National InstrumentsメーカーNI MATRIX Xmath Robust Control Moduleの使用説明書/サービス説明書
ページ先へ移動 of 71
NI MA TRIXx TM Xmath TM Robust Control Module MA TRIXx Xmath Robust Control M odule April 2007 370757C-01.
Support Worldwide Technical Support and Product Info rmation ni.com National Instruments Corporate Headquarters 11500 North Mopac Expressway Aust in, Texas 78759-3504 USA Tel: 512 683 0100 Worldwide O.
Important Information Warranty The media on which you receive National In struments software are warranted not t o fail to execute pr ogramming instructions, due to defects in materials and workmanship, for a period of 90 days from date of shipment, as eviden ced by receipts or other documentation.
Conventions The follo wing conv entions are used in this manual: [ ] Square brackets enclose op tional it ems—for example, [ response ]. » The » symbol leads you th rough nested menu items and dial og box options to a final action.
© National Instruments Corporation v MATRIXx Xmath Robust Control Module Contents Chapter 1 Introduction Using This Manual...................... ......................... ....................... ......................... .......... 1-1 Document Organization.
Contents MATRIXx Xmath R obust Control Mo dule vi ni.com Chapter 3 System Evaluation Singular Value Bode Plots.............. ...................... .......................... ...................... ......... 3-1 L Infinity Norm (linfnorm) ............
© National Instruments Corporation 1-1 MA TRIXx Xmath Robust Co ntrol Module 1 Introduction The Xmath Robust Control Module (RCM) provi des a collection of analysis and synthesis tools that assist in the design of robust control systems. This chapter starts with an outline of the manual and some use notes.
Chapter 1 Introduction MA TRIXx Xmath Robust Control Modul e 1-2 ni.com techniques. The general problem setup is explained together with known limitations; the rest is left to the references. Bibliographic References Throughout this document, biblio graphic references are cited with bracketed entries.
Chapter 1 Introduction © National Instruments Corporation 1-3 MA TRIXx Xmath Robust Co ntrol Module • Xmath Optimization Module • Xmath Robust Control Module • Xmath X μ Module MA TRIXx Help Robust Control Modul e function reference informatio n is available in the MATRIXx Help .
Chapter 1 Introduction MA TRIXx Xmath Robust Control Modul e 1-4 ni.com Figure 1-1. RCM Function Structure Many RCM functions are based on stat e-of-the-art algorithms impl emented in cooperation with researchers at Stanford Uni versity . The robustness analysis functions are based on struct ured singular v alue calculations.
© National Instruments Corporation 2-1 MA TRIXx Xmath Robust Co ntrol Module 2 Robustness Analysis This chapter describes RCM tools used for analyzing the robustness of a closed-loop system. The chapter a ssumes that a controller has been designed for a nominal plant and that the closed-loop performance of this nominal system is acceptable.
Chapter 2 Robustness Analysis MA TRIXx Xmath Robust Control Modul e 2-2 ni.com system, including how the un certain transfer functions are connected to the system and the magnitude bound fun ctions l i ( w ).
Chapter 2 Robustness Analysis © National Instruments Corporation 2-3 MA TRIXx Xmath Robust Co ntrol Module Stability Margin (smargin) Assume that the nominal cl osed-loop system is stable .
Chapter 2 Robustness Analysis MA TRIXx Xmath Robust Control Modul e 2-4 ni.com smargin( ) marg = smargin(SysH, delb {s caling, graph}) The smargin( ) function plots an approximatio n to the stability margin of the system as a function of frequency . For a full discussion of smargin( ) syntax, refer to the MATRIXx Help .
Chapter 2 Robustness Analysis © National Instruments Corporation 2-5 MA TRIXx Xmath Robust Co ntrol Module Figure 2-3. SI SO T racking Syst em with Three Uncertaintie s The H system will ha ve the reference i nput as input1 and the error outpu t as output1 ( w and z , respectiv ely , in Figure 2-2).
Chapter 2 Robustness Analysis MA TRIXx Xmath Robust Control Modul e 2-6 ni.com Figure 2-4. Bound for Sensor Uncertainty Note A value of l 3 at one radian per second of –20 dB indicates that modeling uncertainties of up to 10% (–20 dB = 0.1) are allowed.
Chapter 2 Robustness Analysis © National Instruments Corporation 2-7 MA TRIXx Xmath Robust Co ntrol Module Figure 2-5. Stability Margin Now e xamine the effect on the st ability margin of discretizing H ( s ) at 100 Hz.
