MPEMathematical Problems in Engineering1563-51471024-123XHindawi Publishing Corporation21205310.1155/2009/212053212053Research ArticleComputation of All Stabilizing PID Gain for Second-Order Delay SystemFarkhRihemLaabidiKaoutherKsouriMekkiHeJi HuanResearch Unit on Systems Analysis and ControlNational School Engineering of TunisBP, 37, Le Belvédère, 1002 TunisTunisiaenit.rnu.tn20090809200920090204200919062009150720092009Copyright © 2009This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The problem of stabilizing a second-order delay system using classical proportional-integral-derivative (PID) controller is considered. An extension of the Hermite-Biehler theorem, which is applicable to quasipolynomials, is used to seek the set of complete stabilizing PID parameters. The range of admissible proportional gains is determined in closed form. For each proportional gain, the stabilizing set in the space of the integral and derivative gains is shown to be either a trapezoid or a triangle.

1. Introduction

Dead times are often encountered in various engineering systems and industry processes such as electrical and communication network, chemical process, turbojet engine, nuclear reactor, and hydraulic system. In fact, delays are caused by many phenomena like the time required to transport mass, energy or information, the time processing for sensors, the time needed for controllers to execute a complicated algorithm control, and the accumulation of time lags in a number of simple plants connected in series . Delay makes system analysis and control design much more complex which is explained as follows. First, the effect of the control action takes some time to be felt in the controlled variable. Second, the control action that is applied based on the actual error tries to correct a situation that originated previously. Finally, in the frequency domain, the time delay introduces an extra decrease in the systems phase, which may cause instability [1, 2].

Today, proportional-integral (PI) and proportional-integral-derivative (PID) controller types are the most widely used control strategy. It is estimated that over 90% of process control applications employ PID control thanks to its essential functionality and structural simplicity . Since the minimal requirement for PID controller is to guarantee the system stability. It is desirable to know the complete set of stabilizing PID parameters before tuning and design. Surveys have reported that poor tuning, configuration errors, traditional and empirical techniques such as Ziegler Nichols or Cohen-Coon tuning methods are found in the industrial application of PID . A great deal of academic and industrial effort has focused on improving PID control in the areas of tuning rules to decrease the rising gap between engineering practice and control theory. Recent studies have made use of the generalization of Hermite-Biehler theorem to compute the set of all stabilizing PID controllers for a given plant . Since almost all plants encountered in process control contain time delays, so computing the complete set of PID controllers that stabilize time delay system is of considerable importance. A robust PI/PID design via numerical optimization for delay processes was proposed in . Padma Sree et al.  propose a PI/PID controllers design for first-order delay system by extracting the coefficients of the numerator and denominator of the closed-loop transfer function. In the work presented in , the authors have developed a graphical approach which is based on D-partition method , the borders of absolute and relative stability regions are described in the parameter space. A unified approach and Delta operator are used to obtain unified stability boundaries for PI, PD, and PID controller for an arbitrary order delay system; where the boundaries can be found when only the frequency response and not the parameters of the plant are known . Eriksson and Johansson  consider a design of PID controller for system with varying time delay using multi-objectives’ optimization. Neural network has been widely applied for nonlinear dynamical system identification via PID controller . Zhang et al.  present an intelligent nonlinear PID controller based on the recurrent neural networks (RNNs) to control multivariable plants by using two predictive control schemes. In [14, 15], the authors propose a new PID neural networks (PIDNN) which is a dynamic multilayer network based on P, I, and D neurons. This PIDNN is used to control multivariable plants  and has also the ability to control time delay system . In , Roy and Iqbal have explored PID tuning of first-order delay system using a first-order Padé approximation and the Hermite-Biehler stabilization framework. The set’s characterization of all stabilizing PI/PID parameters using a version of the Hermit-Biehler theorem for first-order delay system is presented in [5, 1720]. However, these results are not applicable to the second-order delay system. Although, the PID stabilization for a given time delay plant has been developed  using the generalized Nyquist criterion, the stabilizing set of the proportional gain cannot be found. In , the characterization of the set of all stabilizing P/PI/PID parameters is given by using the Hermit-Biehler Theorem for polynomials for a class of time delay system which verify the interlacing property at high frequencies. This method, however, is complex and there are difficulties to achieve the stabilizing range of proportional parameter. As shown in [23, Example  2], this latter approach does not determine the precise range of proportional gain which is larger than the exact stabilizing range given by . Furthermore, a known stabilizing value of (Kp,Ki,Kd) is first determined using Nyquist criterion, then the entire stability region is established. In our work, the stabilizing problem of PID controller for second order delay system is analyzed using the extended Hermit-Biehler theorem for quasi-polynomials. The exact range of stabilizing proportional gain is first determined. The set of stabilizing integral and derivative constant values are then derived. Our algorithm is more simple and faster in time computing than the other presented in .

