# 偏微分方程复习总结|Review for Partial Differential Equations

## parabolic equation deduction

h e a t   f l u x   i n = − K A ∂ T ∂ x ∣ x , t ( k   i s   a   p h y s i c s   c o n s t a n t ) h e a t   f l u x   i n = − K A ∂ T ∂ x ∣ x + Δ t , t c h a n g e   o f   h e a t   p e r   t : c v A Δ x ∂ T ∂ t W e   c o u l d   w r i t e   t e h   b a l a n c e   e q u a t i o n : c v A Δ x ∂ T ∂ t = − K A ∂ T ∂ x ∣ x , t + K A ∂ T ∂ x ∣ x + Δ t , t ∂ T ∂ t = k c v 1 Δ x ( ∂ T ∂ x ∣ x + Δ t , t − ∂ T ∂ x ∣ x , t ) t a k e   t h e   l i m i t : ∂ T ∂ t = k c v ∂ 2 T ∂ x 2 s e l e c t   n o n d i m e n s i o n a l   v a r i a b l e s   x ‾   t ‾   T ‾ x ‾ = x L      T ‾ = t t s c a l e    T ‾ = T T 0 ∂ T ∂ t = ∂ T ‾ T 0 ∂ t ‾ t s c a l e = T 0 t s c a l e ∂ T ‾ ∂ t ‾ ∂ 2 T ∂ x 2 = ∂ 2 T ‾ T 0 ∂ ( x ‾ L ) 2 = T 0 L 2 ∂ 2 T ‾ ∂ x ‾ 2 T 0 t s c a l e ∂ T ‾ ∂ t ‾ = T 0 L 2 ∂ 2 T ‾ ∂ x ‾ 2 ∗ k c v ∴ ∂ T ‾ ∂ t ‾ = ∂ 2 T ‾ ∂ x ‾ 2   f o r   t s c a l e = c v L 2 k heat~flux~in=-KA\frac{\partial T}{\partial x}|_{x,t}(k~is~a~physics~constant)\\ heat~flux~in=-KA\frac{\partial T}{\partial x}|_{x+\Delta t,t}\\ change ~of~heat~per~t:c_vA\Delta x\frac{\partial T}{\partial t}\\ We~could~write~teh~balance~equation:\\ c_vA\Delta x\frac{\partial T}{\partial t}=-KA\frac{\partial T}{\partial x}|_{x,t} +KA\frac{\partial T}{\partial x}|_{x+\Delta t,t}\\ \frac{\partial T}{\partial t}=\frac{k}{c_v}\frac{1}{\Delta x}(\frac{\partial T}{\partial x}|_{x+\Delta t,t}-\frac{\partial T}{\partial x}|_{x,t})\\ take~the~limit:\frac{\partial T}{\partial t}=\frac{k}{c_v}\frac{\partial^2 T}{\partial x^2}\\ select~nondimensional~variables~\overline{x} ~\overline{t}~\overline{T}\\ \overline{x}=\frac{x}{L}~~~~\overline{T}=\frac{t}{t_scale}~~\overline{T}=\frac{T}{T_0}\\ \frac{\partial T}{\partial t}=\frac{\partial \overline{T}T_0}{\partial \overline{t}t_scale}=\frac{T_0}{t_scale}\frac{\partial \overline{T}}{\partial\overline{t}} \\ \frac{\partial^2 T}{\partial x^2}=\frac{\partial^2 \overline{T}T_0}{\partial (\overline{x}L)^2}=\frac{T_0}{L^2}\frac{\partial^2 \overline{T}}{\partial\overline{x}^2}\\ \frac{T_0}{t_scale}\frac{\partial \overline{T}}{\partial\overline{t}}=\frac{T_0}{L^2}\frac{\partial^2 \overline{T}}{\partial\overline{x}^2}*\frac{k}{c_v}\\ \therefore \frac{\partial \overline{T}}{\partial\overline{t}}=\frac{\partial^2 \overline{T}}{\partial\overline{x}^2}~for~t_{scale}=\frac{c_vL^2}{k}

