📚 Fourier Series | 傅里叶级数
Fourier series is a cornerstone of mathematical analysis that enables us to represent periodic functions as infinite sums of sines and cosines. Named after Jean-Baptiste Joseph Fourier, this technique bridges the discrete and continuous worlds, finding profound applications in physics, engineering, and signal processing. In IB Mathematics Higher Level, understanding Fourier series deepens your appreciation of function approximation, orthogonality, and convergence.
傅里叶级数是数学分析的基石，它使我们能够将周期函数表示为无限个正弦与余弦的和。这一方法以约瑟夫・傅里叶的名字命名，连接了离散与连续的世界，在物理、工程和信号处理中有着深远应用。在 IB 数学高水平课程中，理解傅里叶级数能加深你对函数逼近、正交性和收敛性的认识。
1. Introduction to Periodic Functions and Fourier Series | 周期函数与傅里叶级数简介
n its domain. The smallest positive T satisfying this condition is the fundamental period.
如果对于定义域内的所有 x 都有 f(x + T) = f(x)，则称函数 f(x) 是以 T 为周期的周期函数。满足该条件的最小正数 T 称为基本周期。
Fourier's key insight was that any reasonably behaved periodic function can be decomposed into a constant term plus harmonically related sine and cosine waves. The resulting representation, called a Fourier series, writes f(x) as a linear combination of sin(nx) and cos(nx) for integer n.
傅里叶的关键洞见是，任何行为良好的周期函数都可以分解为一个常数项加上频率成整数倍关系的正弦波和余弦波。这样得到的表达式称为傅里叶级数，即将 f(x) 表示为 sin(nx) 和 cos(nx)（n 为整数）的线性组合。
This decomposition reveals the frequency spectrum of the function and forms the foundation of harmonic analysis.
这种分解揭示了函数的频谱，构成了调和分析的基础。
2. Orthogonality of Trigonometric Functions | 三角函数的正交性
d. For integers m, n >= 1, the following integral identities hold on the interval [-π, π]:
傅里叶系数的推导依赖于正弦和余弦函数在一个完整周期上的正交性。对于整数 m, n >= 1，在区间 [-π, π] 上有以下积分恒等式：
∫₋π⁺π cos(mx) cos(nx) dx = π δₘₙ, where δₘₙ is the Kronecker delta (1 if m = n, 0 otherwise).
∫₋π⁺π cos(mx) cos(nx) dx = π δₘₙ，其中 δₘₙ 为克罗内克 δ 符号（当 m = n 时为 1，否则为 0）。
∫₋π⁺π sin(mx) sin(nx) dx = π δₘₙ (and 0 if m = 0).
∫₋π⁺π sin(mx) sin(nx) dx = π δₘₙ（当 m = 0 时为 0）。
∫₋π⁺π sin(mx) cos(nx) dx = 0 for all m, n.
对所有 m, n 均有 ∫₋π⁺π sin(mx) cos(nx) dx = 0。
These orthogonality relations allow us to isolate individual coefficients by multiplying the series by a specific trigonometric function and integrating term by term.
利用这些正交关系，我们可以将级数乘以某个特定的三角函数并逐项积分，从而分离出各个系数。
3. Fourier Coefficients for Period 2π | 周期为 2π 的傅里叶系数
For a periodic function f(x) with period 2π, the Fourier series is written as:
对于周期为 2π 的周期函数 f(x)，其傅里叶级数表示为：
f(x) = a₀/2 + Σₙ₌₁∞ (aₙ cos(nx) + bₙ sin(nx))
Multiplying both sides by cos(kx), integrating from -π to π, and using orthogonality yields the Euler formulas for the Fourier coefficients:
将等式两边乘以 cos(kx)，再从 -π 积分到 π，利用正交性即可得到傅里叶系数的欧拉公式：
a₀ = (1/π) ∫₋π⁺π f(x) dx
aₙ = (1/π) ∫₋π⁺π f(x) cos(nx) dx, n >= 1
bₙ = (1/π) ∫₋π⁺π f(x) sin(nx) dx, n >= 1
The constant term a₀/2 represents the average value of f(x) over one period; the aₙ and bₙ coefficients capture the amplitudes of the cosine and sine harmonics, respectively.
常数项 a₀/2 表示 f(x) 在一个周期内的平均值；系数 aₙ 和 bₙ 则分别捕捉余弦和谐波和正弦谐波的振幅。
These formulas are valid whenever the integrals exist, which holds for piecewise continuous functions.