2. Preliminary Results for Analyzing Time Delay System

Several problems in process control engineering are related to the presence of delays. These delays intervene in dynamic models whose characteristic equations are of the following form [5, 25]

δ(s)=d(s)+e-L1sn1(s)+e-L2sn2(s)++e-Lmsnm(s), where: d(s) and ni(s) are polynomials with real coefficients and Li represent time delays. These characteristic equations are recognized as quasi-polynomials. Under the following assumptions:

deg(d(s))=n,deg(ni(s))<n,for  i=1,2,,m,

L1<L2<<Lm.

One can consider the quasi-polynomials δ*(s) described by

δ*(s)=esLmδ(s),δ*(s)=esLmd(s)+es(Lm-L1)n1(s)+es(Lm-L2)n2(s)++nm(s). The zeros of δ(s) are identical to those of δ*(s)  since esLm does not have any finite zeros in the complex plane. However, the quasi-polynomial δ*(s) has a principal term since the coefficient of the term containing the highest powers of s and es is nonzero. If δ*(s)  does not have a principal term, then it has an infinity roots with positive real parts .

The stability of the system with the characteristic equation (2.1) is equivalent to the condition that all the zeros of δ*(s)  must be in the open left half of the complex plan. We said that δ*(s)  is Hurwitz or is stable. The following theorem gives a necessary and sufficient condition for the stability of δ*(s).

Theorem 2.1 (see [<xref ref-type="bibr" rid="B8">5</xref>]).

Let δ*(s)  be given by (2.2), and write δ*(jω)=δr(ω)+jδi(ω), where δr(ω) and δi(ω) represent, respectively, the real and imaginary parts of δ*(jω).

Under conditions (A1) and (A2), δ*(s) is stable if and only if

δr(ω) and δi(ω) have only simple, real roots and these interlace,

δi(ω0)δr(ω0)-δi(ω0)δr(ω0)>0  for some ω0  in [-,+],

where δi(ω) and δr(ω) denote the first derivative with respect to ω of δr(ω) and δi(ω), respectively.

A crucial stage in the application of the precedent theorem is to make sure that δr(ω) and δi(ω) have only real roots. Such a property can be checked while using the following theorem.

Theorem 2.2 (see [<xref ref-type="bibr" rid="B8">5</xref>]).

Let M and N designate the highest powers of s and es which appear in δ*(s). Let η be an appropriate constant such that the coefficient of terms of highest degree in δr(ω) and δi(ω) do not vanish at ω=η. Then a necessary and sufficient condition that δr(ω)and δi(ω) have only real roots is that in each of the intervals -2lπ+η  <ω<  2lπ+η,l=l0,l0+1,l0+2

δr(ω) or δi(ω) have exactly 4lN+M real roots for a sufficiently large l0.

3. PID Control for Second Order Delay System

A second order system with delay can be mathematically expressed by a transfer function having the following form:

G(s)=Ks2+a1s+a0e-Ls, where K is the static gain of the plant, L is the time delay, and a0 and a1 are the plant parameter’s which are always positive. The characteristic equation of the closed-loop system is given by

δ(s)=K(Ki+Kps+Kds2)e-Ls+(s2+a1s+a0)s, we deduce the quasi-polynomial δ*(s):

δ*(s)=eLsδ(s)=K(Ki+Kps+Kds2)+s(s2+a1s+a0)eLs, by replacing s by jω, we get

δ*(jω)=δr(ω)+jδi(ω), with:

δr(ω)=KKi-KKdω+(ω3-a0ω)sin(Lω)-a1ω2cos(Lω),δi(ω)=w[KKp+(a0-ω2)cos(Lω)-a1ωsin(Lω)].

Clearly, the parameters Ki and Kd only affect the real part of δ*(jω) whereas the parameter Kp affects the imaginary part.

Let’s put z=Lω, we get

δr(z)=KKi-KKdz2L2+sin(z)(z3L3-a0zL)-a1z2L2cos(z),δi(z)=zL(KKp+cos(z)(a0-z2L2)-a1zLsin(z)).