## heat-flow problem solution by separation law

B y   s e p e r a t i o n   l a w ,   a s s u m e    u ( x , t ) = X ( t ) T ( t ) u n ( x , t ) = X n ( t ) T n ( t ) u = ∑ n = 1 ∞ u i ( x , t ) = ∑ n = 1 ∞ a i X i ( t ) T i ( t ) ∂ u ∂ t = X d T d t ,    ∂ 2 u ∂ X 2 = T d 2 X d x ,    ∂ u ∂ t = ∂ 2 u ∂ x 2 X d T d t = T d 2 X d x → 1 T d T d t = 1 x d 2 X d x 2 = C   ( w h e r e   C   i s   a   c o n s t a n t ) C a s e   1 : p o s i t i v e   c o n s t a n t   λ 2 > 0 1 T d T d t = 1 x d 2 X d x 2 = λ 2 d T d t = λ 2 T → T ( t ) = A e λ 2 t d 2 X d x 2 = λ 2 x → X ( x ) = B e λ x + C e − λ x ∴ u = A e λ 2 t ( B e λ x + C e − λ x ) u ( 0 , t ) = e λ 2 t ( B + C ) = 0 → B + C = 0 u ( 1 , t ) = e λ 2 t ( B e λ + C e − λ ) = 0 i f λ = 0 , i t   i m p l i e s   u   i s   i d e n t i c a l l y   0 i f λ ≠ 0 , ( B e λ + C e − λ ) = 0 w h i c h   d o e s n ′ t   e x i s t . C a s e   2 : c o n s t a n t = 0 1 T d T d t = 0 → T = A 1 x d 2 X d x 2 = 0 → X = B x + C u ( 0 , t ) = A C = 0 u ( 1 , t ) = A B + A C = 0 → A B = 0 ∴ u ( x , t ) = 0 C a s e   3 : c o n s t a n t   i s   n e g a t i v e 1 T d T d t = − λ 2 → d T d t = − λ 2 T ∴ T ( t ) = e − λ 2 t 1 x d 2 X d x 2 = − λ 2 → d 2 X d x 2 = − λ 2 x ∴ X ( x ) = B c o s ( λ x ) + C s i n ( λ x ) ∴ u ( x , t ) = e − λ 2 t ( B c o s ( λ x ) + C s i n ( λ x ) ) A c o o r d i n g   t o   t h e   b o u n d a r y   c o n d i t i o n   u ( 0 , t ) = e − λ 2 t B = 0 → B = 0 u ( 1 , t ) = e − λ 2 t C c o s ( λ x ) = 0 → C = 0 o r   λ = n π , n = 0 , 1 , 2 , 3... u ( x , t ) = ∑ n = 1 ∞ a n s i n ( λ n x ) e − λ n 2 t A c o o r d i n g   t o   t h e   i n i t i a l   c o n d i t i o n   u ( x , 0 ) = ∑ n = 1 ∞ a n s i n ( λ n x ) = ϕ ( x ) A c c o r d i n g   t o   t h e   t h e o r e m ϕ ( x ) = ∑ n = 1 ∞ a n e 0 s i n ( λ n x ) = ∑ n = 1 ∞ a n s i n ( λ n x ) ∫ 0 ∞ ϕ ( x ) s i n ( λ n x ) d x = ∑ n = 1 ∞ ∫ 0 ∞ a n s i n ( λ n x ) s i n ( λ n x ) d x = ∑ n = 1 ∞ ∫ 0 ∞ a n ( s i n 2 λ n x ) d x = a n 2 ∴ a n = 2 ∫ 0 1 ϕ ( x ) s i n ( λ n x ) d x By~seperation~law, ~assume ~~ u(x,t)=X(t)T(t)\\ u_n(x,t)=X_n(t)T_n(t)\\ u=\sum_{n=1}^{\infin}u_i(x,t)=\sum_{n=1}^{\infin}a_iX_i(t)T_i(t)\\ \frac{\partial u}{\partial t}=X\frac{dT}{dt},~~ \frac{\partial ^2u}{\partial X^2}=T\frac{d^2X}{dx},~~ \frac{\partial u}{\partial t}=\frac{\partial^2 u}{\partial x^2}\\ X\frac{dT}{dt}=T\frac{d^2X}{dx} \rightarrow \frac{1}{T}\frac{dT}{dt}=\frac{1}{x}\frac{d^2X}{dx^2}=C~(where~C~is~a~constant)\\ Case~1: positive~constant~\lambda^2 >0\\ \frac{1}{T}\frac{dT}{dt}=\frac{1}{x}\frac{d^2X}{dx^2}=\lambda^2\\ \frac{dT}{dt}=\lambda^2T \rightarrow T(t)=Ae^{\lambda^2t}\\ \frac{d^2X}{dx^2}=\lambda^2x \rightarrow X(x)=Be^{\lambda x}+Ce^{-\lambda x}\\ \therefore u = Ae^{\lambda^2t}(Be^{\lambda x}+Ce^{-\lambda x})\\ u(0,t)=e^{\lambda^2t}(B+C)=0 \rightarrow B+C=0\\ u(1,t)=e^{\lambda^2t}(Be^{\lambda}+Ce^{-\lambda})=0\\ if \lambda=0,it~implies~u~is~identically~0 \\ if \lambda \neq 0,(Be^{\lambda}+Ce^{-\lambda})=0which~doesn't ~exist.\\ Case~2: constant=0\\ \frac{1}{T}\frac{dT}{dt}=0\rightarrow T=A\\ \frac{1}{x}\frac{d^2X}{dx^2}=0\rightarrow X=Bx+C\\ u(0,t)=AC=0\\ u(1,t)=AB+AC=0\rightarrow AB=0\\ \therefore u(x,t)=0\\ Case~3: constant~is~negative\\ \frac{1}{T}\frac{dT}{dt}=-\lambda^2\rightarrow \frac{dT}{dt}=-\lambda^2T\\ \therefore T(t)=e^{-\lambda^2t}\\ \frac{1}{x}\frac{d^2X}{dx^2}=-\lambda^2\rightarrow \frac{d^2X}{dx^2}=-\lambda^2x\\ \therefore X(x)=Bcos(\lambda x)+Csin(\lambda x)\\ \therefore u(x,t)=e^{-\lambda^2t}(Bcos(\lambda x)+Csin(\lambda x))\\ Acoording~to~the~boundary~condition~\\ u(0,t)=e^{-\lambda^2t}B=0 \rightarrow B=0\\ u(1,t)=e^{-\lambda^2t} Ccos(\lambda x)=0\rightarrow C=0 \\ or ~\lambda= n\pi, n=0,1,2,3...\\ u(x,t)=\sum_{n=1}^{\infin}a_nsin(\lambda_n x)e^{-\lambda _n^2t}\\ Acoording~to~the~initial~condition~\\ u(x,0)=\sum_{n=1}^{\infin}a_nsin(\lambda_n x)=\phi(x)\\ According~ to ~ the~theorem\\ \phi(x)=\sum_{n=1}^{\infin}a_ne^0sin(\lambda_n x)=\sum_{n=1}^{\infin}a_nsin(\lambda_n x)\\ \int_{0}^{\infin}\phi (x)sin(\lambda_n x)dx\\ =\sum_{n=1}^{\infin}\int_{0}^{\infin}a_nsin(\lambda_n x)sin(\lambda_n x)dx\\ =\sum_{n=1}^{\infin}\int_{0}^{\infin}a_n (sin^2\lambda_nx)dx\\ =\frac{a_n }{2}\\ \therefore a_n=2\int_{0}^{1}\phi(x)sin(\lambda_nx)dx

