In the process of large-scale wind farms integration, the time delay is usually caused by the introduction of wide area control signal, which leads to the unstable operation of integrated transmission system. In order to solve this problem, using the control principle of interconnection and damping assignment passivity-based (IDA-PB), this paper puts forward a control method which applies voltage source converter high voltage direct current (VSC-HVDC) technology to the integrated system of time-delay wind farm and keeps the system running stably. In this method, the framework of time-delay port controlled Hamiltonian (PCH) system is constructed, and the energy shaping of the system is carried out by extending the IDA-PB control principle, thus the feedback controller of the system is designed. Around the problem of time-delay stabilization, the stability criterion is obtained by constructing Lyapunov-Krasovskii functional and introducing free weighting matrices. Finally, the simulation results show that the proposed method can effectively solve the time delay problem of the integrated transmission system and avoid the performance deterioration of the system.
Citation: Gaoran Wang, Weiwei Sun, Shuqing Wang. Stabilization of wind farm integrated transmission system with input delay[J]. AIMS Mathematics, 2021, 6(9): 9177-9193. doi: 10.3934/math.2021533
[1] | Dan-Ni Chen, Jing Cheng, Xiao Shen, Pan Zhang . Semi-stable quiver bundles over Gauduchon manifolds. AIMS Mathematics, 2023, 8(5): 11546-11556. doi: 10.3934/math.2023584 |
[2] | Fengxia Zhang, Ying Li, Jianli Zhao . The semi-tensor product method for special least squares solutions of the complex generalized Sylvester matrix equation. AIMS Mathematics, 2023, 8(3): 5200-5215. doi: 10.3934/math.2023261 |
[3] | Shu-Xin Miao, Jing Zhang . On Uzawa-SSI method for non-Hermitian saddle point problems. AIMS Mathematics, 2020, 5(6): 7301-7315. doi: 10.3934/math.2020467 |
[4] | Jiao Xu, Hairui Zhang, Lina Liu, Huiting Zhang, Yongxin Yuan . A unified treatment for the restricted solutions of the matrix equation AXB=C. AIMS Mathematics, 2020, 5(6): 6594-6608. doi: 10.3934/math.2020424 |
[5] | Yuezhen Ren, Ruihu Li, Guanmin Guo . New entanglement-assisted quantum codes constructed from Hermitian LCD codes. AIMS Mathematics, 2023, 8(12): 30875-30881. doi: 10.3934/math.20231578 |
[6] | Shahroud Azami . Monotonicity of eigenvalues of Witten-Laplace operator along the Ricci-Bourguignon flow. AIMS Mathematics, 2017, 2(2): 230-243. doi: 10.3934/Math.2017.2.230 |
[7] | Lv Zhang, Qingbiao Wu . Modified Newton-EHS method for solving nonlinear problems with complex symmetric Jacobian matrices. AIMS Mathematics, 2023, 8(10): 24233-24253. doi: 10.3934/math.20231236 |
[8] | Reena Jain, Hemant Kumar Nashine, Jung Rye Lee, Choonkil Park . Unified relational-theoretic approach in metric-like spaces with an application. AIMS Mathematics, 2021, 6(8): 8959-8977. doi: 10.3934/math.2021520 |
[9] | Ghodratallah Fasihi-Ramandi . Hamilton’s gradient estimate for fast diffusion equations under geometric flow. AIMS Mathematics, 2019, 4(3): 497-505. doi: 10.3934/math.2019.3.497 |
[10] | Sourav Shil, Hemant Kumar Nashine . Positive definite solution of non-linear matrix equations through fixed point technique. AIMS Mathematics, 2022, 7(4): 6259-6281. doi: 10.3934/math.2022348 |
In the process of large-scale wind farms integration, the time delay is usually caused by the introduction of wide area control signal, which leads to the unstable operation of integrated transmission system. In order to solve this problem, using the control principle of interconnection and damping assignment passivity-based (IDA-PB), this paper puts forward a control method which applies voltage source converter high voltage direct current (VSC-HVDC) technology to the integrated system of time-delay wind farm and keeps the system running stably. In this method, the framework of time-delay port controlled Hamiltonian (PCH) system is constructed, and the energy shaping of the system is carried out by extending the IDA-PB control principle, thus the feedback controller of the system is designed. Around the problem of time-delay stabilization, the stability criterion is obtained by constructing Lyapunov-Krasovskii functional and introducing free weighting matrices. Finally, the simulation results show that the proposed method can effectively solve the time delay problem of the integrated transmission system and avoid the performance deterioration of the system.
Let (M,ω) be a complex n-dimensional compact Hermitian manifold and χ be a smooth real (1, 1)-form on (M,ω). Γωk is the set of all real (1, 1)-forms whose eigenvalues belong to the k-positive cone Γk. For any u∈C2(M), we can get a new (1, 1)-form
χu:=χ+√−1∂¯∂u. |
In any local coordinate chart, χu can be expressed as
χu=√−1(χi¯j+ui¯j)dzi∧d¯zj. |
In this article, we study the following form of parabolic Hessian quotient equations
{∂u(x,t)∂t=logCknχku∧ωn−kClnχlu∧ωn−l−logϕ(x,u), (x,t)∈M×[0,T),u(x,0)=u0(x), x∈M, | (1.1) |
where 0≤l<k≤n, [0,T) is the maximum time interval in which the solution exists and ϕ(x,z)∈C∞(M×R) is a given strictly positive function.
The study of the parabolic flows is motivated by complex equations
χku∧ωn−k=ClnCknϕ(x,u)χlu∧ωn−l, χu∈Γωk. | (1.2) |
Equation (1.2) include some important geometry equations, for example, complex Monge-Ampère equation and Donaldson equation [6], which have attracted extensive attention in mathematics and physics since Yau's breakthrough in the Calabi conjecture [28]. Since Eq (1.2) are fully nonlinear elliptic, a classical way to solve them is the continuity method. Using this method, the complex Monge-Ampère equation
χnu=ϕ(x)ωn, χu∈Γωn |
was solved by Yau [28]. Donaldson equation
χnu=∫Mχn∫Mχ∧ωn−1χu∧ωn−1, χu∈Γωn |
was independently solved by Li-Shi-Yao [11], Collins-Szèkelyhidi [3] and Sun [17]. Equation (1.2) also include the complex k-Hessian equation and complex Hessian quotient equation, which, respectively, correspond to
Cknχku∧ωn−k=ϕ(x)ωn, χu∈Γωk, |
χku∧ωn−k=ClnCknϕ(x)χlu∧ωn−l, χu∈Γωk. |
Dinew and Kolodziej [7] proved a Liouville type theorem for m-subharmonic functions in Cn, and combining with the estimate of Hou-Ma-Wu [10], solved the complex k-Hessian equation by using the continuity method. Under the cone condition, Sun [16] solved the complex Hessian quotient equation by using the continuity method. There have been many extensive studies for complex Monge-Ampère equation, Donaldson equation, the complex k-Hessian equation and the complex Hessian quotient equation on closed complex manifolds, see, e.g., [4,12,20,22,23,29,30]. When the right hand side function ϕ in Eq (1.2) depends on u, that is ϕ=ϕ(x,u), it is interesting to ask whether we can solve them. We intend to solve (1.2) by the parabolic flow method.