The application of the second condition of Theorem 2.1, led us to

E(z0)=δi(z0)δr(z0)-δi(z0)δr(z0)>0, from (3.6) we have

δi(z)=KKpL-sin(z)(a0+2a1zL2-z3L3)+cos(z)(a0L-3z2L3-a1z2L2), for z0=0 (a value that annul δi(z)), and we get:

E(z0)=δi(z0)δr(z0)=(KKp+a0L)KKi>0, which implies Kp>-a0/K  since K>0 and Ki>0.

We pass to the verification of the interlacing condition of δr(z) and δi(z) roots. For such purpose, we are going to determine the roots from the imaginary part, which is translated by the following relation:

δi(z)=0z=0orKKp+cos(z)(a0-z2L2)-a1zLsin(z)=0,z=0orKKp+cos(z)(a0-z2L2)=a1zLsin(z),z=0orf(z)=g(z). We notice the z0=0 is a trivial root of the imaginary part. The others are difficult to solve analytically. However, this can be made graphically. The following two cases are presented.

Case 1 (<inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M88"><mml:msub><mml:mrow><mml:mi>K</mml:mi></mml:mrow><mml:mrow><mml:mi>p</mml:mi></mml:mrow></mml:msub><mml:mo>></mml:mo><mml:mrow><mml:mrow><mml:mo>-</mml:mo><mml:msub><mml:mrow><mml:mi>a</mml:mi></mml:mrow><mml:mrow><mml:mn>0</mml:mn></mml:mrow></mml:msub></mml:mrow><mml:mo>/</mml:mo><mml:mrow><mml:mi>K</mml:mi></mml:mrow></mml:mrow></mml:math></inline-formula>).

In this case, we graph the curves of f(z) and of g(z) which are presented in Figure 1. z1[0,π/2],z3[3π/2,2π],z5[7π/2,4π],z2[π/2,π],z4[5π/2,3π],z6[9π/2,5π], that is, zj verifies z1[0,π2],zj[(2j-3)π2,(j-1)π],for  j2.

Representation of the curves of f(z) and of g(z) (K=L=a1=1, a0=2, Kp=1.3, and Ku=1.5884) (Case: Kp<Ku).

Case 2 (<inline-formula><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M101"><mml:msub><mml:mrow><mml:mi>K</mml:mi></mml:mrow><mml:mrow><mml:mi>p</mml:mi></mml:mrow></mml:msub><mml:mo>≥</mml:mo><mml:msub><mml:mrow><mml:mi>K</mml:mi></mml:mrow><mml:mrow><mml:mi>u</mml:mi></mml:mrow></mml:msub></mml:math></inline-formula>).

Figure 2 represents the case where the curves of f(z)=KKu+cos(z)(a0-(z2/L2)) and of g(z) are tangent in [0,π], where Ku is the largest number of Kp.

The plot in Figure 3 corresponds to the case where Kp>Ku and the plot of f(z) does not intersect g(z) twice in the interval [0,π].

Representation of the curves of f(z) and of g(z) (K=L=a1=1, a0=2, Kp=Ku=1.5884) (Case: Kp=Ku).

Representation of the curves of f(z) and of g(z) (K=L=a1=1, a0=2, Kp=2, and Ku=1.5884 (Case: Kp>Ku).

Theorem 2.2 is used to verify that δi(z) possess only simple roots. By replacing Ls by s1 in (2.2), we rewrite δ*(s) as follows:

δ*(s)=eLsδ(s)=es1δ(s1)=es1((s1L)3+a1(s1L)2+a0s1L)+K(Kps1L+Ki). For this new quasi-polynomial, we see that M=3 and N=1 where M and N designate the highest powers of s and es which appear in δ*(s). We choose η=π/4  that satisfies the condition giving by theorem 3 as δr(η)0 and δi(η)0. According to Figure 1, we notice that for -a0/K<Kp<Ku, δi(z) possess four roots in the interval [0,2π-π/4]=[0,7π/4] including the root at origin. As δi(z) is odd function of z, so it possesses seven roots in [-2π+π/4,2π-π/4]=[-7π/4,7π/4]. Hence, we can affirm that δi(z) has exactly 4N+M=7  in [-2π+π/4,2π+π/4]=[-7π/4,9π/4]. In addition, it can be shown that δi(z)  has two real roots in each of the intervals [2lπ+π/4,2(l+1)π+π/4] and [-2(l+1)π+π/4,-2lπ+π/4] for l=1,2,. It fallows that δi(z) has exactly 4lN+M  real roots in [-2lπ+π/4,2lπ+π/4] for -a0/K<Kp<Ku. At the end, according to Theorem 2.2, δi(z) has only real roots for every Kp in [-a0/K,Ku]. For KpKu, corresponding to Figures 2 and 3, the roots of δi(z)  are not real. We pass to determine the superior value of Kp. According to the definition of Ku, if Kp=Ku, then the curves of f(z)  and g(z) are tangent in the point α. Which is translated by