Solution:
B y   s e p e r a t i o n   l a w ,   a s s u m e    u ( x , t ) = X ( t ) T ( t ) u n ( x , t ) = X n ( t ) T n ( t ) u = ∑ n = 1 ∞ u i ( x , t ) = ∑ n = 1 ∞ a i X i ( t ) T i ( t ) ∂ u ∂ t = X d T d t ,    ∂ 2 u ∂ X 2 = T d 2 X d x ,    ∂ u ∂ t = ∂ 2 u ∂ x 2 X d T d t = T d 2 X d x → 1 T d T d t = 1 x d 2 X d x 2 = C   ( w h e r e   C   i s   a   c o n s t a n t ) C a s e   1 : p o s i t i v e   c o n s t a n t   λ 2 > 0 1 T d T d t = 1 x d 2 X d x 2 = λ 2 d T d t = λ 2 T → T ( t ) = A e λ 2 t d 2 X d x 2 = λ 2 x → X ( x ) = B e λ x + C e − λ x ∴ u = A e λ 2 t ( B e λ x + C e − λ x ) u ( 0 , t ) = e λ 2 t ( B + C ) = 0 → B + C = 0 u x ( 1 , t ) = e λ 2 t λ ( B e λ + C e − λ ) = 0 i f λ = 0 , i t   i m p l i e s   u   i s   i d e n t i c a l l y   0 i f λ ≠ 0 , ( B e λ + C e − λ ) = 0 w h i c h   d o e s n ′ t   e x i s t . C a s e   2 : c o n s t a n t = 0 1 T d T d t = 0 → T = A 1 x d 2 X d x 2 = 0 → X = B x + C u ( 0 , t ) = A C = 0 u x ( 1 , t ) = A B + A C = 0 → A B = 0 ∴ u ( x , t ) = 0 C a s e   3 : c o n s t a n t   i s   n e g a t i v e 1 T d T d t = − λ 2 → d T d t = − λ 2 T ∴ T ( t ) = e − λ 2 t 1 x d 2 X d x 2 = − λ 2 → d 2 X d x 2 = − λ 2 x ∴ X ( x ) = B c o s ( λ x ) + C s i n ( λ x ) ∴ u ( x , t ) = e − λ 2 t ( B c o s ( λ x ) + C s i n ( λ x ) ) A c o o r d i n g   t o   t h e   b o u n d a r y   c o n d i t i o n   u ( 0 , t ) = e − λ 2 t B = 0 → B = 0 u x ( 1 , t ) = e − λ 2 t λ C c o s ( λ x ) = 0 → C = 0 o r   λ = π 2 ± n π , n = 0 , 1 , 2 , 3... u ( x , t ) = ∑ n = 1 ∞ a n s i n ( λ n x ) e − λ n 2 t A c o o r d i n g   t o   t h e   i n i t i a l   c o n d i t i o n   u ( x , 0 ) = ∑ n = 1 ∞ a n s i n ( λ n x ) = x A c c o r d i n g   t o   t h e   t h e o r e m ϕ ( x ) = ∑ n = 1 ∞ a n e 0 s i n ( λ n x ) = ∑ n = 1 ∞ a n s i n ( λ n x ) ∫ 0 ∞ ϕ ( x ) s i n ( λ n x ) d x = ∑ n = 1 ∞ ∫ 0 ∞ a n s i n ( λ n x ) s i n ( λ n x ) d x = ∑ n = 1 ∞ ∫ 0 ∞ a n ( s i n 2 λ n x ) d x = a n 2 ∴ a n = 2 ∫ 0 1 ϕ ( x ) s i n ( λ n x ) d x = 2 ∫ 0 1 x s i n ( λ n x ) d x = 2 ( s i n ( λ n ) λ n 2 − c o s ( λ n ) λ n ) = 2 ( s i n ( λ n ) λ n 2 ) By~seperation~law, ~assume ~~ u(x,t)=X(t)T(t)\\ u_n(x,t)=X_n(t)T_n(t)\\ u=\sum_{n=1}^{\infin}u_i(x,t)=\sum_{n=1}^{\infin}a_iX_i(t)T_i(t)\\ \frac{\partial u}{\partial t}=X\frac{dT}{dt},~~ \frac{\partial ^2u}{\partial X^2}=T\frac{d^2X}{dx},~~ \frac{\partial u}{\partial t}=\frac{\partial^2 u}{\partial x^2}\\ X\frac{dT}{dt}=T\frac{d^2X}{dx} \rightarrow \frac{1}{T}\frac{dT}{dt}=\frac{1}{x}\frac{d^2X}{dx^2}=C~(where~C~is~a~constant)\\ Case~1: positive~constant~\lambda^2 >0\\ \frac{1}{T}\frac{dT}{dt}=\frac{1}{x}\frac{d^2X}{dx^2}=\lambda^2\\ \frac{dT}{dt}=\lambda^2T \rightarrow T(t)=Ae^{\lambda^2t}\\ \frac{d^2X}{dx^2}=\lambda^2x \rightarrow X(x)=Be^{\lambda x}+Ce^{-\lambda x}\\ \therefore u = Ae^{\lambda^2t}(Be^{\lambda x}+Ce^{-\lambda x})\\ u(0,t)=e^{\lambda^2t}(B+C)=0 \rightarrow B+C=0\\ u_x(1,t)=e^{\lambda^2t}\lambda(Be^{\lambda}+Ce^{-\lambda})=0\\ if \lambda=0,it~implies~u~is~identically~0 \\ if \lambda \neq 0,(Be^{\lambda}+Ce^{-\lambda})=0which~doesn't ~exist.\\ Case~2: constant=0\\ \frac{1}{T}\frac{dT}{dt}=0\rightarrow T=A\\ \frac{1}{x}\frac{d^2X}{dx^2}=0\rightarrow X=Bx+C\\ u(0,t)=AC=0\\ u_x(1,t)=AB+AC=0\rightarrow AB=0\\ \therefore u(x,t)=0\\ Case~3: constant~is~negative\\ \frac{1}{T}\frac{dT}{dt}=-\lambda^2\rightarrow \frac{dT}{dt}=-\lambda^2T\\ \therefore T(t)=e^{-\lambda^2t}\\ \frac{1}{x}\frac{d^2X}{dx^2}=-\lambda^2\rightarrow \frac{d^2X}{dx^2}=-\lambda^2x\\ \therefore X(x)=Bcos(\lambda x)+Csin(\lambda x)\\ \therefore u(x,t)=e^{-\lambda^2t}(Bcos(\lambda x)+Csin(\lambda x))\\ Acoording~to~the~boundary~condition~\\ u(0,t)=e^{-\lambda^2t}B=0 \rightarrow B=0\\ u_x(1,t)=e^{-\lambda^2t}\lambda Ccos(\lambda x)=0\rightarrow C=0 \\ or ~\lambda=\frac{\pi}{2}\pm n\pi, n=0,1,2,3...\\ u(x,t)=\sum_{n=1}^{\infin}a_nsin(\lambda_n x)e^{-\lambda _n^2t}\\ Acoording~to~the~initial~condition~\\ u(x,0)=\sum_{n=1}^{\infin}a_nsin(\lambda_n x)=x\\ According~ to ~ the~theorem\\ \phi(x)=\sum_{n=1}^{\infin}a_ne^0sin(\lambda_n x)=\sum_{n=1}^{\infin}a_nsin(\lambda_n x)\\ \int_{0}^{\infin}\phi (x)sin(\lambda_n x)dx\\ =\sum_{n=1}^{\infin}\int_{0}^{\infin}a_nsin(\lambda_n x)sin(\lambda_n x)dx\\ =\sum_{n=1}^{\infin}\int_{0}^{\infin}a_n (sin^2\lambda_nx)dx\\ =\frac{a_n }{2}\\ \therefore a_n=2\int_{0}^{1}\phi(x)sin(\lambda_nx)dx=2\int_{0}^{1}xsin(\lambda_nx)dx\\ =2(\frac{sin(\lambda _n)}{\lambda_n^2}-\frac{cos(\lambda _n)}{\lambda_n}) =2(\frac{sin(\lambda _n)}{\lambda_n^2})

