112 年 國立成功大學電機工程學系碩士班丙組《工程數學》

📄 試題原卷 免費註冊後即可對照原始考卷 PDF免費註冊

第 1 題20 分

  1. (20%)
    (a)(10%) Find the fundamental sets y1(x)y_1(x) and y2(x)y_2(x) of the general solutions to y′′−7xy′+16y=0y'' - 7xy' + 16y = 0 for x∈(−8,8)x \in (-8,8).
    (b)(10%) Show that y1(x)y_1(x) and y2(x)y_2(x) are linearly independent with W(x)=y1(x)y2′(x)−y2(x)y1′(x)=0W(x) = y_1(x)y_2'(x) - y_2(x)y_1'(x) = 0 at x=0x = 0 and explain why it is not contradict to the Wronskian test.

登入後即可作答並保存紀錄。

這一題的完整詳解

核心觀念

  1. 二階變係數齊次線性微分方程式之級數解法(Power Series Method):
    對於二階線性微分方程式 y′′+P(x)y′+Q(x)y=0y'' + P(x) y' + Q(x) y = 0,若 x0=0x_0 = 0 處的係數函數 P(x)=−7xP(x) = -7x 與 Q(x)=16Q(x) = 16 皆為解析函數(Analytic Functions),則 x0=0x_0 = 0 為常點(Ordinary Point)。可以在 x=0x = 0 處設冪級數解:
    y(x)=∑n=0∞anxny(x) = \sum_{n=0}^{\infty} a_n x^n
    帶入原式可導出係數的遞迴關係式(Recurrence Relation)。

  2. 基本解集合(Fundamental Set of Solutions):
    二階線性齊次微分方程式的通解由兩個線性獨立的基底解 y1(x)y_1(x) 與 y2(x)y_2(x) 所張成:
    y(x)=c1y1(x)+c2y2(x)y(x) = c_1 y_1(x) + c_2 y_2(x)
    在 x=0x = 0 處,常取規範化初值條件(Normalized Initial Conditions):

    • y1(0)=1, y1′(0)=0  ⟹  (a0,a1)=(1,0)y_1(0) = 1, \, y_1'(0) = 0 \implies (a_0, a_1) = (1, 0) (偶函數解)
    • y2(0)=0, y2′(0)=1  ⟹  (a0,a1)=(0,1)y_2(0) = 0, \, y_2'(0) = 1 \implies (a_0, a_1) = (0, 1) (奇函數解)
  3. 朗斯基行列式(Wronskian)與阿貝爾公式(Abel's Identity):
    兩函數 y1(x),y2(x)y_1(x), y_2(x) 的朗斯基行列式定義為:
    W(y1,y2)(x)=y1(x)y2′(x)−y2(x)y1′(x)W(y_1, y_2)(x) = y_1(x) y_2'(x) - y_2(x) y_1'(x)
    對二階 ODE y′′+P(x)y′+Q(x)y=0y'' + P(x)y' + Q(x)y = 0,由阿貝爾公式:
    W(x)=W(x0)exp⁡(−∫x0xP(t) dt)W(x) = W(x_0) \exp\left( -\int_{x_0}^x P(t) \, dt \right)

  4. 朗斯基測試定理(Wronskian Test / Theorem):
    若 P(x)P(x) 與 Q(x)Q(x) 在區間 II 上均為連續函數,則微分方程式的任意兩解 y1(x),y2(x)y_1(x), y_2(x) 在區間 II 上線性獨立的充要條件為:朗斯基行列式 W(x)W(x) 在區間 II 內處處不為零(W(x)≠0, ∀x∈IW(x) \neq 0, \, \forall x \in I)。


解題方法

(a) 求基本解集合 y1(x)y_1(x) 與 y2(x)y_2(x)

給定微分方程式:
y′′−7xy′+16y=0,x∈(−8,8)y'' - 7x y' + 16 y = 0, \quad x \in (-8,8)