Chapter 2 Robustness Analysis MA TRIXx Xmath Robust Control Modul e 2-8 ni.com W orst-Case Performance Degradation (wcbode) Even if a system is robustly stable, th e uncertain transfer functions still can have a great effect on performance. Co nsider the transfer function from the q th input, w q , to th e p th output, z p .
Chapter 2 Robustness Analysis © National Instruments Corporation 2-9 MA TRIXx Xmath Robust Co ntrol Module wcbode( ) [WCMAG, NOMMAG] = wcbode (Sy sH, delb, {input, output, graph}) The wcbode( ) function computes and plots the worst-case gain of a closed-loop transfer function.
Chapter 2 Robustness Analysis MA TRIXx Xmath Robust Control Modul e 2-10 ni.com Figure 2-6. Performance Degradation of the SISO T rac king System Advanced T opics This section describes the theoretical background on robust ness analysis and performance degradation.
Chapter 2 Robustness Analysis © National Instruments Corporation 2-11 MA TRIXx Xmath Robust Control Modu le for all diagonal Δ such that where μ ( . ) is the structured singular value , introduced by Do yle in [Doy82]. Thus, the margin is the in verse of the structured singular value of H qr diagonally scaled by the magnitude bo unds.
Chapter 2 Robustness Analysis MA TRIXx Xmath Robust Control Modul e 2-12 ni.com Y ou can compare this mar g in to that of the example in the Creating a Nominal Syst em section; the following inputs produce Figure 2-7.
Chapter 2 Robustness Analysis © National Instruments Corporation 2-13 MA TRIXx Xmath Robust Control Modu le of generality—so, roughly speaking, it can be solved.
Chapter 2 Robustness Analysis MA TRIXx Xmath Robust Control Modul e 2-14 ni.com Comparing Scaling Algorithms Using the system from the first ex ample (Figure 2-3), you can compare the results of using.
Chapter 2 Robustness Analysis © National Instruments Corporation 2-15 MA TRIXx Xmath Robust Control Modu le ssv( ) [v,vD] = SSV(M, {scaling}) The ssv( ) function computes an approx imation (and gu aranteed upper bound) to the Scaled Singular V alue of a complex square matrix M , where M can be a reducible matrix.
Chapter 2 Robustness Analysis MA TRIXx Xmath Robust Control Modul e 2-16 ni.com VOPT=ssv(M,{scaling="OPT"}) VOPT (a scalar) = 2.43952 VSVD = max(svd(M)) VSVD (a scalar) = 2.65886 osscale( ) [v, vD] = osscale(M) The osscale( ) function scales a matrix using the Osborne Algorit hm.
Chapter 2 Robustness Analysis © National Instruments Corporation 2-17 MA TRIXx Xmath Robust Control Modu le optscale( ) [v, vOPTD] = optscale (M, {t ol}) The optscale( ) function optimally scales a matri x. An iterativ e optimization (ellipsoid) algorithm which calculates upper and lower bounds on the left side of Equation 2-5 is used.
Chapter 2 Robustness Analysis MA TRIXx Xmath Robust Control Modul e 2-18 ni.com Figure 2-10. Reduction to Separat e Systems In terms of the approximations to th e margin discussed abov e, this reducibility will manifest itself as a pro blem such as di vide-by-zero or nontermination.
Chapter 2 Robustness Analysis © National Instruments Corporation 2-19 MA TRIXx Xmath Robust Control Modu le Using this relation and any of the previously discussed appro ximations for μ ( . ), you can compute an approximation to wcgain( ) . Because the approximations to μ ( .
© National Instruments Corporation 3-1 MA TRIXx Xmath Robust Co ntrol Module 3 System Evaluation This chapter describes system analysis functions that create singu lar value Bode plots, performance plots, and calculate the L ∞ norm of a linear system.
Chapter 3 System E valuation MA TRIXx Xmath Robust Control Modul e 3-2 ni.com Refer to [BoB91 ] in Appendix A, Bibliography . Example 3-1 Creating a Singular Value Plot 1. Let a system H be a 2-input/2-output system: tf=makepoly([1,2],"s")/.
Chapter 3 System Ev aluation © National Instruments Corporation 3-3 MA TRIXx Xmath Robust Co ntrol Module Figure 3-1. Sing ular Value Plot L Infinity Norm (linfnorm) The L ∞ norm of a stable transfer matrix H i s defined as: where is the maximu m singular value and H ( j ω ) is the transfer matr ix under consideration.