F i n a l l y ,   s o l u t i o n   t o   t h e   P D E ： u ( x , t ) = ∑ n = 1 ∞ a n s i n ( λ n x ) e − λ n 2 t w h e r e   a n = 2 s i n   λ n λ n 2 λ n = π 2 ± n π , n = 0 , 1 , 2 , 3... T h e   s o l u t i o n   a g r e e s   w i t h   m y   i n t u i t i o n ， a n d   t h e   s t e a d y   s t a t e   i s   0 Finally, ~solution~to~the~PDE：u(x,t)=\sum_{n=1}^{\infin}a_nsin(\lambda_nx)e^{-\lambda _n^2t}\\ where~a_n=\frac{2sin~ \lambda _n}{\lambda_n^2}\\ \lambda_n=\frac{\pi}{2}\pm n\pi,n=0,1,2,3...\\ The~solution~agrees ~with ~my ~intuition， and~the ~steady~state~is~0

## PDE with non-homogeneous B.Cs

∂ T ‾ ∂ t ‾ = ∂ 2 T ‾ ∂ x ‾ 2    B . C u ( 0 , t ) = u 0 , u ( 1 , t ) = u 1     I . C u ( x , 0 ) = ϕ ( x ) P D E   w i t h   n o n h o m o g e n e o u s   B . C s u ( x , t ) = u s s ( t ) + u t r ( t ) A t   s t r e a d y   s t a t e : ∂ u ∂ t = 0 , ∂ 2 u s s ∂ x 2 = 0 ⇒ d 2 u s s d x 2 = 0 u s s = C 1 x + C 2   a n d   a c c o r d i n g   t o   t h e   i n i t i a l   c o n d i t i o n u s s ∣ x = 0 = u 0 u s s ∣ x = L = u L C 2 = u 0 , C 1 = u l − u 0 L ∴ u s s = u 0 + u l − u 0 L ∗ x ∂ ( u s s ( x ) + u t r ( x , t ) ) ∂ t = ∂ 2 u s s ∂ x 2 + ∂ 2 u t r ∂ x 2 ∂ u s s ∂ t = ∂ 2 u s s ∂ x 2 = 0 ∂ ( u t r ( x , t ) ) ∂ t = ∂ 2 u t r ∂ x 2 S o   w e   j u s t   n e e d   t o   s o l v e   t r a n s i e n t   s u b   p r o b l e m B . C 1 : u s s ∣ x = 0 = 0 ⇒ u t r ∣ x = 0 = 0 B . C 2 : u s s ∣ x = 1 = 1 ⇒ u t r ∣ x = 1 = 0 I . C : u s s ∣ t = 0 + u t r ∣ t = 0 = ϕ ( x ) ⇒ u t r ∣ t = 0 = ϕ ( x ) − [ u 0 + u l − u 0 L ∗ x ] u t r ( x , t ) = ∑ n = 1 ∞ a n s i n ( λ n x ) e − λ n 2 t w h e r e   a n = 2 L ∫ 0 L [ ϕ ( x ) − [ u 0 + u l − u 0 L ∗ x ] ] s i n ( λ n x ) d x λ n = n π L ∴ u ( x , t ) = ∑ n = 1 ∞ a n s i n ( λ n x ) e − λ n 2 t + 1 L w h e r e   a n = 2 L ∫ 0 L [ x 2 − 1 L ] s i n ( λ n x ) d x λ n = n π L T h e   s o l u t i o n   a g r e e s   w i t h   m y   i n t u i t i o n ， a n d   t h e   s t e a d y   s t a t e   i s   1 L , t h e   t r a n s i e n t   s t a t e   i s = ∑ n = 1 ∞ a n s i n ( λ n x ) e − λ n 2 t w h e r e   a n = 2 L ∫ 0 L [ x 2 − 1 L ] s i n ( λ n x ) d x λ n = n π L \frac{\partial \overline{T}}{\partial\overline{t}}=\frac{\partial^2 \overline{T}}{\partial\overline{x}^2}~~ B.C u(0,t)=u_0,u(1,t)=u_1~~~ I.C u(x,0)=\phi(x)\\ PDE~with~nonhomogeneous~B.Cs\\ u(x,t)=u_{ss}(t)+u_{tr}(t)\\ At~stready~state:\\ \frac{\partial u}{\partial t}=0, \frac{\partial^2 u_{ss}}{\partial x^2}=0\\ \Rightarrow \frac{d^2 u_{ss}}{d x^2}=0\\ u_{ss}=C_1x+C_2~and ~according~to~the ~initial ~condition\\ u_{ss}|_{x=0}=u_0\\ u_{ss}|_{x=L}=u_L\\ C_2=u_0,C_1=\frac{u_l-u_0}{L}\\ \therefore u_{ss}=u_0+\frac{u_l-u_0}{L}*x\\ \frac{\partial (u_{ss}(x)+u_{tr}(x,t))}{\partial t}= \frac{\partial^2u_{ss}}{\partial x^2}+\frac{\partial^2u_{tr}}{\partial x^2}\\ \frac{\partial u_{ss}}{\partial t}=\frac{\partial^2u_{ss}}{\partial x^2}=0\\ \frac{\partial (u_{tr}(x,t))}{\partial t}=\frac{\partial^2u_{tr}}{\partial x^2}\\ So~we~just~need~to~solve~transient~sub~problem\\ B.C1:u_{ss}|_{x=0}=0\Rightarrow u_{tr}|_{x=0}=0\\ B.C2:u_{ss}|_{x=1}=1\Rightarrow u_{tr}|_{x=1}=0\\ I.C:u_{ss}|_{t=0}+u_{tr}|_{t=0}=\phi(x)\\ \Rightarrow u_{tr}|_{t=0}=\phi(x)-[u_0+\frac{u_l-u_0}{L}*x]\\ u_{tr}(x,t)=\sum_{n=1}^{\infin}a_nsin(\lambda_nx)e^{-\lambda _n^2t}\\ where~a_n=\frac{2}{L}\int_{0}^{L}[\phi(x)-[u_0+\frac{u_l-u_0}{L}*x]]sin(\lambda_nx)dx\\ \lambda_n=n\frac{\pi}{L}\\ \therefore u(x,t)=\sum_{n=1}^{\infin}a_nsin(\lambda_nx)e^{-\lambda _n^2t}+\frac{1}{L}\\ where~a_n=\frac{2}{L}\int_{0}^{L}[x^2-\frac{1}{L}]sin(\lambda_nx)dx\\ \lambda_n=n\frac{\pi}{L}\\ The~solution~agrees ~with ~my ~intuition， and~the ~steady~state~is~\frac{1}{L},\\ the ~transient ~state~is=\sum_{n=1}^{\infin}a_nsin(\lambda_nx)e^{-\lambda _n^2t}\\ where~a_n=\frac{2}{L}\int_{0}^{L}[x^2-\frac{1}{L}]sin(\lambda_nx)dx\\ \lambda_n=n\frac{\pi}{L}\\