只要上述积分存在，这些公式便成立，这对于分段连续函数总是满足的。
4. Fourier Coefficients for General Period 2L | 一般周期 2L 的傅里叶系数
T = 2L, the Fourier series is adapted by scaling the argument. The series takes the form:
如果 f(x) 的周期 T = 2L，傅里叶级数需要通过对自变量进行缩放来调整。级数形式为：
f(x) = a₀/2 + Σₙ₌₁∞ (aₙ cos(nπx/L) + bₙ sin(nπx/L))
The coefficients are then given by integrals over a symmetric interval of length 2L:
相应系数由长度为 2L 的对称区间上的积分给出：
a₀ = (1/L) ∫₋L⁺L f(x) dx
aₙ = (1/L) ∫₋L⁺L f(x) cos(nπx/L) dx
bₙ = (1/L) ∫₋L⁺L f(x) sin(nπx/L) dx
A useful summary can be arranged as follows:
可以将公式总结为下表：
| Period / 周期 | Argument / 自变量 | aₙ formula / aₙ 公式 | bₙ formula / bₙ 公式 |
|---|---|---|---|
| 2π | nx | (1/π) ∫₋π⁺π f(x) cos(nx) dx | (1/π) ∫₋π⁺π f(x) sin(nx) dx |
| 2L | nπx/L | (1/L) ∫₋L⁺L f(x) cos(nπx/L) dx | (1/L) ∫₋L⁺L f(x) sin(nπx/L) dx |
Choosing an integration interval that matches one full period is often more convenient, such as from 0 to 2L.
实践中，选择一个恰好覆盖一个完整周期的积分区间常常更方便，例如从 0 到 2L。
5. Examples of Fourier Series Expansion | 傅里叶级数展开示例
2π defined by f(x) = 1 for 0 < x < π and f(x) = -1 for -π < x < 0, extended periodically.
考虑周期为 2π 的方波：当 0 < x < π 时 f(x) = 1，当 -π < x < 0 时 f(x) = -1，并周期延拓。
Since the function is odd, a₀ = 0 and aₙ = 0 for all n. The sine coefficients are:
由于该函数为奇函数，所有 a₀ = 0 且 aₙ = 0。正弦系数为：
bₙ = (1/π) [ ∫₋π⁰ (-1) sin(nx) dx + ∫₀⁺π (1) sin(nx) dx ] = (2/π) ∫₀⁺π sin(nx) dx
Evaluating the integral gives bₙ = (2/(nπ))(1 - cos(nπ)) = (2/(nπ))(1 - (-1)ⁿ).
计算积分得 bₙ = (2/(nπ))(1 - cos(nπ)) = (2/(nπ))(1 - (-1)ⁿ)。
1 and bₙ = 0. For odd n = 2k-1, we obtain b₂ₖ₋₁ = 4/(π(2k-1)). Thus the Fourier series is:
对于偶数 n，cos(nπ)=1，bₙ = 0。对于奇数 n = 2k-1，得到 b₂ₖ₋₁ = 4/(π(2k-1))。因此傅里叶级数为：
f(x) = (4/π) Σₖ₌₁∞ sin((2k-1)x) / (2k-1)
This series contains only odd harmonics, and each coefficient decays as 1/(2k-1).
该级数仅包含奇次谐波，每个系数按照 1/(2k-1) 衰减。
Another classic example is the triangular wave, which yields a cosine series with coefficients decaying as 1/n², demonstrating faster convergence.
另一个经典示例是三角波，它产生的是余弦级数，系数按 1/n² 衰减，展现出更快的收敛速度。
6. Even and Odd Functions: Cosine and Sine Series | 偶函数与奇函数：余弦与正弦级数
series simplifies drastically: all bₙ = 0 because sin(nx) is odd. The coefficients become:
如果 f(x) 是偶函数（f(-x) = f(x)），其傅里叶级数将大幅简化：由于 sin(nx) 为奇函数，所有 bₙ = 0。系数变为：
aₙ = (2/π) ∫₀⁺π f(x) cos(nx) dx, a₀ = (2/π) ∫₀⁺π f(x) dx
The resulting series is a Fourier cosine series containing only cosine terms.
这样得到的级数是仅包含余弦项的傅里叶余弦级数。
f(x) is an odd function (f(-x) = -f(x)), all aₙ = 0, and we obtain a Fourier sine series:
类似地，如果 f(x) 是奇函数（f(-x) = -f(x)），则所有 aₙ = 0，我们得到傅里叶正弦级数：
bₙ = (2/π) ∫₀⁺π f(x) sin(nx) dx
For period 2L, the limits are adjusted to 0 to L, and the argument becomes nπx/L, with factor 2/L.
对于周期 2L，积分限调整为 0 到 L，自变量变为 nπx/L，因子变为 2/L。
Recognising the symmetry of the function before computing coefficients can save a significant amount of work.
在计算系数之前识别函数的对称性，可以节省大量的计算工作。
7. Half-Range Expansions | 半范围展开
Sometimes a function is defined only on the interval [0, L] and we wish to represent it using a Fourier series. We can extend the function to [-L, L] either as an even function or as an odd function, leading to half-range cosine and sine series respectively.