Chapter 3 System E valuation MA TRIXx Xmath Robust Control Modul e 3-4 ni.com factor b y wh ich the RMS value of a signal flo wing th rough H can be increased. By comparison, the H 2 norm is defined as: This norm can be interpreted as th e RMS value of the output when the input is unit intensity whit e noise.
Chapter 3 System Ev aluation © National Instruments Corporation 3-5 MA TRIXx Xmath Robust Co ntrol Module •I f A has an imaginary eigen value at j ω 0 , linfno rm( ) retu rns: vOMEGA = SIGMA = Infinity where ω 0 is one of the imaginary eigen values of A .
Chapter 3 System E valuation MA TRIXx Xmath Robust Control Modul e 3-6 ni.com Figure 3-2. Singular Values of H ( j ω) as a Function of ω Note sv is returned in dBs. Check that sigma is within 0.01 (the default value of tol ) of 10**(max(sv,{channels})/20) .
Chapter 3 System Ev aluation © National Instruments Corporation 3-7 MA TRIXx Xmath Robust Co ntrol Module Singular V alue Bode Plots of Subsystems To evaluate the performance achieved by a given cont.
Chapter 3 System E valuation MA TRIXx Xmath Robust Control Modul e 3-8 ni.com The four transfer matrices are labeled e / d , e / n , u / d , and u / n in the f inal plot. The plots in the top ro w , consisting of e / d and e / n , show the regulation or tracking achie ved by the controller .
Chapter 3 System Ev aluation © National Instruments Corporation 3-9 MA TRIXx Xmath Robust Co ntrol Module The system matrix can be calculated using the afeedback( ) function for different v alues of K .
Chapter 3 System E valuation MA TRIXx Xmath Robust Control Modul e 3-10 ni.com Figure 3-5. Per fplots( ) for K = 1 and K =5 clsys( ) SysCL = clsys( Sys, SysC ) The clsys( ) function computes th e state-space realization SysCL , of the closed-loop system from w to z as sho wn in Figu re 3-6.
Chapter 3 System Ev aluation © National Instruments Corporation 3-11 MA TRIXx Xmath Robust Control Modu le Where SysC=system(Ac,Bc,Cc,Dc ) , Sys=system(A,B,C,D) , and nz is the dimension of z and nw is the dimension of w : Given th e ab ove, SysCL is calculated as shown in Figure 3-7.
Chapter 3 System E valuation MA TRIXx Xmath Robust Control Modul e 3-12 ni.com Figure 3-8. I ll-Posed Feedback System Example 3-4 Example of Closed-Loop System a = 1; b = [1,0,1]; c = b'; d = [0,0,0;0,0,1;0,1,0]; Sys = SYSTEM(a,b,c,d); SysC = SYSTEM(-40,2.
© National Instruments Corporation 4-1 MA TRIXx Xmath Robust Co ntrol Module 4 Controller Synthesis This chapter discusses synthe sis tools in two categories, H ∞ and H 2 . This chapter does not explain all of the theory of H ∞ , LQG/LTR, and frequency shaped LQG design techniques.
Chapter 4 Controller Synthe sis MA TRIXx Xmath Robust Control Modul e 4-2 ni.com The functio n hinfcontr( ) can be used to find an optimal H ∞ controller K that is arbitrarily close to solving: (4-2.
Chapter 4 Controller Synthesis © National Instruments Corporation 4-3 MA TRIXx Xmath Robust Co ntrol Module Equi valently , as a transfer matrix: T o enter the extended system , you must kno w the sizes of e and w shown in Figure 4-1. The extended plant P can be cons tructed using the Xmath interconnection functio ns, as shown in Example 4-1.
Chapter 4 Controller Synthe sis MA TRIXx Xmath Robust Control Modul e 4-4 ni.com The transfer matrix G can be viewed as a model of the underlying system dynamics with v and u as generalized forces that produce ef fects in the performance signals z and measured signals y .
Chapter 4 Controller Synthesis © National Instruments Corporation 4-5 MA TRIXx Xmath Robust Co ntrol Module here the weighting matrices are tran sfer matrices, whereas in the LQG setup they are constants.
Chapter 4 Controller Synthe sis MA TRIXx Xmath Robust Control Modul e 4-6 ni.com Selecting these weigh ts has much the same effect he re. Specif ically , let H zv be the closed-loop transfer matr ix (with u = K γ ) from inputs: to outputs: Thus, Suppose that the controller u = K y approximates Eq uation 4-2.
Chapter 4 Controller Synthesis © National Instruments Corporation 4-7 MA TRIXx Xmath Robust Co ntrol Module where and The weights also can be viewed as “design knobs” (for example, [ONR84]).