Solution:
P D E   w i t h   n o n h o m o g e n e o u s   B . C s u ( x , t ) = u s s ( t ) + u t r ( t ) A t   s t r e a d y   s t a t e : ∂ u ∂ t = 0 , ∂ 2 u s s ∂ x 2 = 0 ⇒ d 2 u s s d x 2 = 0 u s s = C 1 x + C 2   a n d   a c c o r d i n g   t o   t h e   i n i t i a l   c o n d i t i o n u s s ∣ x = 0 = 0 u s s ∣ x = 1 = 1 C 2 = 0 , C 1 = 1 ∴ u s s = x ∂ ( u s s ( x ) + u t r ( x , t ) ) ∂ t = ∂ 2 u s s ∂ x 2 + ∂ 2 u t r ∂ x 2 ∂ u s s ∂ t = ∂ 2 u s s ∂ x 2 = 0 ∂ ( u t r ( x , t ) ) ∂ t = ∂ 2 u t r ∂ x 2 S o   w e   j u s t   n e e d   t o   s o l v e   t r a n s i e n t   s u b   p r o b l e m B . C 1 : u s s ∣ x = 0 = 0 ⇒ u t r ∣ x = 0 = 0 B . C 2 : u s s ∣ x = 1 = 1 ⇒ u t r ∣ x = 1 = 0 I . C : u s s ∣ t = 0 + u t r ∣ t = 0 = x 2 ⇒ u t r ∣ t = 0 = x 2 − x f o r   u t r ,   a p p l y   w h a t   w e   k n o w   i n   Q 1 C a s e   1 : p o s i t i v e   c o n s t a n t   λ 2 > 0 u t r   c o u l d   n o t   f i n d   a   s o l u t i o n C a s e   2 :   c o n s t a n t   λ 2 = 0 u t r ( x , t )   i s   i d e n t i c a l l y   0 C a s e   3 : u t r ( x , t ) = ∑ n = 1 ∞ a n s i n ( λ n x ) e − λ n 2 t w h e r e   a n = 2 ∫ 0 1 [ x 2 − x ] s i n ( λ n x ) d x λ n = ± n π , n = 0 , 1 , 2 , 3... ∴ u ( x , t ) = ∑ n = 1 ∞ a n s i n ( λ n x ) e − λ n 2 t + x w h e r e   a n = 2 ∫ 0 1 [ x 2 − 1 ] s i n ( λ n x ) d x = 4 λ n 2 ( s i n λ n + c o s λ n − 1 λ n ) = 4 λ n 3 ( c o s λ n − 1 ) λ n = ± n π , n = 0 , 1 , 2 , 3... T h e   s o l u t i o n   a g r e e s   w i t h   m y   i n t u i t i o n ， a n d   t h e   s t e a d y   s t a t e   i s   x , t h e   t r a n s i e n t   s t a t e   i s = ∑ n = 1 ∞ a n s i n ( λ n x ) e − λ n 2 t w h e r e   a n = 4 λ n 3 ( c o s λ n − 1 ) λ n = ± n π , n = 0 , 1 , 2 , 3... PDE~with~nonhomogeneous~B.Cs\\ u(x,t)=u_{ss}(t)+u_{tr}(t)\\ At~stready~state:\\ \frac{\partial u}{\partial t}=0, \frac{\partial^2 u_{ss}}{\partial x^2}=0\\ \Rightarrow \frac{d^2 u_{ss}}{d x^2}=0\\ u_{ss}=C_1x+C_2~and ~according~to~the ~initial ~condition\\ u_{ss}|_{x=0}=0\\ u_{ss}|_{x=1}=1\\ C_2=0,C_1=1\\ \therefore u_{ss}=x\\ \frac{\partial (u_{ss}(x)+u_{tr}(x,t))}{\partial t}= \frac{\partial^2u_{ss}}{\partial x^2}+\frac{\partial^2u_{tr}}{\partial x^2}\\ \frac{\partial u_{ss}}{\partial t}=\frac{\partial^2u_{ss}}{\partial x^2}=0\\ \frac{\partial (u_{tr}(x,t))}{\partial t}=\frac{\partial^2u_{tr}}{\partial x^2}\\ So~we~just~need~to~solve~transient~sub~problem\\ B.C1:u_{ss}|_{x=0}=0\Rightarrow u_{tr}|_{x=0}=0\\ B.C2:u_{ss}|_{x=1}=1\Rightarrow u_{tr}|_{x=1}=0\\ I.C:u_{ss}|_{t=0}+u_{tr}|_{t=0}=x^2 \Rightarrow u_{tr}|_{t=0}=x^2-x \\ for ~u_{tr},~apply ~what~we~know~in~Q1\\ Case~1: positive~constant~\lambda^2 >0\\ u_{tr}~could~not ~find~a~solution\\ Case~2: ~constant~\lambda^2 =0\\ u_{tr}(x,t)~is~identically~0\\ Case~3:\\ u_{tr}(x,t)=\sum_{n=1}^{\infin}a_nsin(\lambda_nx)e^{-\lambda _n^2t}\\ where~a_n=2\int_{0}^{1}[x^2-x]sin(\lambda_nx)dx\\ \lambda_n=\pm n\pi,n=0,1,2,3...\\ \therefore u(x,t)=\sum_{n=1}^{\infin}a_nsin(\lambda_nx)e^{-\lambda _n^2t}+x\\ where~a_n=2\int_{0}^{1}[x^2-1]sin(\lambda_nx)dx\\ =\frac{4}{\lambda_n^2}(sin\lambda_n+\frac{cos \lambda _n-1}{\lambda_n})=\frac{4}{\lambda_n^3}(cos \lambda _n-1)\\ \lambda_n=\pm n\pi,n=0,1,2,3...\\ The~solution~agrees ~with ~my ~intuition， and~the ~steady~state~is~x,\\ the ~transient ~state~is=\sum_{n=1}^{\infin}a_nsin(\lambda_nx)e^{-\lambda _n^2t}\\ where~a_n=\frac{4}{\lambda_n^3}(cos \lambda _n-1)\\ \lambda_n=\pm n\pi,n=0,1,2,3...\\

