Existence of weak solution to the nonstationary Navier-Stokes equations approximated by pressure stabilization method (Q2154517)

Starting from the nonstationary Navier-Stokes equation posed in a \(C^{2,1}\) domain \(\Omega \) of \(\mathbb{R}^{n}\), \(n\geq 2\), the authors introduce the Helmholtz projections \(P:L_{q}(\Omega )\rightarrow L_{q,\sigma }(\Omega )\) and \(Q:L_{q}(\Omega )\rightarrow \widehat{W}_{q}^{1}(\Omega )\), where \(L_{q,\sigma }(\Omega )\) is the closure of \(C_{0,\sigma }^{\infty }(\Omega )\) with respect to the norm \(\left\Vert \cdot \right\Vert _{q}\), \( \widehat{W}_{q}^{1}(\Omega )=\{\pi \in L_{q,\mathrm{loc}}(\Omega );\) \(\nabla \pi \in L_{q}(\Omega )^{n},\) \(\int_{\Omega }\pi (x)dx=0\}\) and \(L_{q,\mathrm{loc}}(\Omega )=\{\pi ;\) \(\pi \mid _{K}\in L_{q}(K)\) for every compact \(K\) in \(\Omega \}\). They apply the pressure stabilization method with a parameter \(\alpha >0\) and finally consider the problem \(\partial _{t}u_{\alpha }-\Delta u_{\alpha }+B(u_{\alpha },u_{\alpha })+\alpha \nabla Qu_{\alpha }=f\), \(t\in (0,\infty ) \), \(x\in \Omega \), where \(B(u,u)=(Pu\cdot \nabla )u\), with the homogeneous Dirichlet boundary conditions \(u_{\alpha }=0\), \(t\in (0,\infty )\), \(x\in \partial \Omega \), the solution starting from an initial condition \(a\) at \( t=0\). The first main result proves that for any \(a_{\alpha }\in L_{2}(\Omega )\) and \(f\in L_{2}((0,T),L_{2}(\Omega ))\), there exists a weak solution \( u_{\alpha }\) to this problem such that the following estimate holds: \( \left\Vert u_{\alpha }\right\Vert _{\infty ,2,T}^{2}+\left\Vert A_{\alpha }^{1/2}u_{\alpha }\right\Vert _{2,2,T}^{2}\leq \left\Vert a_{\alpha }\right\Vert _{2}^{2}+\int_{0}^{t}(f(s),u_{\alpha }(s))_{\Omega }ds\), where \( A_{\alpha }\) is the approximated Stokes operator \(A_{\alpha }u=-\Delta u+\alpha \nabla Qu\), \(A\) being the usual Stokes operator. Assuming that \(p\) and \(q\) satisfy \(1<p,q<\infty \) and \(2/p+n/q=n+1\), \(a_{\alpha }\in (L_{q},D(A_{\alpha }))_{1-1/p,p}\), \(f\in L_{p}((0,T),L_{q}(\Omega ))\), and \( u_{\alpha }\) is one weak solution to the above problem which does not satisfy the preceding energy estimate, the authors prove estimates on such solution with weight \(e^{\gamma t}\) with \(\gamma \) small enough. In the 2D case, the authors prove that the weak solution to the above problem is unique if \(a_{\alpha }\in H^{1}(\Omega )\) and \(f\in L_{2}((0,T),L_{2}(\Omega ))\). This solution is such that \(A_{\alpha }^{1/2}u_{\alpha }\in L_{2}((0,T),L_{2}(\Omega ))\), \(u_{\alpha }\in W_{2}^{1}((0,T),L_{2}(\Omega ))\cap L_{2}((0,T),H^{2}(\Omega ))\) and \(\pi _{\alpha }\in L_{2}((0,T),W_{2}^{1}(\Omega ))\). For the proofs, the authors use properties of the Helmholtz projection. They introduce the \(L_{p}-L_{q}\) maximal regularity estimate for solutions to the linearized problem associated to the above one. They also use Sobolev's embeddings and various inequalities in such spaces.