Chapter 4 Controller Synthe sis MA TRIXx Xmath Robust Control Modul e 4-8 ni.com •F o r a l l ω ≥ 0, • Condition 1 is a standard cond ition to ensure the existence of a stabilizing controller . Condition 2 en sures that the control signal u is contained in the normalized error v ector e (refer to Figure 4-3).
Chapter 4 Controller Synthesis © National Instruments Corporation 4-9 MA TRIXx Xmath Robust Co ntrol Module If no error message occurs, then is guaranteed. Ho wever , this does not preclude the po ssibility that either or that . For the former c ase, there are two checks: •U s e t h e linfnorm( ) function to compute .
Chapter 4 Controller Synthe sis MA TRIXx Xmath Robust Control Modul e 4-10 ni.com Suppose the i nput/output we ights are as follows: 2. Create the four weights: Wdist = 1/makepoly([1,1],"s" ) Wdist (a transfer function) = 1 ----- s + 1 Wnoise = 0.
Chapter 4 Controller Synthesis © National Instruments Corporation 4-11 MA TRIXx Xmath Robust Control Modu le 4. For this e xample, you will start with gamma=1 as the initial guess and enter: [K,Hew] = hinfcontr(P,1,2,2) ; No error messages are reported.
Chapter 4 Controller Synthe sis MA TRIXx Xmath Robust Control Modul e 4-12 ni.com Figure 4-5. Perfplots for H ew It also is useful to perform perfplots( ) on the unweighted closed-loop system, H zv , wh ich in this case is the closed-loop transfer matrix fr om ( d , n ) into ( x , u ).
Chapter 4 Controller Synthesis © National Instruments Corporation 4-13 MA TRIXx Xmath Robust Control Modu le Figure 4-6. Per fplots for H zv singriccati( ) [P, solstat] = singriccati(A ,Q,R {method}).
Chapter 4 Controller Synthe sis MA TRIXx Xmath Robust Control Modul e 4-14 ni.com Linear -Quadratic-Gaussian Control Synthesis The H 2 Linear-Quadratic-Gaussi an (LQG) control design methods are based on minimizing a quadratic functi on of stat e variables and control inputs.
Chapter 4 Controller Synthesis © National Instruments Corporation 4-15 MA TRIXx Xmath Robust Control Modu le This expression can be con verted into the following form [Gu80]: If R ( j ω ) is not a funct ion of frequency , then C 12 = 0 and D = I . Note The system has a new input v and the old input u is now the output of the system.
Chapter 4 Controller Synthe sis MA TRIXx Xmath Robust Control Modul e 4-16 ni.com fsesti( ) [SysF, vEV] = fsesti(SysA, n s, QWWA, QVVA, {QWVA}) The fsesti( ) function computes a freque ncy-shaped state estimator .
Chapter 4 Controller Synthesis © National Instruments Corporation 4-17 MA TRIXx Xmath Robust Control Modu le fslqgcomp( ) [SysCC, vEV] = fslqgcomp(Sys F, SysC) The fslqgcomp( ) function combines f ilter an d control law to compute a controller from a control law and an estimator .
Chapter 4 Controller Synthe sis MA TRIXx Xmath Robust Control Modul e 4-18 ni.com -0.500025 + 0.866011 j -0.500025 - 0.866011 j 5. T ry t he LQG compensator w ith the full-or der system: [Syscl_fo]=feedback(Sys,Sysc ); poles(Syscl_fo) ans (a column vector) = -0.
Chapter 4 Controller Synthesis © National Instruments Corporation 4-19 MA TRIXx Xmath Robust Control Modu le 0 0 0 1 0 0 0 0 B 0 0 0 1 C 0 0 1 0 D 0 X0 0 0 0 0 System is continuous 7. Frequency-weight the control signal. T ransfer the weight on U from RUU to the third di agonal entry in RXXA .
Chapter 4 Controller Synthe sis MA TRIXx Xmath Robust Control Modul e 4-20 ni.com System is continuous fs_evr (a column vector) = -0.645263 + 0.587929 j -0.
Chapter 4 Controller Synthesis © National Instruments Corporation 4-21 MA TRIXx Xmath Robust Control Modu le 9. Design the LQG compensator . [Sysfs_sc,fs_evc]=fslqgcomp( Sysfs_se,Sysfs_sr) Sysfs_sc (a state space syst em) = A 0 1 0 0 -1 -1.00005 1 0 0 0 0 1 0.