## heat-flow problem with diffusion

T h e   f i r s t   t y p e ∂ T ‾ ∂ t ‾ = α 2 ∂ 2 T ‾ ∂ x ‾ 2 − β ( u − u 0 ) u t = α 2 u x x − β u       B . C u ( 0 , t ) = 0 , u ( 1 , t ) = 0     I . C u ( x , 0 ) = ϕ ( x ) l e t   u = e − β t ∗ w   ∴ w t = α 2 w x x T h e   s e c o n d   t y p e ∂ T ‾ ∂ t ‾ = α 2 ∂ 2 T ‾ ∂ x ‾ 2 − β u x u t = α 2 u x x − β u x       B . C   u ( 0 , t ) = 0 , u ( 1 , t ) = 0     I . C u ( x , 0 ) = ϕ ( x )   l e t u = e β ( x − β t x ) 2 α 2 ∗ w   ∴ w t = α 2 w x x The ~first~type\\ \frac{\partial \overline{T}}{\partial\overline{t}}=\alpha^2\frac{\partial^2 \overline{T}}{\partial\overline{x}^2}-\beta(u-u_0) \\ u_t=\alpha^2u_{xx}-\beta u~~~~~B.C u(0,t)=0,u(1,t)=0~~~ I.C u(x,0)=\phi(x)\\ let~u=e^{-\beta t }*w~\therefore w_t=\alpha^2w_{xx}\\ The~second~type\\ \frac{\partial \overline{T}}{\partial\overline{t}}=\alpha^2\frac{\partial^2 \overline{T}}{\partial\overline{x}^2}-\beta u_x\\ u_t=\alpha^2u_{xx}-\beta u_x~~~~~B.C ~u(0,t)=0,u(1,t)=0~~~ I.C u(x,0)=\phi(x)\\ ~let u=e^{\frac{\beta (x-\frac{\beta t}{x})}{2\alpha^2}}*w~\therefore w_t=\alpha^2w_{xx}\\

u t = u x x − u        B . C u ( 0 , t ) = 0 , u ( 1 , t ) = 0 I . C u ( x , 0 ) = s i n π x + 0.5 s i n ( 3 π x ) l e t   u = e − β t ∗ w   w ( x , t ) = ∑ n = 1 ∞ a n e − ( n π α ) 2 t s i n ( n π x ) a n = 2 ∫ 0 1 ( s i n π x + 0.5 s i n ( 3 π x ) ) s i n ( n π x ) d x n = 1 , a 1 = 2 ∫ 0 1 ( s i n π x ) 2 = 1 n = 3 , a 3 = 2 ∫ 0 1 ( s i n 3 π x ) 2 = 0.5 w ( x , t ) = e − π 2 t s i n π x + 0.5 e − 3 π 2 t s i n 3 π x u_t=u_{xx}- u~~~~\\ ~B.C u(0,t)=0,u(1,t)=0\\ I.C u(x,0)=sin\pi x+0.5 sin(3\pi x)\\ let~u=e^{-\beta t }*w~\\ w(x,t)=\sum_{n=1}^\infin a_n e^{-(n\pi \alpha)^2t}sin(n\pi x)\\ a_n=2\int_0^1 (sin\pi x+0.5 sin(3\pi x))sin(n\pi x)dx\\ n=1,a_1=2\int_0^1 (sin\pi x)^2=1\\ n=3, a_3=2\int_0^1 (sin3\pi x)^2=0.5\\ w(x,t)=e^{-\pi^2t}sin\pi x +0.5e^{-3\pi^2t}sin3\pi x