Equation (1.1) covers some of the important geometric flows in complex geometry. If k=n and l=0, (1.1) is known as the complex Monge-Ampère flow
∂u(x,t)∂t=logχnuωn−logϕ(x), (x,t)∈M×[0,T), |
which is equivalent to the Kähler-Ricci flow. The result of Yau [28] was reproduced by Cao [2] through Kähler-Ricci flow. Using the complex Monge-Ampère flow, similar results on a compact Hermitian manifold and a compact almost Hermitian manifold were proved by Gill [9] and Chu [5], respectively. To study the normalized twisted Chern-Ricci flow
∂ωt∂t=−Ric(ωt)−ωt+η, |
which is equivalent to the following Mong-Ampère flow
∂φ∂t=log(θt+ddcφ)nΩ−φ, |
Tô [25,26] considered the following complex Monge-Ampère flow
∂φ∂t=log(θt+ddcφ)nΩ−F(t,x,φ), |
where Ω is a smooth volume form on M. From this, we can see that the given function ϕ depends on u in some geometric flows. If l=0, (1.1) is called as the complex k-Hessian flow
∂u(x,t)∂t=logCknχku∧ωn−kωn−logϕ(x,u), (x,t)∈M×[0,T). |
The solvability of complex k-Hessian flow was showed by Sheng-Wang [21].
In this paper, our research can be viewed as a generalization of Tô's work in [26] and Sheng-Wang's work in [21]. To solve the complex Hessian quotient flow, the condition of the parabolic C-subsolution is needed. According to Phong and Tô [14], we can give the definition of the parabolic C-subsolution to Eq (1.1).
Definition 1.1. Let u_(x,t)∈C2,1(M×[0,T)) and χu_∈Γωk, if there exist constants δ, R>0, such that for any (x,t)∈M×[0,T),
logσk(λ)σl(λ)−∂tu_≤logϕ(x,u_), λ−λ(u_)+δI∈Γn, |
implies that
|λ|<R, |
then u_ is said to be a parabolic C-subsolution of (1.1), where λ(u_) denotes eigenvalue set of χu_.
Obviously, we can give the equivalent definition of parabolic C-subsolution of (1.1).
Definition 1.2. Let u_(x,t)∈C2,1(M×[0,T)) and χu_∈Γωk, if there exist constant ˜δ>0, for any (x,t)∈M×[0,T), such that
limμ→∞logσk(λ(u_)+μei)σl(λ(u_)+μei)>∂u_∂t+˜δ+logϕ(x,u_), 1≤i≤n, | (1.3) |
then u_ is said to be a parabolic C-subsolution of (1.1).
Our main result is
Theorem 1.3. Let (M,g) a compact Hermitian manifold and χ be a smooth real (1,1)-form on M. Assume there exists a parabolic C-subsolutionu_ for Eq (1.1) and
∂tu_≥max{supM(logσk(λ(u0))σl(λ(u0))−logϕ(x,u0)),0}, | (1.4) |
ϕz(x,z)ϕ>cϕ>0, | (1.5) |
where cϕ is a constant.Then there exits a unique smooth solution u(x,t) to (1.1) all timewith
supx∈M(u0(x)−u_(x,0))=0. | (1.6) |
Moreover, u(x,t) isC∞ convergent to a smooth function u∞, which solves Eq (1.2).
The rest of this paper is organized as follows. In Section 2, we give some important lemmas and estimate on |ut(x,t)|. In Section 3, we prove C0 estimates of Eq (1.1) by the parabolic C-subsolution condition and the Alexandroff-Bakelman-Pucci maximum principle. In Section 4, using the parabolic C-subsolution condition, we establish the C2 estimate for Eq (1.1) by the method of Hou-Ma-Wu [10]. In Section 5, we adapt the blowup method of Dinew and Kolodziej [7] to obtain the gradient estimate. In Section 6, we give the proof of the long-time existence of the solution to the parabolic equation and its convergence, that is Theorem 1.3.
In this section, we give some notations and lemmas. In holomorphic coordinates, we can set
ω=√−1gi¯jdzi∧d¯zj=√−1δijdzi∧d¯zj, χ=√−1χi¯jdzi∧d¯zj, |
χu=√−1(χi¯j+ui¯j)dzi∧d¯zj=√−1Xi¯jdzi∧d¯zj, |
χu_=√−1(χi¯j+u_i¯j)dzi∧d¯zj=√−1X_i¯jdzi∧d¯zj. |
λ(u) and λ(u_) denote the eigenvalue set of {Xi¯j} and {X_i¯j} with respect to {gi¯j}, respectively. In local coordinates, (1.1) can be written as
∂tu=logσk(λ(u))σl(λ(u))−logϕ(x,u). | (2.1) |
For simplicity, set
F(λ(u))=logσk(λ(u))σl(λ(u)), |
then (2.1) is abbreviated as
∂tu=F(λ(u))−logϕ(x,u). | (2.2) |
We use the following notation
Fi¯j=∂F∂Xi¯j, F=∑iFi¯i, Fi¯j,p¯q=∂2F∂Xi¯j∂Xp¯q. |
For any x0∈M, we can choose a local holomorphic coordinates such that the matrix {Xi¯j} is diagonal and X1¯1≥⋯≥Xn¯n, then we have, at x0∈M,
λ(u)=(λ1,⋯,λn)=(X1¯1,⋯,Xn¯n), |
Fi¯j=Fi¯iδij=(σk−1(λ|i)σk−σl−1(λ|i)σl)δij, F1¯1≤⋯≤Fn¯n. |
To prove a priori C0-estimate for solution to Eq (1.1), we need the following variant of the Alexandroff-Bakelman-Pucci maximum principle, which is Proposition 10 in [20].
Lemma 2.1. [20] Let v:B(1)→R be a smooth function, which meets the conditionv(0)+ϵ≤inf∂B(1)v, whereB(1) denotes the unit ball in Rn.Define the set
Ω={x∈B(1):|Dv(x)|<ϵ2, and v(y)≥v(x)+Dv(x)⋅(y−x),∀y∈B(1)}. |
Then there exists a costant c0>0 such that
c0ϵn≤∫Ωdet(D2v). |
Next, we give an estimate on |ut(x,t)|.
Lemma 2.2. Under the assumption of Theorem 1.3, let u(x,t) be a solution to (1.1). Then for any (x,t)∈M×[0,T), we have
min{infMut(x,0),0}≤ut(x,t)≤max{supMut(x,0),0}. | (2.3) |
Furthermore, there is a constant C>0 such that
supM×[0,T)|∂tu(x,t)|≤supM|∂tu(x,0)|≤C, |
where C depends on H=|u0|C2(M) and |ϕ|C0(M×[−H,H]).