Chapter 4 Controller Synthe sis MA TRIXx Xmath Robust Control Modul e 4-22 ni.com 10. Compute the closed-loop system fo r the reduced order plant and the frequency-shaped compensator: [Sysfs_scl]=feedback(Sysr,Sy sfs_sc); poles(Sysfs_scl) ans (a column vector) = -0.
Chapter 4 Controller Synthesis © National Instruments Corporation 4-23 MA TRIXx Xmath Robust Control Modu le Figure 4-8. LQG Feedbac k System for Loop T ransfer Recovery lqgltr( ) [SysC,EV,Kr] = lqgl.
Chapter 4 Controller Synthe sis MA TRIXx Xmath Robust Control Modul e 4-24 ni.com Then ρ is increased so that pointwise in s : Regulator reco very is only guaranteed if G ( s ) is minimum-phase and there are at least as many control signals u as measurements y .
© National Instruments Corporation A -1 MA TRIXx Xmath Robust Control M odule A Bibliography [BBK88] S. Boyd, V . Balakrishnan , and P . Kaba mba. “ A bisectio n method for com puting the L ∞ norm of a transfer matrix and related problems. ” Mathematical Control Sign als, Systems V ol.
Appendix A B ibliograph y MA TRIXx Xmath Robust Control Modul e A-2 ni.com [FaT88] M.K. Fan and A.L. Tits. “m-form Nu merical Range and the Computation of the Structured Singular V alue. ” IEEE Transactions on Auto matic Control , V ol. 33, pp 284 –289, March 198 8.
Appendix A B ibliograph y © National Instruments Corporation A -3 MA TRIXx Xmath Robust Control M odule [SA88] G. Stein and M. At hans. “The LQG/ L TR Procedure for Multivariable Control Design. ” IEEE Transacti ons on Automatic Control , V ol. A C-32 , No.
© National Instruments Corporation B -1 MA TRIXx Xmath Robust Control M odule B T echnical Support and Professional Ser vices Visit the following sections of the National Instruments Web site at ni.com for technical support an d professional services: • Support —Online technical support resources at ni.
© National Instruments Corporation I-1 MA TRIXx Xmath Robust Co ntrol Module Index A Algebraic Riccati Equation (ARE), 4-13 C clsys( ), 3-10 conventions used in the manual, iv D diagnostic tools (NI .
Index MA TRIXx Xmath Robust Control Modul e I-2 ni.com nominal transfer function, 2-8 norm H 2 , 3-4 L ∞ , 3-3 O optscale( ), 2-17 osscale( ), 2-16 P perfplots( ), 3-7 pfscale( ), 2-16 programming e.
デバイスNational Instruments NI MATRIX Xmath Robust Control Moduleの購入後に（又は購入する前であっても）重要なポイントは、説明書をよく読むことです。その単純な理由はいくつかあります：
National Instruments NI MATRIX Xmath Robust Control Moduleをまだ購入していないなら、この製品の基本情報を理解する良い機会です。まずは上にある説明書の最初のページをご覧ください。そこにはNational Instruments NI MATRIX Xmath Robust Control Moduleの技術情報の概要が記載されているはずです。デバイスがあなたのニーズを満たすかどうかは、ここで確認しましょう。National Instruments NI MATRIX Xmath Robust Control Moduleの取扱説明書の次のページをよく読むことにより、製品の全機能やその取り扱いに関する情報を知ることができます。National Instruments NI MATRIX Xmath Robust Control Moduleで得られた情報は、きっとあなたの購入の決断を手助けしてくれることでしょう。
National Instruments NI MATRIX Xmath Robust Control Moduleを既にお持ちだが、まだ読んでいない場合は、上記の理由によりそれを行うべきです。そうすることにより機能を適切に使用しているか、又はNational Instruments NI MATRIX Xmath Robust Control Moduleの不適切な取り扱いによりその寿命を短くする危険を犯していないかどうかを知ることができます。
ですが、ユーザガイドが果たす重要な役割の一つは、National Instruments NI MATRIX Xmath Robust Control Moduleに関する問題の解決を支援することです。そこにはほとんどの場合、トラブルシューティング、すなわちNational Instruments NI MATRIX Xmath Robust Control Moduleデバイスで最もよく起こりうる故障・不良とそれらの対処法についてのアドバイスを見つけることができるはずです。たとえ問題を解決できなかった場合でも、説明書にはカスタマー・サービスセンター又は最寄りのサービスセンターへの問い合わせ先等、次の対処法についての指示があるはずです。
