Part I. Let
\(e^{p}=p(U_{N})-P_{N}\) and
\(e_{I}^{p}=P^{0}_{1,N}e ^{p}\in V^{N}\), where
\(P^{0}_{1,N}\) is the orthogonal projection operator defined as in Lemma
2.2. Note that
\((p(U_{N})-P _{N})(x,T)=0\), hence
$$\begin{aligned} & \int_{0}^{T}-\bigl(p_{t}(U_{N})-P_{Nt},e^{p} \bigr)\,dt \\ &\quad =- \int_{0}^{T} \frac{1}{2}\frac{d}{d t} \bigl\Vert p(U_{N})-P_{N} \bigr\Vert _{L^{2}(\Omega )}^{2}\,dt \\ &\quad=- \frac{1}{2} \bigl( \bigl\Vert \bigl(p(U_{N})-P_{N} \bigr) (x,T) \bigr\Vert _{L^{2}(\Omega )}^{2}- \bigl\Vert \bigl(p(U _{N})-P_{N}\bigr) (x,0) \bigr\Vert _{L^{2}(\Omega )}^{2} \bigr) \\ &\quad=\frac{1}{2} \bigl\Vert \bigl(p(U _{N})-P_{N} \bigr) (x,0) \bigr\Vert _{L^{2}(\Omega )}^{2}\geq 0. \end{aligned}$$
(3.15)
By using Eqs. (
2.27), (
3.6), and (
3.15), we have
$$\begin{aligned} \begin{aligned}[b] & c \bigl\Vert e^{p} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2} \\ &\quad\leq \int_{0}^{T}a\bigl(e ^{p},p(U_{N})-P_{N} \bigr)\,dt+ \int_{0}^{T}\bigl(\phi ' \bigl(y(U_{N})\bigr) \bigl(p(U_{N})-P_{N}\bigr),e ^{p}\bigr)\,dt \\ &\quad\leq \int_{0}^{T}\bigl(\nabla e^{p},A^{*} \nabla \bigl(p(U_{N})-P_{N}\bigr)\bigr)\,dt- \int_{0}^{T}\bigl(p_{t}(U_{N})-P_{Nt},e^{p} \bigr)\,dt \\ &\quad\quad {} + \int_{0}^{T}\bigl(\phi '\bigl(y(U _{N})\bigr) \bigl(p(U_{N})-\tilde{P}_{N} \bigr),e^{p}\bigr)\,dt \\ &\quad\quad {} + \int_{0}^{T}\bigl(\phi '\bigl(y(U _{N})\bigr) (\tilde{P}_{N}-P_{N}),e^{p} \bigr)\,dt \\ &\quad= \int_{0}^{T}\bigl(\nabla e^{p},A ^{*}\nabla \bigl(p(U_{N})-\tilde{P}_{N}\bigr) \bigr)\,dt- \int_{0}^{T}\bigl(p_{t}(U_{N})-P _{Nt},e^{p}\bigr)\,dt \\ &\quad\quad {} + \int_{0}^{T}\bigl(\nabla e^{p},A^{*} \nabla (\tilde{P} _{N}-P_{N})\bigr)\,dt \\ &\quad\quad {} + \int_{0}^{T}\bigl(\phi ' \bigl(y(U_{N})\bigr)p(U_{N})-\phi '( \tilde{Y}_{N})\tilde{P}_{N},e^{p}\bigr)\,dt \\ &\quad\quad {} + \int_{0}^{T}\bigl(\tilde{\phi }''( \tilde{Y}_{N}) \bigl(\tilde{Y}_{N}-y(U_{N}) \bigr)\tilde{P}_{N},e^{p}\bigr)\,dt \\ &\quad\quad {} + \int _{0}^{T}\bigl(\phi ' \bigl(y(U_{N})\bigr) (\tilde{P}_{N}-P_{N}),e^{p} \bigr)\,dt. \end{aligned} \end{aligned}$$
(3.16)
From (
3.16), we can get
$$\begin{aligned} \begin{aligned}[b] & c \bigl\Vert e^{p} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2} \\ &\quad\leq \int_{0}^{T}\bigl( \nabla \bigl(e^{p}-e_{I}^{p} \bigr),A^{*}\nabla \bigl(p(U_{N})-\tilde{P}_{N} \bigr)\bigr)\,dt- \int _{0}^{T}\bigl(p_{t}(U_{N})-P_{Nt},e^{p}-e_{I}^{p} \bigr)\,dt \\ &\quad\quad {} + \int_{0}^{T}\bigl( \phi ' \bigl(y(U_{N})\bigr)p(U_{N})-\phi '( \tilde{Y}_{N})\tilde{P}_{N},e^{p}-e _{I}^{p}\bigr)\,dt \\ &\quad \quad {} + \int_{0}^{T}\bigl(\nabla e_{I}^{p},A^{*} \nabla \bigl(p(U_{N})- \tilde{P}_{N}\bigr)\bigr)\,dt- \int_{0}^{T}\bigl(p_{t}(U_{N})-P_{Nt},e_{I}^{p} \bigr)\,dt \\ &\quad \quad {} + \int_{0}^{T}\bigl(\phi ' \bigl(y(U_{N})\bigr)p(U_{N})-\phi '( \tilde{Y}_{N})\tilde{P} _{N},e_{I}^{p} \bigr)\,dt \\ &\quad \quad {} + \int_{0}^{T}\bigl(\tilde{\phi }''( \tilde{Y}_{N}) \bigl( \tilde{Y}_{N}-y(U_{N}) \bigr)\tilde{P}_{N},e^{p}\bigr)\,dt \\ &\quad\quad {} + \int_{0}^{T}\bigl(\phi '\bigl(y(U _{N})\bigr) (\tilde{P}_{N}-P_{N}),e^{p} \bigr)\,dt+ \int_{0}^{T}\bigl(\nabla e^{p},A^{*} \nabla (\tilde{P}_{N}-P_{N})\bigr)\,dt. \end{aligned} \end{aligned}$$
(3.17)
Thanks to
\(e^{p}-e_{I}^{p}\in V=H_{0}^{1}(\Omega )\), from Eqs. (
2.27), (
3.6), and (
3.17), we can obtain
$$\begin{aligned} \begin{aligned}[b] & c \bigl\Vert e^{p} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2} \\ &\quad= \int_{0}^{T}\bigl(y(U_{N})-y _{d}+\operatorname{div}\bigl(A^{*}\nabla \tilde{P}_{N} \bigr)-\phi '(\tilde{Y}_{N}) \tilde{P}_{N}+P_{Nt},e^{p}-e_{I}^{p} \bigr)\,dt \\ &\quad\quad {} + \int_{0}^{T}\bigl(\tilde{\phi }''( \tilde{Y}_{N}) \bigl(\tilde{Y}_{N}-y(U_{N}) \bigr)\tilde{P}_{N},p(U_{N})-P_{N}\bigr)\,dt \\ &\quad\quad {} + \int_{0}^{T}\bigl(y(U_{N})- \hat{Y}_{N},e_{I}^{p}\bigr)\,dt+ \int_{0}^{T}\bigl( \hat{y}_{d}-y_{d},e_{I}^{p} \bigr)\,dt \\ &\quad\quad {} + \int_{0}^{T}\bigl(\phi ' \bigl(y(U_{N})\bigr) ( \tilde{P}_{N}-P_{N}),p(U_{N})-P_{N} \bigr)\,dt \\ &\quad\quad {}+ \int_{0}^{T}\bigl(\nabla e^{p},A ^{*}\nabla (\tilde{P}_{N}-P_{N})\bigr)\,dt. \end{aligned} \end{aligned}$$
(3.18)
Then we have
$$\begin{aligned} \begin{aligned}[b] & c \bigl\Vert e^{p} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2} \\ &\quad\leq \int_{0}^{T}\bigl( \hat{Y}_{N}- \hat{y}_{d}+\operatorname{div}\bigl(A^{*}\nabla \tilde{P}_{N}\bigr)-\phi '( \tilde{Y}_{N}) \tilde{P}_{N}+P_{Nt},e^{p}-e_{I}^{p} \bigr)\,dt \\ &\quad\quad {} + \int_{0} ^{T}\bigl(\phi ' \bigl(y(U_{N})\bigr) (\tilde{P}_{N}-P_{N}),p(U_{N})-P_{N} \bigr)\,dt+ \int_{0} ^{T}\bigl(\nabla e^{p},A^{*} \nabla (\tilde{P}_{N}-P_{N})\bigr)\,dt \\ &\quad\quad {} + \int_{0} ^{T}\bigl(\hat{y}_{d}-y_{d},e^{p} \bigr)\,dt+ \int_{0}^{T}\bigl(\tilde{\phi }''( \tilde{Y}_{N}) \bigl(\tilde{Y}_{N}-y(U_{N}) \bigr)\tilde{P}_{N},P(U_{N})-P_{N}\bigr)\,dt \\ &\quad\quad {} + \int_{0}^{T}\bigl(y(U_{N})- \hat{Y}_{N},e^{p}\bigr)\,dt \\ &\quad \equiv \sum_{i=1}^{6}I_{i}. \end{aligned} \end{aligned}$$
(3.19)
By using Lemma
2.2, we can estimate the first term
\(I_{1}\) as follows:
$$\begin{aligned} \begin{aligned}[b] I_{1}&= \int_{0}^{T}\bigl(\hat{Y}_{N}- \hat{y}_{d}+\operatorname{div}\bigl(A^{*}\nabla \tilde{P}_{N}\bigr)-\phi '(\tilde{Y}_{N}) \tilde{P}_{N}+P_{Nt},e^{p}-e_{I} ^{p}\bigr)\,dt \\ &\leq C(\delta ) \int_{0}^{T}N^{-2} \int_{\Omega }\bigl(\hat{Y} _{N}-\hat{y}_{d}+ \operatorname{div}\bigl(A^{*}\nabla \tilde{P}_{N}\bigr)-\phi '( \tilde{Y}_{N})\tilde{P}_{N}+P_{Nt} \bigr)^{2}\,dx\,dt \\ &\quad {} +\delta \int_{0}^{T} \bigl\Vert e ^{p} \bigr\Vert _{H^{1}(\Omega )}^{2}\,dt \\ &\leq C(\delta )\eta_{2}^{2}+\delta \bigl\Vert p(U_{N})-P_{N} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2}, \end{aligned} \end{aligned}$$
(3.20)
where
δ is an arbitrary positive number,
\(C(\delta )\) is a constant dependent on
δ. Note that
\(\phi '(\cdot )\geq 0\), we can obtain
$$\begin{aligned} I_{2}&= \int_{0}^{T}\bigl(\phi ' \bigl(y(U_{N})\bigr) (\tilde{P}_{N}-P_{N}),p(U_{N})-P _{N}\bigr)\,dt \\ &\leq C(\delta ) \Vert \tilde{P}_{N}-P_{N} \Vert _{L^{2}(0,T;L^{2}( \Omega ))}^{2}+\delta \bigl\Vert p(U_{N})-P_{N} \bigr\Vert _{L^{2}(0,T;L^{2}(\Omega ))} ^{2} \\ &\leq C(\delta ) \int_{0}^{T} \int_{\Omega } \bigl\vert A^{*}\nabla ( \tilde{P}_{N}-P_{N}) \bigr\vert ^{2}\,dx\,dt+\delta \bigl\Vert p(U_{N})-P_{N} \bigr\Vert _{L^{2}(0,T;H ^{1}(\Omega ))}^{2} \\ &\leq C(\delta )\eta_{3}^{2}+\delta \bigl\Vert p(U_{N})-P _{N} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2}. \end{aligned}$$
(3.21)
For the third term
\(I_{3}\), we can derive
$$\begin{aligned} \begin{aligned}[b] I_{3}&= \int_{0}^{T}\bigl(\nabla e^{p},A^{*} \nabla (\tilde{P}_{N}-P_{N})\bigr)\,dt \\ &\leq C(\delta ) \int_{0}^{T} \int_{\Omega } \bigl\vert A^{*}\nabla (\tilde{P} _{N}-P_{N}) \bigr\vert ^{2}\,dx\,dt+\delta \int_{0}^{T} \int_{\Omega } \bigl\vert \nabla e^{p} \bigr\vert ^{2}\,dx\,dt \\ &\leq C(\delta )\eta_{3}^{2}+\delta \bigl\Vert p(U_{N})-P_{N} \bigr\Vert _{L^{2}(0,T;H ^{1}(\Omega ))}^{2}. \end{aligned} \end{aligned}$$
(3.22)
Similarly, for
\(I_{4}\) we have the following estimate:
$$\begin{aligned} \begin{aligned}[b] I_{4}&= \int_{0}^{T}\bigl(\hat{y}_{d}-y_{d},e^{p} \bigr)\,dt \\ &\leq C(\delta ) \Vert y _{d}-\hat{y}_{d} \Vert _{L^{2}(0,T;L^{2}(\Omega ))}^{2}+\delta \bigl\Vert p(U_{N})-P _{N} \bigr\Vert _{L^{2}(0,T;L^{2}(\Omega ))}^{2} \\ &\leq C(\delta )\eta^{2}_{4}+ \delta \bigl\Vert p(U_{N})-P_{N} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2}. \end{aligned} \end{aligned}$$
(3.23)
Due to
\(\phi (\cdot )\in W^{2,\infty }(-R,-R)\), we can get