Proof. Differentiating (2.2) on both sides simultaneously at t, we obtain
(ut)t=Fi¯jXi¯jt−ϕzϕut=Fi¯j(ut)i¯j−ϕzϕut. | (2.4) |
Set uεt=ut−εt, ε>0. For any T′∈(0,T), suppose uεt achieves its maximum Mt at (x0,t0)∈M×[0,T′]. Without loss of generality, we may suppose Mt≥0. If t0>0, From the parabolic maximum principle and (2.4), we get
0≤(uεt)t−Fi¯j(uεt)i¯j+ϕzϕuεt≤(ut)t−ε−Fi¯j(ut)i¯j+ϕzϕut−εϕzϕt0≤−ε−εϕzϕt0. |
This is obviously a contradiction, so t0=0 and
supM×[0,T′]uεt(x,t)=supMut(x,0), |
that is
supM×[0,T′]ut(x,t)=supM×[0,T′](uεt(x,t)+εt)≤supMut(x,0)+εT′. |
Letting ε→0, we obtain
supM×[0,T′]ut(x,t)≤supMut(x,0). |
Since T′∈(0,T) is arbitrary, we have
supM×[0,T)ut(x,t)≤supMut(x,0). | (2.5) |
Similarly, setting uεt=ut+εt, ε>0, we obtain
infM×[0,T)ut(x,t)≥infMut(x,0). | (2.6) |
(2.1) yields
|ut(x,0)|=|logσk(λ(u0))σl(λ(u0))−logϕ(x,u0)|≤C. | (2.7) |
Combining (2.5)–(2.7), we complete the proof of Proposition 2.2.
From the concavity of F(λ(u)) and the condition of the parabolic C-subsolution, we give the following lemma, which plays an important role in the estimation of C2.
Lemma 2.3. Under the assumption of Theorem 1.3 and assuming that X1¯1≥⋯≥Xn¯n, there exists two positive constantsN and θ such that we have either
Fi¯i(u_i¯i−ui¯i)−∂t(u_−u)≥θ(1+F) | (2.8) |
or
F1¯1≥θN(1+F). | (2.9) |
Proof. Since u_ is a parabolic C-subsolution to Eq (1.1), from Definition 1.2, there are uniform constants ˜δ>0 and N>0, such that
logσk(λ(u_)+Ne1)σl(λ(u_)+Nue1)>∂u_∂t+˜δ+logϕ(x,u_). | (2.10) |
If ϵ>0 is sufficiently small, it can be obtained from (2.10)
logσk(λ(u_)−ϵI+Ne1)σl(λ(u_)−ϵI+Nue1)≥∂u_∂t+˜δ+logϕ(x,u_). |
Set λ′=λ(u_)−ϵI+Ne1, then
F(λ′)≥∂u_∂t+˜δ+logϕ(x,u_). | (2.11) |
Using the concavity of F(λ(u)) gives
Fi¯i(u_i¯i−ui¯i)=Fi¯i({X_i¯i−Xi¯i)=Fi¯i({X_i¯i−ϵδii+Nδi1−Xi¯i)+ϵF−NF1¯1≥F(λ′)−F(λ(u))+ϵF−NF1¯1. | (2.12) |
From Lemma 2.2 and (1.4), we obtain
u_t(x,t)≥ut(x,t), ∀(x,t)∈M×[0,T). | (2.13) |
In addition, it can be obtained from the condition (1.6)
u_(x,0)≥u(x,0), ∀x∈M×[0,T). | (2.14) |
(2.13) and (2.14) deduce that
u_(x,t)≥u(x,t), ∀(x,t)∈M×[0,T). | (2.15) |
It follows from this that
ϕ(x,u_)≥ϕ(x,u). | (2.16) |
Combining (2.2), (2.11) and (2.16) gives that
F(λ′)−F(λ(u))≥u_t(x,t)−ut(x,t)+˜δ+logϕ(x,u_)−logϕ(x,u)≥u_t(x,t)−ut(x,t)+˜δ. | (2.17) |
Put (2.17) into (2.12)
Fi¯i(u_i¯i−ui¯i)≥u_t(x,t)−ut(x,t)+˜δ+ϵF−NF1¯1≥˜δ+ϵF−NF1¯1. |
Let
θ=min{˜δ2, ϵ2}. |
If F1¯1N≤θ(1+F), Inequality (2.8) is obtained, otherwise Inequality (2.9) must be true.
In this section, we prove the C0 estimates by the existence of the parabolic C-subsolution and the Alexandroff-Bakelman-Pucci maximum principle.
Proposition 3.1. Under the assumption of Theorem1.3, let u(x,t) be a solution to (1.1). Then there exists a constant C>0 such that
|u(x,t)|C0(M×[0,T))≤C, |
where C depends on |u0|C2(M) and |u_|C2(M×[0,T)).