KKu+cos(α)(a0-α2L2)=a1αLsin(α),ddz[KKu+cos(z)(a0-z2L2)]z=α=ddz[a1zLsin(z)]z=α,-2αcos(α)(1+a1L)+sin(α)(α2-a0L2-a1L)=0tan(α)=α(2+a1L)(α2-a0L2-a1L), once α is determined, the parameter Ku is expressed by the relation:

Ku=1K(a1αLsin(α)-cos(α)(a0-α2L2)).

Theorem 3.1.

The δi(z) have real roots if and only if: -a0K<Kp<1K(a1αLsin(α)-cos(α)(a0-α2L2)), where α is the solution of the equation tan(α)=α(2+a1L)/(α2-a1L-a0L2) in the interval [0,π].

After determination of the roots of the imaginary part δi(z), we pass to the evaluation of these roots by the real part δr(z)

δr(z)=KKi-KKdz2L2+sin(z)(z3L3-a0zL)-a1z2L2cos(z)=Kz2L2[-Kd+KiL2z2+LKz[-a1zLcos(z)+sin(z)(z2L2-a0)]]=Kz2L2[-Kd+m(z)Ki+b(z)], where:

m(z)=L2z2,b(z)=LKz[-a1zLcos(z)+sin(z)(z2L2-a0)]. From condition 1 of Theorem 2.1, the roots of δr(z) and δi(z) have to interlace for δ*(s) to be stable. We evaluate δr(z) at the roots of the imaginary part δi(z).

For z0=0, we have

δr(z0)=KKi>0, for zjz0, where j=1,2,3,, we get:

δr(zj)=Kzj2L2[-Kd+m(zj)Ki+b(zj)]=Kzj2L2[-Kd+mjKi+bj], where:

m(zj)=mj,b(zj)=bj. Interlacing the roots of δr(z) and δi(z) is equivalent to δr(z0)>0 (since Ki>0), δr(z1)<0,δr(z2)>0 .

We can use the interlacing property and do it as δi(z) which has only real roots to reach that δr(z) possess real roots too.

From the previous equations we get the following inequalities:

{δr(z0)>0δr(z1)<0δr(z2)>0δr(z3)<0δr(z4)>0{Ki>0Kd>m1Ki+b1Kd<m2Ki+b2Kd>m3Ki+b3Kd<m4Ki+b4 Thus intersecting all these regions in the (Ki,Kd) space, we get the set of (Ki,Kd) values for which the roots of δr(z) and δi(z) interlace for a fixed value of Kp. We notice that all these regions are half planes with their boundaries being lines with positive slopes mj.

Example 3.2.

Consider the plant given by relation (3.1) with the following parameters K=L=a1=1 and a0=2: G(s)=1s2+s+2e-s. The imaginary part δi(z) has only simple real roots if and only if -1<Kp<1.58.

We set the controller parameter Kp to 1.3, which is inside the previous range. For this Kp, δi(z) takes the form: z(1.3+cos(z)(2-z2)-zsin(z))=0, We next compute some of the positive real roots of this equation and arrange them in increasing order of magnitude: z0=0,z1=1.3608,z2=1.8905,z3=4.9829,z4=7.9619,z5=11.0976,z6=14.2017. Using (3.18) we calculate the parameters mj and bj for j=1,,6: {m1=0.54,m2=0.2798,m3=0.0403,m4=0.0158,{b1=0.54,b2=-0.3150,b3=1.1047,b4=-4.6822. Interlacing the roots of the real and the imaginary part occurs for Kp=1.3, if and only if the following inequalities are satisfied: Ki>0,Kd>0.54Ki-0.3150,Kd<0.2798Ki+1.1047,Kd>0.0403Ki-4.6822,Kd<0.0158Ki+7.7736. The boundaries of these regions are illustrated in Figure 4.