## non-homogeneous PDE

u t = α 2 u x x + f ( x , t ) B . C   α 1 u x ( 0 , t ) + β 1 u ( 0 , t ) = 0 α 2 u x ( 1 , t ) + β 2 u ( 1 , t ) = 0 I . C   u ( x , 0 ) = ϕ ( x ) f ( x , t ) ≠ 0 , h a s   a   h e a t   d a m p i n g   e f f e c t S T E P   1 f ( x , t ) = ∑ f n ( t ) X n ( x ) a n d   f i n d   t h e   r e s p o n s e   u n ( x , t ) t o   e a c h   o f   t h e s e   i n d i v i d u a l   c o m p o n e n t s u ( x , t ) = ∑ n = 1 ∞ u n ( x , t ) S T E P   2 m u l t i p l y   e a c h   s i d e   o f   t h e   e q u a t i o n   b y   s i n ( m π x )   a n d   i n t e g r a t e   f r o m   0   t o   1   w i t h   r e s p e c t   t o   X ∫ 0 1 f ( x , t ) s i n ( m π x ) d x = ∑ n = 1 ∞ f n ( t ) ∫ 0 1 s i n ( m π x ) s i n ( n π x ) d x = 1 2 f m ( t ) c h a n g e   f r o m   m   t o   n → f n ( t ) = 2 ∫ 0 1 f n ( t ) s i n ( n π x ) S T E P   3 f ( x , t ) = ∑ f n ( t ) X n ( x ) a n d   f i n d   t h e   r e s p o n s e   u n ( x , t ) t o   e a c h   o f   t h e s e   i n d i v i d u a l   c o m p o n e n t s s u b s t i t u i n g   u ( x , t ) = ∑ n = 1 ∞ u n ( x , t ) i n t o   t h e   s y s t e m u t = α 2 u x x + ∑ f n ( t ) X n ( x ) B . C   α 1 u x ( 0 , t ) + β 1 u ( 0 , t ) = 0 α 2 u x ( 1 , t ) + β 2 u ( 1 , t ) = 0 I . C   u ( x , 0 ) = ϕ ( x ) ∑ n = 1 ∞ T n ′ ( t ) s i n n π x = − α 2 ∑ n = 1 ∞ T n ( t ) ( n π ) 2 s i n n π x + ∑ n = 1 ∞ f n ( t ) s i n n π x ∑ n = 1 ∞ [ T n ′ ( t ) + ) ( α n π ) 2 s i n n π x − f n ( t ) ] s i n n π x = 0 I . C   ∑ n = 1 ∞ T n ( 0 ) s i n n π x = ϕ ( x ) B . C   ∑ n = 1 ∞ T n ′ ( t ) s i n 0 = 0    ∑ n = 1 ∞ T n ′ ( t ) s i n n π = 0 ( s a t i s f y ) T n ′ ( t ) + ) ( α n π ) 2 s i n n π x − f n ( t ) = 0 a n = 2 ∫ 0 1 ϕ ( x ) s i n n π x d x T n ( t ) = a n e − ( n π α ) 2 + ∫ 0 1 e − ( n π α ) 2 ( t − τ ) f n ( τ ) d τ H e n c e , t h e   s o l u t i o n   o f   t h e   p r o b l e m u ( x , t ) = ∑ n = 1 ∞ T n ( t ) s i n n π x = ∑ n = 1 ∞ [ a n e − ( n π α ) 2 s i n n π x + ∫ 0 1 e − ( n π α ) 2 ( t − τ ) f n ( τ ) d τ ( s i n n π ) ] t r a n s i e n t   p a r t   f r o m   i n i t i a l   c o n d i t i o n s t e a d y   s t a t e   f r o m   h e a t   s o u r c e u_t=\alpha^2 u_{xx}+f(x,t)\\ B.C~\alpha_1u_x(0,t)+\beta_1u(0,t)=0\\ \alpha_2u_x(1,t)+\beta_2u(1,t)=0\\ I.C~u(x,0)=\phi(x)\\ f(x,t) \neq 0,has~a~heat~damping~effect\\ STEP~1\\ f(x,t) =\sum f_n(t)X_n(x)and~find~the~response~u_n(x,t)\\ to~each~of~these~individual~components\\ u(x,t) =\sum_{n=1}^\infin u_n(x,t)\\ STEP~2\\ multiply~each~side~of ~the~equation~by~sin(m\pi x)~and ~integrate~\\ from~0~to~1~with~respect~to~X\\ \int_0^1f(x,t)sin(m\pi x)dx=\sum_{n=1}^\infin f_n(t) \int_0^1 sin(m\pi x)sin(n\pi x)dx\\ =\frac{1}{2}f_m(t)change~from~m~to~n\\ \rightarrow f_n(t)=2\int_0^1f_n(t)sin(n\pi x)\\ STEP~3\\ f(x,t) =\sum f_n(t)X_n(x)and~find~the~response~u_n(x,t)\\ to~each~of~these~individual~components\\ substituing~u(x,t) =\sum_{n=1}^\infin u_n(x,t)into~the~system\\ u_t=\alpha^2 u_{xx}+\sum f_n(t)X_n(x)\\ B.C~\alpha_1u_x(0,t)+\beta_1u(0,t)=0\\ \alpha_2u_x(1,t)+\beta_2u(1,t)=0\\ I.C~u(x,0)=\phi(x)\\ \sum_{n=1}^\infin T_n^{'}(t)sinn\pi x=-\alpha^2\sum_{n=1}^\infin T_n(t)(n\pi)^2sinn\pi x +\sum_{n=1}^\infin f_n(t)sinn\pi x\\ \sum_{n=1}^\infin[T_n^{'}(t)+)(\alpha n\pi)^2sinn\pi x-f_n(t)]sinn\pi x=0\\ I.C~\sum_{n=1}^\infin T_n(0)sin n \pi x=\phi(x)\\ B.C~\sum_{n=1}^\infin T_n^{'}(t)sin0=0~~\sum_{n=1}^\infin T_n^{'}(t)sinn\pi=0(satisfy)\\ T_n^{'}(t)+)(\alpha n\pi)^2sinn\pi x-f_n(t)=0\\ a_n=2\int_0^1\phi(x)sin n \pi x dx\\ T_n(t)=a_ne^{-(n\pi \alpha)^2}+\int_0^1e^{-(n\pi \alpha)^2(t-\tau )}f_n(\tau)d\tau\\ Hence, the~solution~of~the~problem\\ u(x,t)=\sum_{n=1}^\infin T_n(t)sin n\pi x\\ =\sum_{n=1}^\infin [a_ne^{-(n\pi \alpha)^2}sin n\pi x+\int_0^1e^{-(n\pi \alpha)^2(t-\tau )}f_n(\tau)d\tau(sin n\pi)]\\ transient ~part~from~initial~condition\\ steady~state~from~heat~source\\