Proof. Combining (2.13), (2.14) and ∂ϕ(x,z)∂z≥0 yields
u_t(x,t)+logϕ(x,u_)≥ut(x,t)+logϕ(x,u). | (3.1) |
Let's rewrite Eq (2.2) as
F(λ(u))=∂tu+logϕ(x,u). | (3.2) |
when fix t∈[0,T), Eq (3.2) is elliptic. From (3.1), we see that the parabolic C-subsolution u_(x,t) is a C-subsolution to Eq (3.2) in the elliptic sense. From (2.15), we have
supM×[0,T)(u−u_)=0. |
Our goal is to obtain a lower bound for L=infM×t(u−u_). Note that λ(u)∈Γk, which implies that λ(u)∈Γ1, then Δ(u−u_)≥−˜C, where Δ is the complex Laplacian with respect to ω. According to Tosatti-Weinkove's method [22], we can prove that ‖u−u_‖L1(M) is bounded uniformly. Let G:M×M→R be the associated Green's function, then, by Yau [28], there is a uniform constant K such that
G(x,y)+K≥0, ∀(x,y)∈M×M, and ∫y∈MG(x,y)ωn(y)=0. |
Since
supM×[0,T)(u−u_)=0, |
then for fixed t∈[0,T) there exists a point x0∈M such that (u−u_)(x0,t)=0. Thus
(u−u_)(x0,t)=∫M(u−u_)dμ−∫y∈MG(x0,y)Δ(u−u_)(y)ωn(y)=∫M(u−u_)dμ−∫y∈M(G(x0,y)+K)Δ(u−u_)(y)ωn(y)≤∫M(u−u_)dμ+˜CK∫Mωn, |
that is
∫M(u_−u)dμ=∫M|(u−u_)|dμ≤˜CK∫Mωn. |
Let us work in local coordinates, for which the infimum L is achieved at the origin, that is L=u(0,t)−u_(0,t). We write B(1)={z:|z|<1}. Let v=u−u_+ϵ|z|2, for a small ϵ>0. We have infv=L=v(0), and v(z)≥L+ϵ for z∈∂B(1). From Lemma 2.1, we get
c0ϵ2n≤∫Ωdet(D2v). | (3.3) |
At the same time, if x∈Ω, then D2v(x)≥0 implies that
ui¯j(x)−u_i¯j(x)+ϵδij≥0. |
If ϵ is sufficiently small, then
λ(u)∈λ(u_)−δI+Γn. |
Set μ=λ(u)−λ(u_). Since λ(u) satisfies Eq (3.2), then
F(λ(u_)+μ)=∂tu+logϕ(x,u), μ+δI∈Γn. | (3.4) |
u_ is a C-subsolution to Eq (3.2) in the elliptic sense, so there is a uniform constant R>0, such that
|μ|≤R, |
which means |vi¯j|≤C, for any x∈Ω. As in Blocki [1], for x∈Ω, we have D2v(x)≥0 and so
D2v(x)≤22ndet(vi¯j)2≤C′. |
From this and (3.3), we obtain
c0ϵ2n≤∫Ωdet(D2v)≤C′⋅vol(Ω). | (3.5) |
On the other hand, by the definition of Ω in Lemma 2.1, for x∈Ω, we get
v(0)≥v(x)−Dv(x)⋅x>v(x)−ϵ2, |
and so
|v(x)|>|L+ϵ2|. |
It follows that
∫M|v(x)|≥∫Ω|v(x)|≥|L+ϵ2|⋅vol(Ω). | (3.6) |
Since ‖u−u_‖L1(M) is bounded uniformly, ∫M|v(x)| is also bounded uniformly. If L is very large, Inequality (3.6) contradicts (3.5), which means that L has a lower bound. For any t∈[0,T), Inequality (3.1) holds, thus
|u(x,t)|C0(M×[0,T))≤|L|+supM×[0,T)|u_|≤C. |
In this section, we prove that the second-order estimates are controlled by the square of the gradient estimate linearly. Our calculation is a parabolic version of that in Hou-Ma-Wu [10].
Proposition 4.1. Under the assumption of Theorem1.3, let u(x,t) be a solution to (1.1). Then there exists a constant ˜C such that
supM×[0,T)|√−1∂¯∂u|≤˜C(supM×[0,T)|∇u|2+1), |
where ˜C depends χ, ω, |ϕ|C2(M×[−C,C]), |u_|C2(M×[0,T)), |∂tu_|C0(M×[0,T)) and |u0|C2(M).
Proof. Let λ(u)=(λ1,…,λn) and λ1 is the maximum eigenvalue. For any T′<T, we consider the following function
W(x,t)=logλ1+φ(|∇u(x,t)|2)+ψ(u(x,t)−u_(x,t)), (x,t)∈M×[0,T′], | (4.1) |
where φ and ψ are determined later. We want to apply the maximum principle to the function W. Since the eigenvalues of the matrix {Xi¯j} with respect to ω need not be distinct at the point where W achieves its maximum, we will perturb {Xi¯j} following the technique of [20]. Let W achieve its maximum at (x0,t0)∈M×[0,T′]. Near (x0,t0), we can choose local coordinates such that {Xi¯j} is diagonal with X1¯1≥⋯≥Xn¯n, and λ(u)=(X1¯1,⋯,Xn¯n). Let D be a diagonal matrix such that D11=0 and 0<D22<⋯<Dnn are small, satisfying Dnn<2D22. Define the matrix ˜X=X−D. At (x0,t0), ˜X has eigenvalues
˜λ1=λ1, ˜λi=λi−Dii, n≥i≥2. |
Since all the eigenvalues of ˜X are distinct, we can define near (x0,t0) the following smooth function
˜W=log˜λ1+φ(|∇u|2)+ψ(u−u_), | (4.2) |
where
φ(s)=−12log(1−s2K), 0≤s≤K−1, |
ψ(s)=−Elog(1+s2L), −L+1≤s≤L−1, |
K=supM×[0,T′]|∇u|2+1, |
L=supM×[0,T′]|u|+supM×[0,T′]|u_|+1, |
E=2L(C1+1), |
and C1>0 is to be chosen later. Direct calculation yields
0<14K≤φ′≤12K, φ″=2(φ′)2>0, | (4.3) |
and
C1+1≤−ψ′≤2(C1+1), ψ″≥4ϵ1−ϵ(ψ′)2,∀ ϵ≤14E+1. | (4.4) |
Without loss of generality, we can assume that λ1>1. From here on, all calculations are done at (x0,t0). From the maximum principle, calculating the first and second derivatives of the function ˜W gives
0=˜Wi=˜λ1,iλ1+φ′(|∇u|2)i+ψ′(u−u_)i,1≤i≤n, | (4.5) |
0≥˜Wi¯i=˜λ1,i¯iλ1−˜λ1,i˜λ1,¯iλ21+φ′(|∇u|2)i¯i+φ″|(|∇u|2)i|2+ψ′(u−u_)i¯i+ψ″|(u−u_)i|2. | (4.6) |
0≤˜Wt=˜λ1,tλ1+φ′(|∇u|2)t+ψ′(u−u_)t. | (4.7) |
Define
L:=Fi¯j∇∂∂¯zj∇∂∂zi−∂t. |
Obviously,
0≥L˜W=Llog˜λ1+Lφ(|∇u|2)+Lψ(u−u_). | (4.8) |
Next, we will estimate the terms in (4.8). Direct calculation shows that
Llog˜λ1=Fi¯i˜λ1,i¯iλ1−Fi¯i|˜λ1,i|2λ21−˜λ1,tλ1. | (4.9) |