步驟一:代入冪級數及其導函數
設 y(x)=∑n=0∞anxny(x) = \sum_{n=0}^{\infty} a_n x^n,對其逐項求導:
y′(x)=∑n=1∞nanxn−1y'(x) = \sum_{n=1}^{\infty} n a_n x^{n-1}
y′′(x)=∑n=2∞n(n−1)anxn−2=∑n=0∞(n+2)(n+1)an+2xny''(x) = \sum_{n=2}^{\infty} n(n-1) a_n x^{n-2} = \sum_{n=0}^{\infty} (n+2)(n+1) a_{n+2} x^n

代入原微分方程式:
∑n=0∞(n+2)(n+1)an+2xn−7x∑n=1∞nanxn−1+16∑n=0∞anxn=0\sum_{n=0}^{\infty} (n+2)(n+1) a_{n+2} x^n - 7x \sum_{n=1}^{\infty} n a_n x^{n-1} + 16 \sum_{n=0}^{\infty} a_n x^n = 0
∑n=0∞[(n+2)(n+1)an+2−7nan+16an]xn=0\sum_{n=0}^{\infty} \left[ (n+2)(n+1) a_{n+2} - 7n a_n + 16 a_n \right] x^n = 0

步驟二:求遞迴關係式
令 xnx^n 之係數為零,得到遞迴關係式:
(n+2)(n+1)an+2+(16−7n)an=0(n+2)(n+1) a_{n+2} + (16 - 7n) a_n = 0
  ⟹  an+2=7n−16(n+2)(n+1)an,n≥0\implies a_{n+2} = \frac{7n - 16}{(n+2)(n+1)} a_n, \quad n \ge 0

步驟三:求基本解 y1(x)y_1(x) 與 y2(x)y_2(x)

  1. 解 y1(x)y_1(x)(取 a0=1,a1=0a_0 = 1, a_1 = 0,僅含偶次項):

    • n=0n = 0: a2=7(0)−162⋅1a0=−8a_2 = \frac{7(0)-16}{2 \cdot 1} a_0 = -8
    • n=2n = 2: a4=7(2)−164⋅3a2=−212(−8)=43a_4 = \frac{7(2)-16}{4 \cdot 3} a_2 = \frac{-2}{12} (-8) = \frac{4}{3}
    • n=4n = 4: a6=7(4)−166⋅5a4=1230(43)=815a_6 = \frac{7(4)-16}{6 \cdot 5} a_4 = \frac{12}{30} \left(\frac{4}{3}\right) = \frac{8}{15}
    • 一般通項(n=2kn = 2k):
      a2k=(∏j=0k−114j−16(2j+2)(2j+1))a0a_{2k} = \left( \prod_{j=0}^{k-1} \frac{14j - 16}{(2j+2)(2j+1)} \right) a_0
      故偶函數基底解為:
      y1(x)=1−8x2+43x4+815x6+⋯=1+∑k=1∞a2kx2ky_1(x) = 1 - 8x^2 + \frac{4}{3}x^4 + \frac{8}{15}x^6 + \cdots = 1 + \sum_{k=1}^{\infty} a_{2k} x^{2k}
  2. 解 y2(x)y_2(x)(取 a0=0,a1=1a_0 = 0, a_1 = 1,僅含奇次項):

    • n=1n = 1: a3=7(1)−163⋅2a1=−96=−32a_3 = \frac{7(1)-16}{3 \cdot 2} a_1 = \frac{-9}{6} = -\frac{3}{2}
    • n=3n = 3: a5=7(3)−165⋅4a3=520(−32)=−38a_5 = \frac{7(3)-16}{5 \cdot 4} a_3 = \frac{5}{20} \left(-\frac{3}{2}\right) = -\frac{3}{8}
    • n=5n = 5: a7=7(5)−167⋅6a5=1942(−38)=−19112a_7 = \frac{7(5)-16}{7 \cdot 6} a_5 = \frac{19}{42} \left(-\frac{3}{8}\right) = -\frac{19}{112}
    • 一般通項(n=2k+1n = 2k+1):
🔒

後續完整解題步驟與【答案】

免費註冊,享三天全站完整詳解閱覽。

免費註冊

第 2 題20 分

  1. (20%) Let a function f(x)f(x) defined over [−L,L][-L, L].
    (a)(5%) What are the conditions for the Fourier series to uniformly and absolutely converge to f(x)f(x)?
    (b)(15%) Under the conditions you give in (a), prove that the uniformly and absolutely convergence of the Fourier series.