$$\begin{aligned} \begin{aligned}[b] I_{5}&= \int_{0}^{T}\bigl(\tilde{\phi }''( \tilde{Y}_{N}) \bigl(\tilde{Y}_{N}-y(U _{N}) \bigr)\tilde{P}_{N},p(U_{N})-P_{N}\bigr)\,dt \\ &\leq C(\delta ) \int_{0}^{T} \int_{\Omega } \bigl\vert y(U_{N})- \tilde{Y}_{N} \bigr\vert ^{2}\,dx\,dt+\delta \bigl\Vert p(U_{N})-P _{N} \bigr\Vert _{L^{2}(0,T;L^{2}(\Omega ))}^{2} \\ &\leq C(\delta ) \bigl\Vert y(U_{N})-Y _{N} \bigr\Vert _{L^{2}(0,T;L^{2}(\Omega ))}^{2}+C(\delta ) \Vert Y_{N}- \tilde{Y} _{N} \Vert _{L^{2}(0,T;L^{2}(\Omega ))}^{2} \\ & \quad {} +\delta \bigl\Vert p(U_{N})-P_{N} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2} \\ &\leq C(\delta )\eta_{5}^{2}+C( \delta ) \bigl\Vert y(U_{N})-Y_{N} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2}+ \delta \bigl\Vert p(U _{N})-P_{N} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2}. \end{aligned} \end{aligned}$$
(3.24)
Then we can estimate the last term
\(I_{6}\) of (
3.19) as follows:
$$\begin{aligned} \begin{aligned}[b] I_{6}&= \int_{0}^{T}\bigl(y(U_{N})- \hat{Y}_{N},e^{p}\bigr)\,dt \\ &\leq C(\delta ) \int_{0}^{T} \int_{\Omega } \bigl\vert y(U_{N})- \hat{Y}_{N} \bigr\vert ^{2}\,dx\,dt+\delta \bigl\Vert p(U _{N})-P_{N} \bigr\Vert _{L^{2}(0,T;L^{2}(\Omega ))}^{2} \\ &\leq C(\delta ) \bigl\Vert y(U _{N})-Y_{N} \bigr\Vert _{L^{2}(0,T;L^{2}(\Omega ))}^{2}+C(\delta ) \Vert Y_{N}- \hat{Y}_{N} \Vert _{L^{2}(0,T;L^{2}(\Omega ))}^{2} \\ &\quad {} +\delta \bigl\Vert p(U_{N})-P _{N} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2} \\ &\leq C(\delta )\eta_{6}^{2}+C( \delta ) \bigl\Vert y(U_{N})-Y_{N} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2}+ \delta \bigl\Vert p(U _{N})-P_{N} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2} . \end{aligned} \end{aligned}$$
(3.25)
Therefore, let
δ be small enough, from (
3.16)–(
3.25), we obtain
$$ \bigl\Vert p(U_{N})-P_{N} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2}\leq C(\delta ) \sum_{i=2}^{6} \eta_{i}^{2}+C(\delta ) \bigl\Vert y(U_{N})-Y_{N} \bigr\Vert _{L^{2}(0,T;H ^{1}(\Omega ))}^{2}. $$
(3.26)
Part II. Let
\(e^{y}=y(U_{N})-Y_{N}\),
\(e_{I}^{y}=P^{0}_{1,N} e ^{y}\in V^{N}\), where
\(P^{0}_{1,N}\) is the orthogonal projection operator defined as in Lemma
2.2. Note that
$$\begin{aligned} \begin{aligned}[b] \int_{0}^{T}\bigl(y_{t}(U_{N})-Y_{Nt},e^{y} \bigr)\,dt&= \int_{\Omega } \int_{0}^{T}e ^{y}\bigl(y_{t}(U_{N})-Y_{Nt} \bigr)\,dt\,dx \\ &= \int_{\Omega } \int_{0}^{T}e^{y}\frac{d(y(U_{N})-Y_{N})}{dt}\,dt\,dx \\ &= \int_{\Omega } \int_{0}^{T}e^{y}\,d\bigl(y(U_{N})-Y_{N} \bigr)\,dx \\ &= \int_{\Omega }\bigl(\bigl(y(U_{N})-Y_{N}\bigr) (x,T)\bigr)^{2}\,dx- \int_{\Omega }\bigl(\bigl(y(U _{N})-Y_{N} \bigr) (x,0)\bigr)^{2}\,dx \\ &\quad {} - \int_{0}^{T}\bigl(y_{t}(U_{N})-Y_{Nt},e^{y} \bigr)\,dt, \end{aligned} \end{aligned}$$