According to Inequality (78) in [20], we have
˜λ1,i¯i≥Xi¯i1¯1−2Re(Xi¯11¯T1i1)−C0λ1, | (4.10) |
where C0 depending χ, ω, |ϕ|C2(M×[−C,C]), |u_|C2(M×[0,T)), |∂tu_|C0(M×[0,T)) and |u0|C2(M)). From here on, C0 can always absorb the constant it represents before, and can change from one line to the next, but it does not depend on the parameter we choose later. By calculating the covariant derivatives of (4.7) in the direction ∂∂z1 and ∂∂¯z1, we obtain
ut1=Fi¯iXi¯i1−(logϕ)1−(logϕ)uu1, | (4.11) |
and
ut1¯1=Fi¯j,p¯qXi¯j1Xp¯q¯1+Fi¯iXi¯i1¯1−(logϕ)1¯1−(logϕ)1uu¯1−(logϕ)u¯1u1−(logϕ)uu|u1|2−(logϕ)uu1¯1. | (4.12) |
Notice that
X1¯1i=χ1¯1i+u1¯1i=(χ11i−χi11+Tpi1χp¯1)+Xi¯11−T1i1λ1, | (4.13) |
therefore
|X1¯1i|2≤|Xi¯11|2−2λ1Re(Xi¯11¯T1i1)+C0(λ21+|X1¯1i|). | (4.14) |
Combining (4.14) with
˜λ1,i=X1¯1i−(D11)i, |
gives
−Fi¯i|˜λ1,i|2λ21=−Fi¯i|X1¯1i|2λ21+2λ21Fi¯iRe(X1¯1i(D11)¯i)−Fi¯i|(D11)i|2λ2i≥−Fi¯i|X1¯1i|2λ21−C0λ21Fi¯i|X1¯1i|−C0F≥−Fi¯i|Xi¯11|2λ21+2λ1Fi¯iRe(Xi¯11¯T1i1)−C0λ21Fi¯i|X1¯1i|−C0F. | (4.15) |
Let's set λ1≥K, otherwise the proof is completed. Putting 4.10–4.12 and (4.15) into (4.9) yields
Llog˜λ1≥−Fi¯j,p¯qXi¯j1Xp¯q¯1λ1−Fi¯i|Xi¯11|2λ21−C0λ1Fi¯i|X1¯1i|λ1−C0F+(logϕ)1¯1+(logϕ)1uu¯1+(logϕ)u¯1u1+(logϕ)uu|u1|2−(logϕ)uχ1¯1λ1≥−Fi¯j,p¯qXi¯j1Xp¯q¯1λ1−Fi¯i|Xi¯11|2λ21−C0λ1Fi¯i|X1¯1i|λ1−C0F−C0 | (4.16) |
A simple computation gives
Lφ(|∇u|2)=φ′Fi¯i(|∇u|2)i¯i+φ″Fi¯i|(|∇u|2)i|2−φ′(|∇u|2)t. | (4.17) |
Next, we estimate the formula (4.17). Differentiating Eq (2.2), we have
(ut)p=Fi¯iXi¯ip−(logϕ)p−(logϕ)uup, | (4.18) |
and
(ut)¯p=Fi¯iXi¯i¯p−(logϕ)¯p−(logϕ)uu¯p. | (4.19) |
It follows from (4.18) and (4.19) that
∂t|∇u|2=∑putpu¯p+∑pupu¯t¯p=Fi¯i∑p(Xi¯ipu¯p+Xi¯i¯pup)−∑p(logϕ)pu¯p−∑p(logϕ)¯pup−2∑p(logϕ)u|∇u|2. | (4.20) |
By commuting derivatives, we have the identity
Fi¯iXi¯ip=Fi¯iui¯ip+Fi¯iχi¯ip=Fi¯iupi¯i−Fi¯iTqpiuq¯i−Fi¯iuqR qi¯ip+Fi¯iχi¯ip. | (4.21) |
Direct calculation gives
Fi¯i(|∇u|2)i¯i=∑pFi¯i(upi¯iu¯p+u¯pi¯iup)+∑pFi¯i(upiu¯p¯i+up¯iu¯pi). | (4.22) |
It follows from (4.21) that
Fi¯iupi¯iu¯p−Fi¯iXi¯ipu¯p=Fi¯iTqpiuq¯iu¯p+Fi¯iu¯puqR qi¯ip−Fi¯iχi¯ipu¯p≥−C0K12Fi¯iXi¯i−C0K12F−C0KF. | (4.23) |
Noticing that
Fi¯iXi¯i=σk−1(λ|i)σkλi−σl−1(λ|i)σlλi=k−l, | (4.24) |
from this and (4.23), we obtain
Fi¯iupi¯iu¯p−Fi¯iXi¯ipu¯p≥−C0K12−C0K12F−C0KF. | (4.25) |
In the same way, we can get
Fi¯iu¯pi¯iup−Fi¯iXi¯i¯pup≥−C0K12−C0K12F−C0KF. | (4.26) |
Using (4.20)–(4.26) in (4.17), we have
Lφ(|∇u|2)≥φ″Fi¯i|(|∇u|2)i|2+φ′(−C0K12−C0K12F−C0KF)+φ′∑p((logϕ)pu¯p+(logϕ)¯pup+2(logϕ)u|∇u|2)+∑pFi¯i(|upi|2+|u¯pi|2)≥φ″Fi¯i|(|∇u|2)i|2+∑pFi¯i(|upi|2+|u¯pi|2)−C0−C0F. | (4.27) |
A simple calculation gives
Lψ(u−u_)=ψ″Fi¯i|(u−u_)i|2+ψ′[Fi¯i(u−u_)i¯i−∂t(u−u_)]. | (4.28) |
Substituting (4.27), (4.28) and (4.16) into (4.8),
0≥−Fi¯j,p¯qXi¯j1Xp¯q¯1λ1−Fi¯i|Xi¯11|2λ21−C0λ1Fi¯i|X1¯1i|λ1+φ″Fi¯i|(|∇u|2)i|2+φ′∑pFi¯i(|upi|2+|u¯pi|2)−C0−C0F+ψ″Fi¯i|(u−u_)i|2+ψ′[Fi¯i(u−u_)i¯i−∂t(u−u_)]. | (4.29) |
Let δ>0 be a sufficiently small constant to be chosen later and satisfy
δ≤min{11+4E,12}. | (4.30) |
We separate the rest of the calculations into two cases.
Case 1: λn<−δλ1.
Using(4.5), we find that
−Fi¯i|X1¯1i|2λ21=−Fi¯i|φ′(|∇u|2)i+ψ′(u−u_)i−(D11)iλ1|2≥−2(φ′)2Fi¯i|(|∇u|2)i|2−2Fi¯i|ψ′(u−u_)i−(D11)iλ1|2≥−2(φ′)2Fi¯i|(|∇u|2)i|2−C0|ψ′|2KF−C0F. | (4.31) |
From (4.13), we have
|Xi¯11|2λ21≤|X1¯1i|2λ21+C0(1+|X1¯1i|λ1). | (4.32) |
Combining (4.31) with (4.32), we conclude that
−Fi¯i|Xi¯11|2λ21≥−2(φ′)2Fi¯i|(|∇u|2)i|2−C0|ψ′|2KF−C0F−C0Fi¯i|X1¯1i|λ1. | (4.33) |
Note that the operator F is concave, which implies that
−Fi¯j,p¯qXi¯j1Xp¯q¯1λ1≥0. | (4.34) |
Applying (4.33) and (4.34) to (4.29) and using φ″=2(φ′)2 yield that
0≥−C0Fi¯i|X1¯1i|λ1−C0λ1Fi¯i|X1¯1i|λ1+φ′∑pFi¯i(|upi|2+|u¯pi|2)−C0|ψ′|2KF−C0F−C0+ψ″Fi¯i|(u−u_)i|2+ψ′[Fi¯i(u−u_)i¯i−∂t(u−u_)]. | (4.35) |
Note that the fact
|X1¯1i|λ1=|−φ′(upiu¯p+upu¯pi)−ψ′(u−u_)i+(D11)iλ1|, |
It follows that
−C0Fi¯i|X1¯1i|λ1≥−C0φ′K−12Fi¯i(|upi|+|u¯pi|)+C0ψ′K12F−C0F. | (4.36) |
Using the following inequality
K−12(|upi|+|u¯pi|)≤14C0(|upi|2+|u¯pi|2)+C0K. |
deduces
−C0Fi¯i|X1¯1i|λ1≥−14φ′Fi¯i(|upi|2+|u¯pi|2)+C0ψ′K12F−C0F. | (4.37) |
Note that λ1>1, we have
−C0λ1Fi¯i|X1¯1i|λ1≥−14φ′Fi¯i(|upi|2+|u¯pi|2)+C0ψ′K12F−C0F. | (4.38) |
Since ψ″>0, which implies that
ψ″Fi¯i|(u−u_)i|2≥0. | (4.39) |
According to Lemma 2.3, there are at most two possibilities:
(1) If (2.8) holds true, then
ψ′[Fi¯i(u−u_)i¯i−∂t(u−u_)]≥θ(1+F)|ψ′|. | (4.40) |
Substituting (4.37)—(4.40) into (4.35) and using φ′≥14K yield that
0≥18K∑pFi¯i(|upi|2+|u¯pi|2)−C0|ψ′|2KF−C0(F+1)+θ(1+F)|ψ′|≥18KFi¯iλ2i−C0(C1+1)2KF+θ(C1+1)(F+1)−C0(F+1)≥δ2λ218nKF−C0(C1+1)2KF+θ(C1+1)(F+1)−C0(F+1). | (4.41) |
We may set θC1≥C0. It follows from (4.41) that λ1≤˜CK.
