|
105 | 105 | </section> |
106 | 106 | <section id="time-integration"> |
107 | 107 | <span id="marching"></span><h3>Time Integration<a class="headerlink" href="#time-integration" title="Link to this heading"></a></h3> |
108 | | -<p>Consider the initial‐value problem for passive tracer advection in a continuous velocity field</p> |
| 108 | +<p>Consider the initial-value problem for passive tracer advection in a continuous velocity field</p> |
109 | 109 | <div class="math notranslate nohighlight"> |
110 | 110 | \[\frac{d\mathbf{x}}{dt} = \sigma\,\mathbf{u}(\mathbf{x},t)\,, |
111 | 111 | \mathbf{x}(t_n)=\mathbf{x}_n\,,\]</div> |
112 | 112 | <p>where <span class="math notranslate nohighlight">\(\sigma = \pm1\)</span> selects forward or backward integration.</p> |
113 | 113 | <p><strong>Explicit Euler Method</strong></p> |
114 | | -<p>The first‐order explicit Euler scheme advances the position by sampling the velocity at the beginning of the time step:</p> |
| 114 | +<p>The first-order explicit Euler scheme advances the position by sampling the velocity at the beginning of the time step:</p> |
115 | 115 | <div class="math notranslate nohighlight"> |
116 | 116 | \[\begin{split}\mathbf{u}_n = \mathbf{u}(\mathbf{x}_n,t_n),\\ |
117 | 117 | \mathbf{x}_{n+1} = \mathbf{x}_n + \sigma\,\Delta t\,\mathbf{u}_n.\end{split}\]</div> |
118 | 118 | <p>This method incurs a global error of order <span class="math notranslate nohighlight">\(O(\Delta t)\)</span> and requires only one velocity evaluation per step.</p> |
119 | | -<p><strong>Second‐Order Runge–Kutta (Heun’s Method)</strong></p> |
120 | | -<p>Heun’s method attains second‐order accuracy by combining predictor and corrector slopes:</p> |
| 119 | +<p><strong>Second-Order Runge-Kutta (Heun’s Method)</strong></p> |
| 120 | +<p>Heun’s method attains second-order accuracy by combining predictor and corrector slopes:</p> |
121 | 121 | <div class="math notranslate nohighlight"> |
122 | 122 | \[\begin{split}k_1 = \sigma\,\mathbf{u}(\mathbf{x}_n,t_n),\\ |
123 | 123 | \mathbf{x}^* = \mathbf{x}_n + \Delta t\,k_1,\\ |
124 | 124 | k_2 = \sigma\,\mathbf{u}(\mathbf{x}^*,t_n + \Delta t),\\ |
125 | 125 | \mathbf{x}_{n+1} = \mathbf{x}_n + \tfrac{\Delta t}{2}\,(k_1 + k_2).\end{split}\]</div> |
126 | 126 | <p>This scheme yields a global error of order <span class="math notranslate nohighlight">\(O(\Delta t^2)\)</span> with two velocity evaluations per step.</p> |
127 | | -<p><strong>Classical Fourth‐Order Runge–Kutta (RK4)</strong></p> |
128 | | -<p>The classical RK4 method achieves fourth‐order accuracy via four slope evaluations at intermediate points:</p> |
| 127 | +<p><strong>Classical Fourth-Order Runge-Kutta (RK4)</strong></p> |
| 128 | +<p>The classical RK4 method achieves fourth-order accuracy via four slope evaluations at intermediate points:</p> |
129 | 129 | <div class="math notranslate nohighlight"> |
130 | 130 | \[\begin{split}k_1 = \mathbf{u}(\mathbf{x}_n,t_n),\\ |
131 | 131 | k_2 = \mathbf{u}\!\bigl(\mathbf{x}_n + \tfrac{\Delta t}{2}k_1,\;t_n + \tfrac{\Delta t}{2}\bigr),\\ |
132 | 132 | k_3 = \mathbf{u}\!\bigl(\mathbf{x}_n + \tfrac{\Delta t}{2}k_2,\;t_n + \tfrac{\Delta t}{2}\bigr),\\ |
133 | 133 | k_4 = \mathbf{u}(\mathbf{x}_n + \Delta t\,k_3,\;t_n + \Delta t),\\ |
134 | 134 | \mathbf{x}_{n+1} = \mathbf{x}_n + \tfrac{\Delta t}{6}\,(k_1 + 2k_2 + 2k_3 + k_4).\end{split}\]</div> |
135 | 135 | <p>This yields a global error of order <span class="math notranslate nohighlight">\(O(\Delta t^4)\)</span> with four velocity evaluations per step.</p> |
136 | | -<p><strong>Sixth‐Order Runge–Kutta (RK6)</strong></p> |
137 | | -<p>The six‐stage scheme uses non‐uniform weights to attain sixth‐order accuracy:</p> |
| 136 | +<p><strong>Sixth-Order Runge-Kutta (RK6)</strong></p> |
| 137 | +<p>The six-stage scheme uses non-uniform weights to attain sixth-order accuracy:</p> |
138 | 138 | <div class="math notranslate nohighlight"> |
139 | 139 | \[\begin{split}k_1 = \mathbf{u}(\mathbf{x}_n,t_n),\\ |
140 | 140 | k_2 = \mathbf{u}\!\bigl(\mathbf{x}_n + \tfrac{\Delta t}{3}k_1,\;t_n + \tfrac{\Delta t}{3}\bigr),\\ |
|
144 | 144 | k_6 = \mathbf{u}\!\bigl(\mathbf{x}_n + \Delta t(-\tfrac{3}{2}k_1 + 2k_2 - \tfrac{1}{2}k_3 + k_4),\;t_n + \Delta t\bigr),\\ |
145 | 145 | \mathbf{x}_{n+1} = \mathbf{x}_n + \Delta t\bigl(\tfrac{1}{20}k_1 + \tfrac{1}{4}k_4 + \tfrac{1}{5}k_5 + \tfrac{1}{2}k_6\bigr).\end{split}\]</div> |
146 | 146 | <p>This scheme incurs a global error of order <span class="math notranslate nohighlight">\(O(\Delta t^6)\)</span> with six velocity evaluations.</p> |
147 | | -<p>All methods assume a continuous velocity interpolation (e.g., tricubic) to supply <span class="math notranslate nohighlight">\(\mathbf{u}\)</span> at arbitrary particle positions and times.</p> |
148 | 147 | </section> |
149 | 148 | </section> |
150 | 149 | <section id="ftle-computation"> |
|
0 commit comments