登入後即可作答並保存紀錄。

這一題的完整詳解

(a) 統一、絕對收斂的充分條件
令 ff 為 2L2L 週期函數,定義於 [−L,L][-L,L]。若滿足下列條件,則其傅里葉級數

f(x)∼a02+∑n=1∞(ancos⁡nπxL+bnsin⁡nπxL)f(x)\sim\frac{a_{0}}2+\sum_{n=1}^{\infty}\bigl(a_{n}\cos\frac{n\pi x}{L}+b_{n}\sin\frac{n\pi x}{L}\bigr)

同時在 [−L,L][-L,L] 上一致收斂且絕對收斂:

  1. ff 在 [−L,L][-L,L] 連續 (包含端點的周期延伸)。
  2. ff 的導函數 f′f' 在 [−L,L][-L,L] 分段連續,且僅有有限個跳躍不連續點,亦即 f′∈PC[−L,L]f'\in PC[-L,L]。
  3. f′f' 為有界變差(等價於 f′∈BV[−L,L]f'\in BV[-L,L])。

此條件等價於「f∈C1f\in C^{1} except finit點跳躍」或「ff 為 Lipschitz (α>0)」的常用敘述。


(b) 證明

  1. 係數衰減速度
    對於 n≥1n\ge1,利用分部積分(週期性邊界項消失)得
an=1L∫−LLf(x)cos⁡nπxL dx=Lnπ1L∫−LLf′(x)sin⁡nπxL dx,bn=1L∫−LLf(x)sin⁡nπxL dx=−Lnπ1L∫−LLf′(x)cos⁡nπxL dx.\begin{aligned} a_n &=\frac1L\int_{-L}^{L}f(x)\cos\frac{n\pi x}{L}\,dx =\frac{L}{n\pi}\frac1L\int_{-L}^{L}f'(x)\sin\frac{n\pi x}{L}\,dx,\\[2mm] b_n &=\frac1L\int_{-L}^{L}f(x)\sin\frac{n\pi x}{L}\,dx =-\frac{L}{n\pi}\frac1L\int_{-L}^{L}f'(x)\cos\frac{n\pi x}{L}\,dx . \end{aligned}

再次對 f′f' 作分部積分(因 f′f' 只在有限點有跳躍,跳躍貢獻為有界常數),可得到

∣an∣,  ∣bn∣≤Cn2,C:=Lπ(Var⁡[−L,L](f′)),|a_n|,\;|b_n|\le\frac{C}{n^{2}},\qquad C:=\frac{L}{\pi}\bigl(\operatorname{Var}_{[-L,L]}(f')\bigr),

🔒

後續完整解題步驟與【答案】

免費註冊,享三天全站完整詳解閱覽。

免費註冊

第 3 題10 分

  1. (10%) Determine the Laplace transform of e−3t∫0te3τcos⁡(2τ)dτe^{-3t} \int_{0}^{t} e^{3\tau} \cos(2\tau) d\tau.

登入後即可作答並保存紀錄。

這一題的完整詳解

核心觀念

本題旨在考驗對**拉普拉斯轉換(Laplace Transform)**基本性質的熟練度與觀察能力。核心涵蓋下列觀念與定理:

  1. 基本轉換對(Basic Laplace Transforms):

    • L{eat}=1s−a\mathcal{L}\{e^{at}\} = \frac{1}{s-a}
    • L{cos⁡(bt)}=ss2+b2\mathcal{L}\{\cos(bt)\} = \frac{s}{s^2+b^2}
  2. 摺積定理(Convolution Theorem):

    • 若函數 f(t)=(g∗h)(t)=∫0tg(t−τ)h(τ)dτf(t) = (g * h)(t) = \int_{0}^{t} g(t-\tau) h(\tau) d\tau,則其拉普拉斯轉換為 L{f(t)}=G(s)⋅H(s)\mathcal{L}\{f(t)\} = G(s) \cdot H(s)。
  3. 頻域位移定理(ss-shifting Property)與時域積分定理(Integration Property):