(2) If (2.9) holds true,
F1¯1>θN(1+F). | (4.42) |
According to ψ′<0 and the concavity of the operator F, we have
ψ′[Fi¯i(u−u_)i¯i−∂t(u−u_)]=ψ′[Fi¯i(Xi¯i−X_i¯i)−∂t(u−u_)]≥ψ′[F(χu)−F(χu_)−∂tu+∂tu_]=ψ′[ϕ(x,u)+∂tu_−F(χu_)]≥C0ψ′. | (4.43) |
Using (4.37)—(4.39) and (4.43) in (4.35), together with (4.42), we find that
0≥18KFi¯iλ2i−C0|ψ′|2KF+C0ψ′−C0(F+1)≥θλ218NK(1+F)+δ2λ218nKF−C0(C1+1)2KF−C0(C1+1)−C0(F+1). | (4.44) |
Let λ1 be sufficiently large, so that
θλ218NK(1+F)−C0(C1+1)−C0(F+1)≥0, |
It follows from (4.44)that λ1≤˜CK.
Case 2: λn≥−δλ1.
Let
I={i∈{1,⋯,n}|Fi¯i>δ−1F1¯1}. |
Let us first treat those indices which are not in I. Similar to (4.31), we obtain
−∑i∉IFi¯i|X1¯1i|2λ21≥−2(φ′)2∑i∉IFi¯i|(|∇u|2)i|2−C0Kδ|ψ′|2F1¯1−C0F. | (4.45) |
Using (4.32) yields that
−∑i∉IFi¯i|Xi¯11|2λ21≥−2(φ′)2∑i∉IFi¯i|(|∇u|2)i|2−C0Kδ|ψ′|2F1¯1−C0∑i∉IFi¯i|X1¯1i|λ1−C0F. | (4.46) |
Since
−Fi¯1,1¯i=Fi¯i−F1¯1X1¯1−Xi¯i andλi≥λn≥−δλ1, |
which implies that
−∑i∈IFi¯1,1¯i≥1−δ1+δ1λ1∑i∈IFi¯i, |
It follows that
−Fi¯1,1¯i|Xi¯11|2λ1≥1−δ1+δ∑i∈IFi¯i|Xi¯11|2λ21. | (4.47) |
Recalling φ″=2(φ′)2 and 0<δ≤12, we obtain from (4.5) that
∑i∈Iφ″Fi¯i|(|∇u|2)i|2=2∑i∈IFi¯i|Xi¯11λ1+ψ′(u−u_)i+χ11i−χi11+Tpi1χp¯1−(D11)iλ1|2≥2∑i∈IFi¯i(δ|Xi¯11λ1|2−2δ1−δ(ψ′)2|(u−u_)i|2−C0)≥2δ∑i∈IFi¯i|Xi¯11λ1|2−4δ1−δ(ψ′)2Fi¯i|(u−u_)i|2−C0F. | (4.48) |
Notice that ψ″≥4ϵ1−ϵ(ψ′)2 if ϵ=14E+1. Since 14E+1≥δ, we get that
ψ″Fi¯i|(u−u_)i|2−4δ1−δ(ψ′)2Fi¯i|(u−u_)i|2≥0. | (4.49) |
Take (4.46)–(4.49) into (4.29),
0≥−C0∑i∉IFi¯i|X1¯1i|λ1−C0λ1Fi¯i|X1¯1i|λ1+φ′∑pFi¯i(|upi|2+|u¯pi|2)−C0F−C0−C0Kδ|ψ′|2F1¯1+ψ′[Fi¯i(u−u_)i¯i−∂t(u−u_)]. | (4.50) |
Similar to (4.37) and (4.38), by using the third term of (4.50) to absorb the first two terms of it, We get that
0≥18K∑pFi¯i(|upi|2+|u¯pi|2)−C0Kδ|ψ′|2F1¯1−C0F−C0+ψ′[Fi¯i(u−u_)i¯i−∂t(u−u_)]≥18K∑pFi¯iλ2i−C0Kδ|ψ′|2F1¯1−C0(F+1)+ψ′[Fi¯i(u−u_)i¯i−∂t(u−u_)]. | (4.51) |
According to Lemma2.3, there are at most two possibilities:
(1) If (2.8) holds true, then
ψ′[Fi¯i(u−u_)i¯i−∂t(u−u_)]≥θ(1+F)|ψ′|. | (4.52) |
Put (4.52) into (4.51)
0≥18K∑pFi¯iλ2i−C0Kδ|ψ′|2F1¯1−C0(F+1)+θ(1+F)|ψ′|≥18K∑pF1¯1λ21−C0Kδ(1+C1)2F1¯1−C0(F+1)+θ(1+F)(1+C1). | (4.53) |
Here, C1 is determined finally, such that
θC1≥C0. |
It follows from (4.53) that
λ1≤˜CK. |
(2) If (2.9) holds true,
F1¯1>θN(1+F). | (4.54) |
Substituting (4.43) into (4.51) and using (4.54) give that
0≥18Kλ21−C0Kδ|(1+C1)|2−NθC0(1+C1)−NθC0 | (4.55) |
It follows that
λ1≤˜CK. |
To obtain the gradient estimates, we adapt the blow-up method of Dinew and Kolodziej [7] and reduce the problem to a Liouville type theorem which is proved in [7].