In this case, the stability region is defined by only two boundaries: Kd=m1Ki+b1,Kd=m2Ki+b2, because we have the following inequalities: bj<bj+2,forj=2,4,6,,bj>bj+2,forj=1,3,5,7,,mj<mj+2,forj1.

Region boundaries of Example 3.2.

Example 3.3.

Consider the plant (3.1) with the following parameters K=2,a0=3,a1=1,L=2:G(s)=2s2+s+3e-2s. The imaginary part δi(z) has only simple real roots if and only if -1.5<Kp<0.98.

We now set the controller parameter Kp to 0.5, which is inside the previous range. For this Kp, δi(z)takes the form z(0.5+cos(z)(2-z24)-0.5zsin(z))=0. We next compute some of the positive real roots of this equation and arrange them in increasing order of magnitude z0=0,z1=1.6480,z2=2.9830,z3=5.4560,z4=8.077,z5=11.22,z6=14.26,z7=17.41. Using (3.18), we calculate the parameters mj and bj for j=1,,7:  {m1=1.4728,m2=0.4495,m3=0.1344,m4=0.0613,m5=0.0318,m6=0.0197,m7=0.0132,{b1=-1.3656,b2=0.4527,b3=-0.9377,b4=1.7176,b5=-2.5853,b6=3.3906,b7=-4.2097. Interlacing the roots of the real and imaginary part occurs for Kp=0.5, if and only if the following inequalities are satisfied: Ki>0,Kd>1.4728Ki-1.3656,Kd<0.4495Ki+0.4527,Kd>0.1344Ki-0.9377,Kd<0.0613Ki+1.7176,Kd>0.0318Ki-2.5853,Kd<0.0197Ki+3.3906,Kd>0.0132Ki-4.2097. The boundaries of these regions are illustrated in Figure 5.

In this case, the stability region is defined by only two boundaries: Kd=m1Ki+b1,Kd=m2Ki+b2,Kd=m3Ki+b3, because we have the following inequalities: bj<bj+2,forj=2,4,6,,bj>bj+2,forj=3,5,7,,b1<b3,mj<mj+2,forj1.

Region boundaries of Example 3.3.

Plots of δi(z) (dotted line), (-a1(z/L)) cos(z) (thick solid line) and sin(z)(z2/L2-a0) (thin solid line).

As pointed out in Examples 3.2 and 3.3, the inequalities given by (3.22) represent half planes in the space of Ki and Kd. Their boundaries are given by lines with the following equations:

Kd=mjKi+bj,forj=1,2,3,. We notice that the difference between Examples 3.2 and 3.3 resides in change of the sign of (b1-b3) which is explained by the localization of z1 and z3 in different intervals according to (3.12).

We now state an important technical lemma that allows us to develop an algorithm for solving the PID stabilization problem. This lemma shows the behavior of the parameter bj, j=2,3,4, for different values of the parameter Kp inside the range proposed by Theorem 3.1.

Lemma 3.4.

If -a0K<Kp<1K(a1αLsin(α)-cos(α)(a0-α2L2)), where α is the solution of the equation tan(α)=α(2+a1L)/(α2-a1L-a0L2) in the interval [0,π].Then:

bj<bj+2forj=2,4,6,,

bj>bj+2forj=3,5,7,,

mj>mj+2forj1,andlimjmj=0.

Proof.

From (3.18) we have m(zj)=L2zj2,0<zj<zj+2L2zj2>L2zj+22,forj1, then m(zj)>m(zj+2),limjm(zj)=0.

Remark 3.5.

As we can see from Figure 1, the odd roots of δi(z), that is, zj for j=3,5,7,, verify zj[(2j-3)π/2,(j-1)π] and are getting closer to (2j-3)π/2 as j increase, so we have: limjcos(zj)=0,limjsin(zj)=-1. Moreover, since cosine function and sine function are monotonically increasing between (2j-3)π/2 and (j-1)π, we have {cos(z3)>cos(z5)>cos(z7)>sin(z3)>sin(z5)>sin(z7)>{cos(zj)>cos(zj+2)>0,sin(zj+2)<sin(zj)<0.

Remark 3.6.