    • L{eatg(t)}=G(s−a)\mathcal{L}\{e^{at} g(t)\} = G(s-a)
    • L{∫0tg(τ)dτ}=G(s)s\mathcal{L}\left\{\int_{0}^{t} g(\tau) d\tau\right\} = \frac{G(s)}{s}

解題方法

【解法一:摺積定理(最精簡切入點)】

觀察被積式與外圍指數項 e−3te^{-3t} 的關係,將 e−3te^{-3t} 搬入積分符號內:
f(t)=e−3t∫0te3τcos⁡(2τ)dτ=∫0te−3(t−τ)cos⁡(2τ)dτf(t) = e^{-3t} \int_{0}^{t} e^{3\tau} \cos(2\tau) d\tau = \int_{0}^{t} e^{-3(t-\tau)} \cos(2\tau) d\tau

此形式完全符合兩函數摺積(Convolution)的定義 (g∗h)(t)=∫0tg(t−τ)h(τ)dτ(g * h)(t) = \int_{0}^{t} g(t-\tau) h(\tau) d\tau,其中:

  • g(t)=e−3t  ⟹  G(s)=L{e−3t}=1s+3g(t) = e^{-3t} \implies G(s) = \mathcal{L}\{e^{-3t}\} = \frac{1}{s+3}
  • h(t)=cos⁡(2t)  ⟹  H(s)=L{cos⁡(2t)}=ss2+22=ss2+4h(t) = \cos(2t) \implies H(s) = \mathcal{L}\{\cos(2t)\} = \frac{s}{s^2+2^2} = \frac{s}{s^2+4}

根據摺積定理,兩函數時域摺積的拉普拉斯轉換等於各自拉普拉斯轉換在頻域上的乘積:
L{f(t)}=L{(g∗h)(t)}=G(s)⋅H(s)\mathcal{L}\{f(t)\} = \mathcal{L}\{(g * h)(t)\} = G(s) \cdot H(s)

代入 G(s)G(s) 與 H(s)H(s):
L{f(t)}=1s+3⋅ss2+4=s(s+3)(s2+4)\mathcal{L}\{f(t)\} = \frac{1}{s+3} \cdot \frac{s}{s^2+4} = \frac{s}{(s+3)(s^2+4)}


【解法二:位移定理與積分定理(標準定理推導)】

設 g(t)=e3tcos⁡(2t)g(t) = e^{3t} \cos(2t),先求其拉普拉斯轉換 G(s)G(s)。根據頻域位移定理:
G(s)=L{e3tcos⁡(2t)}=L{cos⁡(2t)}∣s→s−3=s−3(s−3)2+4=s−3s2−6s+13G(s) = \mathcal{L}\{e^{3t} \cos(2t)\} = \left. \mathcal{L}\{\cos(2t)\} \right|_{s \to s-3} = \frac{s-3}{(s-3)^2 + 4} = \frac{s-3}{s^2 - 6s + 13}

設 h(t)=∫0tg(τ)dτh(t) = \int_{0}^{t} g(\tau) d\tau,根據時域積分定理:
H(s)=L{∫0tg(τ)dτ}=G(s)s=s−3s(s2−6s+13)H(s) = \mathcal{L}\left\{\int_{0}^{t} g(\tau) d\tau\right\} = \frac{G(s)}{s} = \frac{s-3}{s(s^2 - 6s + 13)}

目標函數為 f(t)=e−3th(t)f(t) = e^{-3t} h(t),再次套用頻域位移定理(將 ss 替換為 s+3s+3):
F(s)=L{e−3th(t)}=H(s+3)=(s+3)−3(s+3)[(s+3)2−6(s+3)+13]F(s) = \mathcal{L}\{e^{-3t} h(t)\} = H(s+3) = \frac{(s+3)-3}{(s+3) \left[ (s+3)^2 - 6(s+3) + 13 \right]}

展開分母中括號項:
(s+3)2−6(s+3)+13=s2+6s+9−6s−18+13=s2+4(s+3)^2 - 6(s+3) + 13 = s^2 + 6s + 9 - 6s - 18 + 13 = s^2 + 4