Proposition 5.1. Under the assumption of Theorem 1.3, let u(x,t) be a solution to (1.1). Then there exists a uniform constant ˜C such that
supM×[0,T)|∇u|≤˜C. | (5.1) |
Proof. Suppose that the gradient estimate (5.1) does not hold. Then there exists a sequence (xm,tm)∈M×[0,T) with tm→T such that
supM×[0,tm]|∇u(x,t)|=|∇u(xm,tm)| andlimm→∞|∇u(xm,tm)|=∞. |
After passing to a subsequence, we may assume that limm→∞xm=x0∈M. We choose a coordinate chart {U,(z1,⋯,zn)} at x0, which we identify with an open set in Cn, and such that ω(0)=β=√−1∑idzi∧d¯zi. We may assume that the open set contains ¯B1(0) and m is sufficiently large so that zm=z(xm)∈B1(0). Define
|∇u(xm,tm)|=Cm, ˜um(z)=u(zCm,tm). |
From this and Proposition 4.1, we have
supM|∇˜um|=1, supM|√−1 ¯∂∂˜um|≤˜C. |
This yields that ˜um is contained in the Hölder space C1,γ(Cn) with a uniform. Along with a standard application of Azela-Ascoli theorem, we may suppose ˜um has a limit ˜u∈C1,γ(Cn) with
|˜u|+|∇˜u|<C and |∇˜u(0)|≠0, | (5.2) |
in particular ˜u is not constant. On the other hand, similar to the method of Dinew and Kolodziej [7], we have
[χu(zCm)]k∧[ω(zCm)]n−k=e∂tuϕm(zCm,u)[χu(zCm)]l∧[ω(zCm)]n−l. |
Fixing z, we obtain
C2(k−l)m[O(1C2m)β+√−1 ¯∂∂˜um(z)]k∧[(1+O(|z|2C2m))β]n−k=e∂tumϕm(zCm,um)[O(1C2m)β+√−1 ¯∂∂˜um(z)]l∧[(1+O(|z|2C2m))β]n−l. | (5.3) |
Lemma 2.2 gives ∂tu is bounded uniformly. Since
ϕm(zCm,um)≤supM×[−C,C]ϕ, |
which implies that ϕm(zCm,um) is bounded uniformly. Taking the limits on both sides of 5.3 by m→∞ yields that
(√−1∂¯∂˜u)k∧βn−k=0. | (5.4) |
which is in the pluripotential sense. Moreover, a similar reasoning tells us that for any 1≤p≤k,
(√−1∂¯∂˜u)p∧βn−p≥0. | (5.5) |
Then, (5.4) and (5.5) imply ˜u is a k-subharmonic. By a result of Blocki [1], ˜u is a maximal k-subharmonic function in Cn. Applying the Liouville theorem in [7], we find that ˜u is a constant, which contradicts with (5.2).
In this section, we shall give a proof of the long-time existence to the flow and its convergence, that is Theorem 1.3.
From Lemma 2.2, Proposition 3.1, Propositions 4.1 and 5.1, we conclude that Eq 1.1 is uniformly parabolic. Therefore by Evans-Krylov's regularity theory [8,13,19,27] for uniformly parabolic equation, we obtain higher order derivative estimates. By the a priori estimates which don't depend on time, one can prove that the short time existence on [0,T) extends to [0,∞), that is the smooth solution exists at all time t>0. After proving C∞ estimates on [0,∞), we are able to show the convergence of the solution flow.
Let v=eγtut, where 0<γ<cϕ. Commuting derivative of v with respect to t and using (2.4), we obtain
vt=eγtutt+γv=γv+eγt(Fi¯juti¯j−ϕzϕut)=Fi¯jvi¯j+(γ−ϕzϕ)v. |
Using the condition (1.5) yields γ−ϕzϕ<0, According to the parabolic maximum principle, it follows that
supM×[0,∞)|v(x,t)|≤supM|ut(x,0)|≤supM|F(λ(u0))−ϕ(x,u0)|≤C, |
which means that |ut| decreases exponentially, in particular
∂t(u+Cγe−γt)≤0. |
According to Proposition 3.1, it follows that u+Cγe−γt is bounded uniformly and decreasing in t. Thus it converges to a smooth function u∞. From the higher order prior estimates, we can see that the function u(x,t) converges smoothly to u∞. Letting t→∞ in Eq (2.1),
σk(λ(u∞))σl(λ(u∞))=ϕ(x,u∞). |
In this paper, we have considered the parabolic Hessian quotient equation (1.1), in which the right hand side function ϕ depends on u. Firstly, we prove C0 estimates of Eq (1.1) by the parabolic C-subsolution condition and the Alexandroff-Bakelman-Pucci maximum principle. Secondly, we establish the C2 estimate for Eq (1.1) by using the parabolic C-subsolution condition. Thirdly, we obtain the gradient estimate by adapting the blowup method. Finally we give the proof of the long-time existence of the solution to the parabolic equation and its convergence. As an application, we show the solvability of a class of complex Hessian quotient equations, which generalizes the relevant results.