有时函数仅仅定义在区间 [0, L] 上，而我们希望用傅里叶级数来表示它。我们可以将函数延拓到 [-L, L] 上，既可以选择偶延拓，也可以选择奇延拓，分别得到半范围余弦级数和半范围正弦级数。
For a half-range cosine expansion (even extension), the coefficients are:
对于半范围余弦展开（偶延拓），系数为：
a₀ = (2/L) ∫₀⁺L f(x) dx, aₙ = (2/L) ∫₀⁺L f(x) cos(nπx/L) dx
For a half-range sine expansion (odd extension), the coefficients are:
对于半范围正弦展开（奇延拓），系数为：
bₙ = (2/L) ∫₀⁺L f(x) sin(nπx/L) dx
nds on the boundary conditions of the problem or the desired behaviour at x = 0 and x = L.
选择哪种延拓方式取决于问题的边界条件或在 x = 0 和 x = L 处所期望的性质。
8. Convergence and Dirichlet Conditions | 收敛性与狄利克雷条件
Not every periodic function possesses a convergent Fourier series in the strict sense. The Dirichlet conditions provide sufficient criteria for pointwise convergence:
并非每一个周期函数都具有严格意义下收敛的傅里叶级数。狄利克雷条件给出了逐点收敛的充分判据：
- The function f(x) must be periodic and piecewise continuous on the interval.
- 函数 f(x) 必须是周期的，且在区间上分段连续。
- It must have only a finite number of maxima, minima, and discontinuities in one period.
- 在一个周期内，它只能有有限多个极大值、极小值和间断点。
- The integrals defining the Fourier coefficients must be finite.
- 定义傅里叶系数的积分必须有限。
When these conditions are satisfied, the Fourier series converges to f(x) at all points of continuity. At a jump discontinuity, it converges to the average of the left-hand and right-hand limits: (f(x⁻) + f(x⁺))/2.
当这些条件满足时，傅里叶级数在所有连续点上收敛到 f(x)。在跳跃间断点处，它收敛到左极限与右极限的平均值：(f(x⁻) + f(x⁺))/2。
0 converges to (1 + (-1))/2 = 0, matching the observed average value at the discontinuity.
例如，方波的级数在 x = 0 处收敛到 (1 + (-1))/2 = 0，这与在该间断点处的平均值一致。
9. Gibbs Phenomenon and Approximation | 吉布斯现象与逼近
Even when a Fourier series converges pointwise, the convergence near a jump discontinuity is not uniform. The Gibbs phenomenon refers to the overshoot that occurs when we truncate the series to a finite number of terms N.
即使傅里叶级数逐点收敛，在跳跃间断点附近的收敛也不是一致的。吉布斯现象指的是，当我们用有限项 N 截断级数时，会出现过冲的现象。
As N increases, the partial sum oscillates near the jump, and the maximum overshoot approaches approximately 9% of the jump magnitude, regardless of how large N becomes.
随着 N 增大，部分和在跳跃点附近振荡，最大过冲量约等于跳跃幅度的 9%，无论 N 变得多大，这一比例基本保持不变。
This highlights the fact that infinite series can approximate a function accurately in the mean-squared sense, yet fail to capture the exact behaviour near abrupt changes.
这突出表明，无穷级数可以在均方意义下精确逼近函数，却无法完美捕捉突变附近的精确行为。
In practical applications, we use smoothing techniques or consider the implications of Gibbs phenomenon when processing signals.
在实际应用中，我们在处理信号时会使用平滑技术或考虑吉布斯现象的影响。
10. Applications and Significance | 应用与意义
Fourier series provides a powerful method for solving partial differential equations such as the heat equation, wave equation, and Laplace's equation, by separating variables and expressing solutions as superpositions of normal modes.
傅里叶级数为求解热传导方程、波动方程和拉普拉斯方程等偏微分方程提供了强大的工具，通过分离变量并将解表示为简正模式的叠加。
In signal processing, any periodic signal can be analysed in terms of its harmonic content, which is the cornerstone of Fourier analysis and filter design.
在信号处理中，任何周期信号都可以根据其谐波成分进行分析，这是傅里叶分析和滤波器设计的基石。
Furthermore, the concept of expressing a function in an orthogonal basis underpins many areas of modern mathematics and physics, including quantum mechanics and image compression.
此外，在正交基中展开函数的概念支撑着包括量子力学和图像压缩在内的许多现代数学与物理分支。
Mastering Fourier series equips you with a lens through which to view complex oscillatory phenomena as simple, elegant sums of sine and cosine waves.
掌握傅里叶级数让你拥有一个独特的视角，能够将复杂的振荡现象视为正弦波与余弦波的简洁、优雅之和。
Published by TutorHao | Mathematics Revision Series | aleveler.com
更多咨询请联系16621398022（同微信）
或截图后用「扫一扫 → 相册」选择本图