因此:
F(s)=s(s+3)(s2+4)F(s) = \frac{s}{(s+3)(s^2+4)}


【解法三:定積分直接展開法(計算與驗算)】

🔒

後續完整解題步驟與【答案】

免費註冊,享三天全站完整詳解閱覽。

免費註冊

第 4 題

  1. Let r(x)r(x) be a periodic triangular wave function with fundamental period PaP_a as shown in Figure 1 below, and then the phase angle form of the Fourier series of the function rr can be known.
    🖼️【此處有附圖,請對照原卷】
    Figure 1. Periodic triangular wave function r(x)r(x).
    (a) Here we want to approximate the function with a partial sum of the Fourier series of rr. In order to build the partial sum, the strategy here is to include a group of nthn^{th} harmonics cncos⁡(nω0x+δn)c_n \cos(n\omega_0 x + \delta_n) with amplitudes not smaller than 15% of the largest amplitude, where cnc_n is the nthn^{th} harmonic amplitude, δn\delta_n is the nthn^{th} phase angle, and ω0=2πPa\omega_0 = \frac{2\pi}{P_a}. Suggest what frequencies nω0n\omega_0 should be included in this partial sum. The frequency nω0n\omega_0 can be zero or positive. Show the details. (25%)
    (b) Now we have a nonhomogeneous 2nd-order differential equation y′′(x)+3y′(x)+2y(x)=r(x)y''(x) + 3y'(x) + 2y(x) = r(x), where r(x)r(x) is illustrated in Figure 1. Our strategy here is to first approximate the periodic function rr with the partial sum of the Fourier series of rr obtained in (a) and then to solve this differential equation. If PbP_b is fundamental period of the particular solution obtained based on this approach, find the value of 2πPb\frac{2\pi}{P_b}. Show the details. (25%)
🖼️ 本題含圖表,以下為原卷對應頁面:
原卷第 2 頁原卷第 3 頁

登入後即可作答並保存紀錄。

這一題的完整詳解

核心觀念

本題先由圖形讀出三角波的基本週期與對稱性,再用傅立葉係數判斷哪些諧波振幅達到門檻。接著將保留的諧波當作微分方程的輸入,求出週期特解的角頻率。

解題方法

圖中波形在偶數整點取 −k/2-k/2、奇數整點取 k/2k/2,相鄰波峰(或波谷)相距 22,因此 Pa=2P_a=2、ω0=π\omega_0=\pi。波形關於 yy 軸偶對稱,且一個週期內正負面積相抵。

(a) 判斷應保留的頻率

在 −1≤x≤1-1\le x\le 1 上,圖形可寫成

r(x)=k∣x∣−k2.r(x)=k|x|-\frac{k}{2}.

令傅立葉級數為

r(x)=a02+∑n=1∞[ancos⁡(nπx)+bnsin⁡(nπx)].r(x)=\frac{a_0}{2}+\sum_{n=1}^{\infty} \left[a_n\cos(n\pi x)+b_n\sin(n\pi x)\right].

偶對稱使 bn=0b_n=0,正負面積相抵使 a0=0a_0=0。對 n≥1n\ge1,

an=2∫01(kx−k2)cos⁡(nπx) dx=2k[(−1)n−1]n2π2.\begin{aligned} a_n &=2\int_0^1\left(kx-\frac{k}{2}\right)\cos(n\pi x)\,dx\\ &=\frac{2k\left[(-1)^n-1\right]}{n^2\pi^2}. \end{aligned}

因此偶數 nn 的係數為 00;所有奇數 nn 的係數均為

an=−4kπ2n2.a_n=-\frac{4k}{\pi^2n^2}.

寫成振幅—相位形式時,奇數諧波的振幅為 cn=∣an∣=4k/(π2n2)c_n=|a_n|=4k/(\pi^2n^2),相位角可取 δn=π\delta_n=\pi。故

r(x)=−4kπ2∑m=0∞cos⁡((2m+1)πx)(2m+1)2.r(x)=-\frac{4k}{\pi^2} \sum_{m=0}^{\infty}\frac{\cos\big((2m+1)\pi x\big)}{(2m+1)^2}.
🔒

後續完整解題步驟與【答案】

免費註冊,享三天全站完整詳解閱覽。

免費註冊

其他考古題