This work was supported by the Natural Science Foundation of Anhui Province Education Department (Nos. KJ2021A0659, gxgnfx2018017); Quality Enginering Project of Anhui Province Education Department (Nos. 2018jyxm0491, 2019mooc205, 2020szsfkc0686); Science Research Project of Fuyang Normal University (No. 2021KYQD0011)
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
[1] | T. An, G. Tang, W. Wang, Research and application on multi-terminal and DC grids based on VSC-HVDC technology in China, High Volt., 2 (2017), 1–10. |
[2] | Y. F. Guo, H. L. Gao, Q. W. Wu, H. R. Zhao, J. Ostergaard, M. Shahidehpour, Enhanced voltage control of VSC-HVDC connected offshore wind farms based on model predictive control, IEEE T. Sustain. Energ., 9 (2018), 474–487. |
[3] | H. Z. Liu, Z. Chen, Contribution of VSC-HVDC to frequency regulation of power systems with offshore wind generation, IEEE T. Energy Conver., 30 (2015), 918–926. |
[4] | J. Lin, Integrating the first HVDC-based offshore wind power into PJM system–A real project case study, IEEE T. Ind. Appl., 52 (2016), 1970–1978. |
[5] | B. B. Shao, S. Q. Zhao, Y. H. Yang, B. F. Gao, F. Blaabjerg, Sub-synchronous oscillation characteristics and analysis of direct-drive wind farms with VSC-HVDC systems, IEEE T. Sustain. Energ., 12 (2021), 1127–1140. |
[6] | S. B. Wang, X. Y. Meng, T. W. Chen, Wide-area control of power systems through delayed network communication, IEEE T. Contr. Syst. T., 20 (2012), 495–503. |
[7] | B. Chaudhuri, R. Majumder, B. C. Pal, Wide-area measurement-based stabilizing control of power system considering signal transmission delay, IEEE T. Power Syst., 19 (2004), 1971–1979. |
[8] | B. Yang, Y. Z. Sun, Damping factor based delay margin for wide area signals in power system damping control, IEEE T. Power Syst., 28 (2013), 3501–3502. |
[9] | M. Li, Y. Chen, Wide-area stabiliser on sliding mode control for cross-area power systems with random delay and packet dropouts, Asian J. Control, 20 (2018), 2130–2142. |
[10] | K. T. Yu, Y. M. Li, Adaptive fuzzy control for nonlinear systems with sampled data and time-varying input delay, AIMS Mathematics, 5 (2020), 2307–2325. |
[11] | D. Yang, X. D. Li, J. L. Qiu, Output tracking control of delayed switched systems via state-dependent switching and dynamic output feedback, Nonlinear Anal.-Hybri., 32 (2019), 294–305. |
[12] | X. Y. Yang, X. D. Li, Q. Xi, P. Y. Duan, Review of stability and stabilization for impulsive delayed systems, Math. Biosci. Eng., 15 (2018), 1495–1515. |
[13] | W. W. Sun, M. Y. Qiu, X. Y. Lv, H∞ filter design for a class of delayed Hamiltonian systems with fading channel and sensor saturation, AIMS Mathematics, 5 (2020), 2909–2922. |
[14] | Y. Liu, N. Chopra, Gravity-compensation-driven position regulation for robotic systems under input/output delays, IEEE T. Contr. Syst. T., 22 (2014), 995–1005. |
[15] | S. Ahmed, H. P. Wang, M. S. Aslam, I. Ghous, I. Qaisar, Robust adaptive control of robotic manipulator with input time-varying delay, Int. J. Control, Autom. Syst., 17 (2019), 2193–2202. |
[16] | Y. Kang, Z. J. Li, X. Q. Cao, D. H. Zhai, Robust control of motion/force for robotic manipulators with random time delays, IEEE T. Contr. Syst. T., 21 (2013), 1708–1718. |
[17] | Q. He, J. K. Liu, An observer for a velocity-sensorless VTOL aircraft with time-varying measurement delay, Int. J. Syst. Sci., 47 (2016), 652–661. |
[18] | S. W. Su, Y. Lin, Output tracking control for a velocity-sensorless VTOL aircraft with measurement delays, Int. J. Syst. Sci., 46 (2015), 885–895. |
[19] | X. J. Jin, G. D. Yin, Y. J. Li, J. Q. Li, Stabilizing vehicle lateral dynamics with considerations of state delay of AFS for electric vehicles via robust gain-scheduling control, Asian J. Control, 18 (2016), 89–97. |
[20] | M. Z. Huang, W. N. Gao, Z. P. Jiang, Connected cruise control with delayed feedback and disturbance: An adaptive dynamic programming approach, Int. J. Adapt. Control., 33 (2019), 356–370. |
[21] | S. Y. Han, G. Y. Tang, Y. H. Chen, X. X. Yang, X. Yang, Optimal vibration control for vehicle active suspension discrete-time systems with actuator time delay, Asian J. Control, 15 (2013), 1579–1588. |
[22] | D. Soudbakhsh, A. Chakrabortty, A. M. Annaswamy, A delay-aware cyber-physical architecture for wide-area control of power systems, Control Eng. Pract., 60 (2017), 171–182. |
[23] | D. Efimov, J. Schiffer, R. Ortega, Robustness of delayed multistable systems with application to droop-controlled inverter-based microgrids, Int. J. Control, 89 (2016), 909–918. |
[24] | J. Li, Z. H. Chen, D. S. Cai, W. Zhen, Q. Huang, Delay-dependent stability control for power system with multiple time-delays, IEEE T. Power Syst., 31 (2016), 2316–2326. |
[25] | X. Li, R. Wang, S. Wu, S. N. Wu, G. M. Dimirovski, Exponential stability for multi-area power systems with time delays under load frequency controller failures, Asian J. Control, 19 (2017), 787–791. |
[26] | H. Luo, I. A. Hiskens, Z. Hu, Stability analysis of load frequency control systems with sampling and transmission delay, IEEE T. Power Syst., 35 (2020), 3603–3615. |
[27] | S. P. Wen, X. H. Yu, Z. G. Zeng, J. J. Wang, Event-triggering load frequency control for multiarea power systems with communication delays, IEEE T. Ind. Electron., 63 (2016), 1308–1317. |
[28] | Y. H. Gui, C. H. Kim, C. C. Chung, Improved low-voltage ride through capability for PMSG wind turbine based on port-controlled hamiltonian system, Int. J. Control Autom. Syst., 14 (2016), 1195–1204. |
[29] | H. H. Song, Y. B. Qu, Energy-based modelling and control of wind energy conversion system with DFIG, Int. J. Control, 84 (2011), 281–292. |
[30] | W. W. Sun, Y. Wu, L. P. Wang, Trajectory tracking of constrained robotic systems via a hybrid control strategy, Neurocomputing, 330 (2019), 188–195. |
[31] | Y. Ren, W. W. Sun, Robust adaptive control for robotic systems with input time-varying delay using Hamiltonian method, IEEE/CAA J. Automat., 5 (2018), 852–859. |
[32] | Y. Guo, Z. Xi, D. Cheng, Speed regulation of permanent magnet synchronous motor via feedback dissipative Hamiltonian realisation, IET Control Theory A., 1 (2007), 281–290. |
[33] | A. Yaghmaei, M. J. Yazdanpanah, Structure preserving observer design for port-Hamiltonian systems, IEEE T. Automat. Contr, 64 (2019), 1214–1220. |
[34] | X. M. Fan, L. Guan, C. J. Xia, T. Y. Ji, IDA-PB control design for VSC-HVDC transmission based on PCHD model, Int. T. Electr. Energy., 25 (2015), 2133–2143. |
[35] | S. Y. Dai, X. F. Koutsoukos, Safety analysis of integrated adaptive cruise and lane keeping control using multi-modal port-Hamiltonian systems, Nonlinear Anal.-Hybri., 35 (2020), 100816. |
[36] | S. Aoues, F. L. Cardoso-Ribeiro, D. Matignon, D. Alazard, Modeling and control of a rotating flexible spacecraft: A port-Hamiltonian approach, IEEE T. Contr. Syst. T., 27 (2019), 355–362. |
[37] | W. W. Sun, B. Z. Fu, Adaptive control of time-varying uncertain nonlinear systems with input delay: A Hamiltonian approach, IET Control Theory A., 10 (2016), 1844–1858. |
[38] | W. W. Sun, X. Y. Lv, K. L. Wang, L. P. Wang, Observer-based output feedback stabilisation and L2-disturbance attenuation of uncertain Hamiltonian systems with input and output delays, Int. J. Syst. Sci., 50 (2019), 2565–2578. |
[39] | J. Schiffer, E. Fridman, R. Ortega, J. Raischde, Stability of a class of delayed port-Hamiltonian systems with application to microgrids with distributed rotational and electronic generation, Automatica, 74 (2016), 71–79. |