控制法已被設(shè)計(jì)在連續(xù)時(shí)間域通過(guò)使用一個(gè)Lyapunov為基礎(chǔ)的方法，在離散時(shí)間實(shí)現(xiàn)而不需要速度和壓力的直接測(cè)量。通過(guò)選擇一個(gè)快速采樣率，在連續(xù)時(shí)間域設(shè)計(jì)的控制工程師的滑動(dòng)模式控制器在離散時(shí)間實(shí)施的設(shè)計(jì)在實(shí)踐中普遍被接受。模型的不確定性，事實(shí)上，壓力并不是直接測(cè)量和摩擦沒(méi)有補(bǔ)償，并且時(shí)間的延遲使抖振臨界降低。為了避免這種不良現(xiàn)象，采納邊界層的方法。 圖1 控制架構(gòu)的概述 第二章 實(shí)驗(yàn)設(shè)置 2.1控制結(jié)構(gòu) 這項(xiàng)工作中的液壓執(zhí)行機(jī)構(gòu)是一個(gè)線(xiàn)性單桿缸，線(xiàn)性單桿缸活塞的位置由一個(gè)線(xiàn)性可變地微分變壓器（LVDT）傳感器測(cè)量。數(shù)字控制結(jié)構(gòu)，有以下特點(diǎn)： ?powerpc CPU運(yùn)行在500 MHz； ?采樣率是500 Hz； ?12位A / D和D / A轉(zhuǎn)換器； ?10 kHz帶寬的電壓-電流和電流-電壓轉(zhuǎn)換器； ?250Hz帶寬的電流調(diào)節(jié)和發(fā)射機(jī)； ?100Hz的二階巴特沃斯抗混疊濾波器； ?液壓系統(tǒng)包括一個(gè)LVDT傳感器的位置測(cè)量 注意 ?唯一的變量是活塞的位置； ?數(shù)字實(shí)現(xiàn)和有限的帶寬的電路存在和過(guò)濾器可以忽略所有組件在100赫茲； ?量化器產(chǎn)生的測(cè)量噪聲的存在，事實(shí)上，測(cè)量分辨率為2.3 · 10?5米； ?執(zhí)行器在一個(gè)廣泛的和可變的溫度范圍內(nèi)工作。 2.2實(shí)施問(wèn)題 關(guān)于真實(shí)結(jié)構(gòu)實(shí)驗(yàn)測(cè)試，注意一些重要的關(guān)于線(xiàn)性和魯棒控制器的設(shè)計(jì)方面問(wèn)題 1.時(shí)間延遲反饋鏈：一個(gè)約6毫秒的已嚴(yán)重影響動(dòng)態(tài)的時(shí)間反饋鏈延遲被檢測(cè)出測(cè)量，而且這是設(shè)計(jì)控制器必須考慮的問(wèn)題。 2.位置測(cè)量的噪聲：LVDT傳感器的非理想特性，與量化器的連接處，產(chǎn)生一個(gè)不可忽略的測(cè)量噪聲。然后，為了避免致動(dòng)器的振動(dòng)，驗(yàn)證控制器的過(guò)濾能力是很重要的。 3.不變量測(cè)量：可用的唯一變量是執(zhí)行器的位置。因此，建議的滑動(dòng)模式控制器在沒(méi)有腔內(nèi)壓力作為一般工作條件進(jìn)行工作。執(zhí)行機(jī)構(gòu)的速度和加速度通過(guò)使用微分濾波器測(cè)量得到，這種方式定義為傳遞函數(shù)： 濾波器的參數(shù)在沒(méi)有放大高頻組件的情況下已經(jīng)實(shí)現(xiàn)導(dǎo)數(shù)的行為上升到100赫茲 第三章 模型描述 電液伺服執(zhí)行器的結(jié)構(gòu)，由一個(gè)噴嘴擋板式流量控制閥和一個(gè)非對(duì)稱(chēng)液壓作動(dòng)器，如圖3所示。圖中閥門(mén)和執(zhí)行器為了看得清楚，調(diào)大了比例。所有的符號(hào)在圖3報(bào)告表1所示，將用下面的模型描述： 圖3 液壓伺服作動(dòng)器布局 3.1伺服閥模型 3.1.1伺服閥的非線(xiàn)性模型 噴嘴擋板閥的非線(xiàn)性模型可以表示在圖4中，其主要內(nèi)容是所示的方案。輸入的轉(zhuǎn)矩電動(dòng)機(jī)我這電流轉(zhuǎn)化成扭矩在電樞： 當(dāng)電樞旋轉(zhuǎn)是非常小的，通常被認(rèn)為是一個(gè)簡(jiǎn)化的線(xiàn)性關(guān)系： 電樞擋板的動(dòng)力學(xué)與電樞轉(zhuǎn)矩?fù)醢逦灰品匠蹋ㄔ谛〗嵌刃D(zhuǎn)的情況下）是： 噴嘴的流量與擋板位移對(duì)閥芯的側(cè)流方程： 12