As we can see from Figure 1, the even roots of δi(z) that is, zj for j=2,4,6, verify zj[(2j-3)π/2,(j-1)π] and are getting closer to (2j-3)π/2 as j increase, so we have: limj cos(zj)=0,limjsin(zj)=1. Moreover, since cosine function and sine function are monotonically decreasing between (2j-3)π/2 and (j-1)π, we have: {cos(z2)<cos(z4)<cos(z6)<sin(z2)<sin(z4)<sin(z6)<{cos(zj)<cos(zj+2)<0,0<sin(zj)<sin(zj+2).

The change of b(zj) can be found with graphical approach. Given a stabilizing Kp value inside the admissible range by using Theorem 3.7, the curve of δi(z) is plotted to obtain their roots denoted by zj,j=1,2,3,, the curves of (-a1(z/L))cos(z) and sin(z)(z2/L2-a0) are plotted too in order to understand the behavior of b(zj).

As far as the odd roots of δi(z) are concerned, the corresponding sin(z)(z2/L2-a0) is decreasing by large magnitude, and for the even ones, the corresponding sin(z)(z2/L2-a0) is increasing by large magnitude. However, compared to the change of sin(zj)(zj2/L2-a0), the difference between the values of (-a1(zj+2/L))cos(zj+2) and (-a1(zj/L))cos(zj) is much smaller than both for odd and even j. Thus, the b(zj) have the similar change rules as sin(zj)(zj2/L2-a0),

bj+2-bj=LK[-a1zj+2Lcos(zj+2)+sin(zj+2)(zj+22L2-a0)]-LK[-a1zjLcos(zj)+sin(zj)(zj2L2-a0)],bj+2-bj=a1L[cos(zj)-cos(zj+2)]+LK[sin(zj+2)zj+2(zj+22L2-a0)-sin(zj)zj(zj2L2-a0)]. Using Remarks 3.5 and 3.6, we can affirm that

bj>bj+2andbj-asj+ for odd values of j,

bj<bj+2andbj+asj+ for even values of j.

We are ready to state the main results of our work.

Theorem 3.7.

Under the above assumptions on K, L, a0, and a1, the range of Kp values for which a solution exists to the PID stabilization problem of an open-loop stable plant with transfer function G(s) is given by -a0K<Kp<1K(a1αLsin(α)-cos(α)(a0-α2L2)), where α is the solution of the equation tan(α)=α(2+a1L)/(α2-a1L-a0L2) in the interval [0,π].

For Kp values outside this range, there are no stabilizing PID controllers. The complete stabilizing region is given by

the cross-section of the stabilizing region in the (Ki,Kd) space is the Trapezoid T if b3>b1,

the cross-section of the stabilizing region in the (Ki,Kd) space is the triangle if b1>b3,

The parameters bj,mj,j=1,2,3 necessary for determining the boundaries, can be determined using (3.21) where zj,j=1,2,3 are the positive-real solutions of δi(z) arranged in ascending order of magnitude.

In view of Theorem 3.7, we propose an algorithm to determine the set of all stabilizing parameters for a second order delay system.

An algorithm for determining PID parameters is given as follows

Choose Kp in the interval suggested by Theorem 3.7 and initialize j=1,

Find the roots zj of δi(z),

Compute the parameter bj,mj associated with the zj for j=1,2,3 founded,

Determine the stability region in the plane (Ki,Kd) using Figure 7 (Theorem 3.7).

Go to step 1.

The stabilizing region of (Kp,Ki), (a): b3>b1, (b): b1>b3.

4. Conclusion

In this work, we have proposed an extension of Hermit-Biehler theorem to compute the stability region for second-order delay system controlled by PID controller. The procedure is based first on determining the range of proportional gain value Kp for which a solution to PID stabilization exists. Then, it is shown that for a fixed Kp inside this range, the stabilizing integral Ki and derivative gain Kd values lie inside a region with known shape and boundaries.