then we have
$$\begin{aligned} \begin{aligned} \int_{0}^{T}\bigl(y_{t}(U_{N})-Y_{Nt},e^{y} \bigr)\,dt&=\frac{1}{2} \int_{\Omega }\bigl(\bigl(y(U _{N})-Y_{N} \bigr) (x,T)\bigr)^{2}\,dx\\&\quad{} -\frac{1}{2} \int_{\Omega }\bigl(\bigl(y(U_{N})-Y_{N}\bigr) (x,0)\bigr)^{2}\,dx.\end{aligned} \end{aligned}$$
Thus
$$\begin{aligned} \int_{0}^{T}\bigl(y_{t}(U_{N})-Y_{Nt},e^{y} \bigr)\,dt+\frac{1}{2} \bigl\Vert y_{0}(x)-Y_{N}(x,0) \bigr\Vert _{L^{2}(\Omega )}^{2}\geq 0. \end{aligned}$$
(3.27)
From (
2.30), it is easy to show that
$$\begin{aligned} \bigl(\phi \bigl(y(U_{N})\bigr)-\phi (Y_{N}),e^{y}\bigr)=\bigl(\tilde{\phi }' \bigl(y(U_{N})\bigr) \bigl(y(U _{N})-Y_{N} \bigr),e^{y}\bigr)\geq 0. \end{aligned}$$
(3.28)
By using (
3.27), we can get
$$\begin{aligned} \begin{aligned}[b] c \bigl\Vert e^{y} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2}&\leq \int_{0}^{T}a\bigl(y(U_{N})-Y _{N},e^{y}\bigr)\,dt+ \int_{0}^{T}\bigl(\phi \bigl(y(U_{N}) \bigr)-\phi (Y_{N}),e^{y}\bigr)\,dt \\ &\leq \int_{0}^{T}a\bigl(y(U_{N})-Y_{N},e^{y} \bigr)\,dt+ \int_{0}^{T}\bigl(\phi \bigl(y(U _{N}) \bigr)-\phi (\hat{Y}_{N}),e^{y}\bigr)\,dt \\ & \quad {} + \int_{0}^{T}\bigl(y_{t}(U_{N})-Y_{Nt},e ^{y}\bigr)\,dt+ \int_{0}^{T}\bigl(\phi (\hat{Y}_{N})-\phi (Y_{N}),e^{y}\bigr)\,dt \\ & \quad {} + \frac{1}{2} \bigl\Vert y_{0}(x)-Y_{N}(x,0) \bigr\Vert _{L^{2}(\Omega )}^{2} \\ &= \int_{0} ^{T}\bigl(A\nabla \bigl(y(U_{N})- \hat{Y}_{N}\bigr),\nabla e^{y}\bigr)\,dt+ \int_{0}^{T}\bigl( \phi \bigl(y(U_{N}) \bigr)-\phi (\hat{Y}_{N}),e^{y}\bigr)\,dt \\ & \quad {} + \int_{0}^{T}\bigl(A\nabla (\hat{Y}_{N}-Y_{N}), \nabla e^{y}\bigr)\,dt+ \int_{0}^{T}\bigl(y_{t}(U_{N})-Y_{Nt},e ^{y}\bigr)\,dt \\ &\quad {} + \int_{0}^{T}\bigl(\phi (\hat{Y}_{N})-\phi (Y_{N}),e^{y}\bigr)\,dt+ \frac{1}{2} \bigl\Vert y_{0}(x)-Y_{N}(x,0) \bigr\Vert _{L^{2}(\Omega )}^{2}. \end{aligned} \end{aligned}$$
(3.29)
Combining (
3.28) and (
3.29), note that
\(e^{y}-e_{I}^{y} \in H_{0}^{1}(\Omega )\), from (
2.25) and (
3.4), we can obtain
$$\begin{aligned} c \bigl\Vert e^{y} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2}&\leq \int_{0}^{T}\bigl(\hat{f}+BU _{N}+ \operatorname{div}(A\nabla \hat{Y}_{N})-\phi (\tilde{Y}_{N})-Y_{Nt},e^{y}-e _{I}^{y}\bigr)\,dt \\ & \quad {} + \int_{0}^{T}\bigl(\tilde{\phi }'( \hat{Y}_{N}) (\hat{Y}_{N}-Y _{N}),e^{y} \bigr)\,dt+ \int_{0}^{T}\bigl(A\nabla (\hat{Y}_{N}-Y_{N}), \nabla e^{y}\bigr)\,dt \\ & \quad {} + \int_{0}^{T}\bigl(f-\hat{f},e^{y} \bigr)\,dt+\frac{1}{2} \bigl\Vert Y_{N}(x,0)-y_{0}(x) \bigr\Vert _{L^{2}(\Omega )}^{2} \\ &\equiv J_{1}+J_{2}+J_{3}+J_{4}+ \frac{1}{2} \eta_{10}^{2}. \end{aligned}$$