u t = α 2 u x x + ∑ f n ( t ) X n ( x ) B . C   α 1 u x ( 0 , t ) + β 1 u ( 0 , t ) = 0 α 2 u x ( 1 , t ) + β 2 u ( 1 , t ) = 0 I . C   u ( x , 0 ) = ϕ ( x ) ∑ n = 1 ∞ T n ′ ( t ) s i n n π x = − α 2 ∑ n = 1 ∞ T n ( t ) ( n π ) 2 s i n n π x + ∑ n = 1 ∞ f n ( t ) s i n n π x = 1 ∑ n = 1 ∞ [ T n ′ ( t ) + ) ( α n π ) 2 s i n n π x − f n ( t ) ] s i n n π x = 0 I . C   ∑ n = 1 ∞ T n ( 0 ) s i n n π x = s i n π x T n ( 0 ) = 2 ∫ 0 1 s i n π x s i n π x d x = 1   f o r   n = 1    ; 0    f o r   n ≠ 1 B . C   ∑ n = 1 ∞ T n ′ ( t ) s i n 0 = 0    ∑ n = 1 ∞ T n ′ ( t ) s i n n π = 0 ( s a t i s f y ) T n ′ ( t ) + ( α n π ) 2 s i n n π x = f n ( t ) = 2 ∫ 0 1 s i n 3 π x s i n n π x d x = 1   f o r   n = 3    ; 0    f o r   n ≠ 3 n = 1 T 1 ′ + ( π α ) 2 T 1 = 0 , T 1 ( 0 ) = 1 → T 1 ( t ) = e − ( π α ) 2 t n = 2 T 2 ′ + ( π α ) 2 T 2 = 0 , T 2 ( 0 ) = 0 → T 2 ( t ) = 0 n = 3 T 3 ′ + ( π α ) 2 T 3 = 1 , T 3 ( 0 ) = 0 → T 3 ( t ) = 1 ( 3 π α ) 2 [ 1 − e − ( 3 π α ) 2 t ] n > 3 T n ′ + ( π α ) 2 T n = 0 , T n ( 0 ) = 0 → T n ( t ) = 0 u ( x , t ) = ∑ n = 1 ∞ T n ( t ) s i n n π x = ∑ n = 1 ∞ [ a n e − ( n π α ) 2 s i n n π x + ∫ 0 1 e − ( n π α ) 2 ( t − τ ) f n ( τ ) d τ ( s i n n π ) ] = e − ( π α ) 2 t s i n π x + 1 ( 3 π α ) 2 [ 1 − e − ( 3 π α ) 2 t ] s i n 3 π x f o r   α = 1 u ( x , t ) = e − π 2 t s i n π x + 1 ( 3 π ) 2 [ 1 − e − ( 3 π ) 2 t ] s i n 3 π x T h i s   s o l u t i o n   a g r e e   w i t h   m y   i n t u i t i o n ,   a n d   t h e   f i r s t   p a r t   i s   d u e   t o   t h e   I C   a n d   t h e   s e c o n d   p a r t   t h a t   i s   d u e   t o   t h e   h e a t   s o u r c e   f ( x , t ) . T h e   p h r a s e   s t e a d y   s t a t e   i s   n o t   t h e   b e s t   p h r a s e   t o   d e s c r i e b   t h e   s e c o n d   p a r t , s i n c e   i t   d o e s n ′ t   n e c e s s a r i l y   c o m e   t o   r e s t ( i t   m a y   a p p r o a c h   a   p e r i o d i c   s t e a d y   s t a t e , i f   f ( x , t )   i s   p e r i o d i c   i n   t ) . u_t=\alpha^2 u_{xx}+\sum f_n(t)X_n(x)\\ B.C~\alpha_1u_x(0,t)+\beta_1u(0,t)=0\\ \alpha_2u_x(1,t)+\beta_2u(1,t)=0\\ I.C~u(x,0)=\phi(x)\\ \sum_{n=1}^\infin T_n^{'}(t)sinn\pi x=-\alpha^2\sum_{n=1}^\infin T_n(t)(n\pi)^2sinn\pi x +\sum_{n=1}^\infin f_n(t)sinn\pi x =1\\ \sum_{n=1}^\infin[T_n^{'}(t)+)(\alpha n\pi)^2sinn\pi x-f_n(t)]sinn\pi x=0\\ I.C~ \sum_{n=1}^\infin T_n(0)sin n \pi x=sin\pi x\\ T_n(0)=2\int_0^1sin\pi x sin \pi x dx=1~for~n=1 ~~; 0~~for ~n \neq 1\\ B.C~ \sum_{n=1}^\infin T_n^{'}(t)sin0=0~~\sum_{n=1}^\infin T_n^{'}(t)sinn\pi=0(satisfy)\\ T_n^{'}(t)+(\alpha n\pi)^2sinn\pi x=f_n(t)=2\int_0^1sin3\pi xsinn\pi xdx\\ =1~for~n=3 ~~; 0~~for ~n \neq 3\\ n=1\\ T_1^{'}+(\pi\alpha)^2T_1=0,T_1(0)=1\\ \rightarrow T_1(t)=e^{-(\pi\alpha)^2t}\\ n=2\\ T_2^{'}+(\pi\alpha)^2T_2=0,T_2(0)=0\\ \rightarrow T_2(t)=0\\ n=3\\ T_3^{'}+(\pi\alpha)^2T_3=1,T_3(0)=0\\ \rightarrow T_3(t)=\frac{1}{(3\pi \alpha)^2}[1-e^{-(3\pi\alpha)^2t}]\\ n>3\\ T_n^{'}+(\pi\alpha)^2T_n=0,T_n(0)=0\\ \rightarrow T_n(t)=0\\ u(x,t)=\sum_{n=1}^\infin T_n(t)sin n\pi x\\ =\sum_{n=1}^\infin [a_ne^{-(n\pi \alpha)^2}sin n\pi x+\int_0^1e^{-(n\pi \alpha)^2(t-\tau )}f_n(\tau)d\tau(sin n\pi)]\\ =e^{-(\pi\alpha)^2t}sin\pi x+\frac{1}{(3\pi \alpha)^2}[1-e^{-(3\pi\alpha)^2t}]sin3\pi x\\ for~\alpha=1\\ u(x,t)=e^{-\pi^2t}sin\pi x+\frac{1}{(3\pi )^2}[1-e^{-(3\pi)^2t}]sin3\pi x\\ This ~solution~ agree~ with~ my ~intuition, ~and ~the ~first ~part ~is~due~to~the~IC~\\and~the~second~part~that~is~due~to~the~heat~source~f(x,t). \\ The~phrase~steady~state~is~not~the~best~phrase~to~descrieb~the~second~part,\\ since~it~doesn't~necessarily~come~to~rest(it~may~approach~a~periodic~steady~state, \\ if~f(x,t)~is~periodic~in~t).