ZhongQ. C.Robust Control of Time Delay System2006London, UKSpringerNiculescuS.-I.Delay Effects on Stability2001269London, UKSpringer383Lecture Notes in Control and Information SciencesMR1880658KnospeC.PID controlIEEE Control Systems Magazine20062613031TakatsuH.ItohT.ArakiM.Future needs for the control theory in industries—report and topics of the control technology survey in Japanese industryJournal of Process Control199885-6369374EID2-s2.0-0032186695SilvaG. J.DattaA.BhattacharyyaS. P.PID Controllers for Time-Delay Systems2005Boston, Mass, USABirkhäuser330Control EngineeringMR2197784ToscanoR.A simple robust PI/PID controller design via numerical optimization approachJournal of Process Control20051518188EID2-s2.0-003290866110.1016/j.jprocont.2004.03.005Padma SreeR.SrinivasM. N.ChidambaramM.A simple method of tuning PID controllers for stable and unstable FOPTD systemsComputers and Chemical Engineering2004281122012218EID2-s2.0-003554629910.1016/j.compchemeng.2004.04.004HwangC.HwangJ. H.Stabilisation of first-order plus dead-time unstable processes using PID controllersIEE Proceedings: Control Theory and Applications200415118994EID2-s2.0-003647503210.1049/ip-cta:20040035McKayJ.The D-partition method applied to systems with dead time and distributed lagMeasurement and Control1970310293294EID2-s2.0-0014857365LeeT.WatkinsJ. M.EmamiT.SujoldžićS.A unified approach for stabilization of arbitrary order continuous-time and discrete-time transfer functions with time delay using a PID controllerProceedings of the IEEE Conference on Decision and Control (CDC '07)200721002105EID2-s2.0-3954912265710.1109/CDC.2007.4434729ErikssonL. M.lasse.eriksson@tkk.fiJohanssonM.mikaelj@s3.kth.sePID controller tuning rules for varying time-delay systemsProceedings of the American Control ConferenceJuly 2007619625EID2-s2.0-003207351810.1109/ACC.2007.4282655LiS.-J.LiuY.-X.An improved approach to nonlinear dynamical system identification using PID neural networksInternational Journal of Nonlinear Sciences and Numerical Simulation200672177182EID2-s2.0-0032997179ZhangY.WangF.SongY.ChenZ.chenzq@nankai.edu.cnYuanZ.Recurrent neural networks-based multivariable system PID predictive controlFrontiers of Electrical and Electronic Engineering in China200722197201EID2-s2.0-003259824810.1007/s11460-007-0037-4ShuH.PiY.PID neural networks for time-delay systemsComputers and Chemical Engineering2000242–7859862EID2-s2.0-003466130310.1016/S0098-1354(00)00340-9ShuH.GuoX.ShuH.PID neural networks in multivariable systemsProceedings of the IEEE International Symposium on Intelligent Control2002440444EID2-s2.0-0012148772RoyA.IqbalK.PID controller tuning for the first-order-plus-dead-time process model via Hermite-Biehler theoremISA Transactions2005443363378EID2-s2.0-0036033762SilvaG. J.DattaA.BhattacharyyaS. P.Stabilization of time delay systems2Proceedings of the American Control Conference2000963970EID2-s2.0-0034545271SilvaG. J.DattaA.BhattacharyyaS. P.PI stabilization of first-order systems with time delayAutomatica2001371220252031EID2-s2.0-002872724210.1016/S0005-1098(01)00165-0SilvaG. J.DattaA.BhattacharyyaS. P.Stabilization of first-order systems with time delay using the PID controllerProceedings of the American Control Conference2001646504655EID2-s2.0-0028727242SilvaG. J.DattaA.BhattacharyyaS. P.New results on the synthesis of PID controllersIEEE Transactions on Automatic Control2002472241252MR188189110.1109/9.983352XuH.DattaA.BhattacharyyaS. P.PID stabilization of LTI plants with time-delay4Proceedings of the 42nd IEEE Conference on Decision and Control (DC '03)200340384043EID2-s2.0-0036475032OliveiraV. A.TeixeiraM. C. M.CossiL. V.Stabilizing a class of time delay systems using the Hermite-Biehler theoremLinear Algebra and Its Applications2003369203216MR198848710.1016/S0024-3795(02)00721-8ZBL1109.93363OliveiraV. A.CossiL. V.SilvaA. M. F.TeixeiraM. C. M.PID stabilization of a class of time delay systemsProceedings of the 44th IEEE Conference on Decision and Control, and the European Control Conference (CDC-ECC '05)200513671372EID2-s2.0-003554634410.1109/CDC.2005.1582349OliveiraV. A.CossiL. V.TeixeiraM. C. M.SilvaA. M. F.Synthesis of PID controllers for a class of time delay systemsAutomatica200945717781782EID2-s2.0-3394737295610.1016/j.automatica.2009.03.018BhattacharyyaS. P.ChapellatH.KeelL. H.Robust Control: The Parametric Approach1995Upper Saddle River, NJ, USAPrentice-Hall