(3.30)
From Lemma (
2.2), we have
$$\begin{aligned} \begin{aligned}[b] J_{1}&= \int_{0}^{T}\bigl(\hat{f}+BU_{N}+ \operatorname{div}(A\nabla \hat{Y}_{N})- \phi (\tilde{Y}_{N})-Y_{Nt},e^{y}-e_{I}^{y} \bigr)\,dt \\ &\leq C(\delta ) \int _{0}^{T}N^{-2} \int_{\Omega }\bigl(\hat{f}+BU_{N}+\operatorname{div}(A\nabla \hat{Y} _{N})-\phi (\tilde{Y}_{N})-Y_{Nt} \bigr)^{2}\,dx\,dt \\ & \quad {} +\delta \int_{0}^{T} \bigl\Vert e ^{y} \bigr\Vert _{H^{1}(\Omega )}^{2}\,dt \\ &\leq C(\delta )\eta_{7}^{2}+\delta \bigl\Vert y(U_{N})-Y_{N} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2}. \end{aligned} \end{aligned}$$
(3.31)
For
\(J_{2}\), by using the assumption of
ϕ, we can get
$$\begin{aligned} \begin{aligned}[b] J_{2}&= \int_{0}^{T}\bigl(\tilde{\phi }'( \hat{Y}_{N}) (\hat{Y}_{N}-Y_{N}),e ^{y}\bigr)\,dt \\ &\leq C(\delta ) \Vert Y_{N}-\hat{Y}_{N} \Vert _{L^{2}(0,T;L^{2}( \Omega ))}^{2}+\delta \bigl\Vert e^{y} \bigr\Vert _{L^{2}(0,T;L^{2}(\Omega ))}^{2} \\ &\leq C(\delta )\eta_{6}^{2}+\delta \bigl\Vert e^{y} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2}. \end{aligned} \end{aligned}$$
(3.32)
For
\(J_{3}\), it is clear that
$$\begin{aligned} \begin{aligned}[b] J_{3}&= \int_{0}^{T}\bigl(A\nabla (\hat{Y}_{N}-Y_{N}), \nabla e^{y}\bigr)\,dt \\ &\leq C(\delta ) \int_{0}^{T} \int_{\Omega } \bigl\vert A\nabla (\hat{Y}_{N}-Y_{N}) \bigr\vert ^{2}\,dx\,dt+ \delta \int_{0}^{T} \int_{\Omega } \bigl\vert A\nabla e^{y} \bigr\vert ^{2}\,dx\,dt \\ &\leq C( \delta )\eta_{8}^{2}+\delta \bigl\Vert y(U_{N})-Y_{N} \bigr\Vert _{L^{2}(0,T;H^{1}( \Omega ))}^{2}. \end{aligned} \end{aligned}$$
(3.33)
For
\(J_{4}\), we can obtain
$$\begin{aligned} \begin{aligned}[b] J_{4}&= \int_{0}^{T}\bigl(f-\hat{f},e^{y}\bigr)\,dt \\ &\leq C(\delta ) \Vert f-\hat{f} \Vert _{L^{2}(0,T;L^{2}(\Omega ))}^{2}+\delta \bigl\Vert y(U_{N})-Y_{N} \bigr\Vert _{L^{2}(0,T;L ^{2}(\Omega ))}^{2} \\ &\leq C(\delta )\eta_{9}^{2}+\delta \bigl\Vert y(U_{N})-Y _{N} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2}. \end{aligned} \end{aligned}$$
(3.34)
By simplifying both sides of (
3.30), we get
$$ \bigl\Vert y(U_{N})-Y_{N} \bigr\Vert _{L^{2}(0,T;H^{1}(\Omega ))}^{2}\leq C(\delta ) \sum_{i=6}^{10} \eta_{i}^{2}. $$
(3.35)
Then we derive (
3.14) follows from (
3.26) and (
3.35). □