NeumannBoundaryCondition<T>
Models & Types · AiDotNet.PhysicsInformed.PDEs
Pins the solution’s slope on the boundary rather than its value: ∂u/∂n = flux.
Implements: IBoundaryCondition<T>
For Beginners
A Dirichlet condition says what the value is at the edge. A Neumann condition says how fast it is changing there — which is what you know when the edge is insulated, or when a fixed amount is flowing in. An insulated rod end has zero flux: no heat crosses it, so the temperature gradient there is zero. That is the default here.
How It Works
The other half of what a boundary-value problem usually needs, alongside
DirichletBoundaryCondition (#2099).
Example
var architecture = new NeuralNetworkArchitecture<double>(
InputType.OneDimensional, NeuralNetworkTaskType.Regression, inputSize: 2, outputSize: 1);
// Insulated at both ends: no flux across the boundary
var insulated = new NeumannBoundaryCondition<double>(flux: 0.0, lowerBound: 0.0, upperBound: 1.0);
var pinn = new PhysicsInformedNeuralNetwork<double>(
architecture, new HeatEquation<double>(), new IBoundaryCondition<double>[] { insulated });
Constructors
| Constructor | Summary |
|---|---|
NeumannBoundaryCondition(double flux, double lowerBound, double upperBound, int coordinate, double tolerance) |
Initializes a Neumann condition on one input coordinate. |
Properties
| Property | Summary |
|---|---|
string Name { get; } |
Methods
| Method | Summary |
|---|---|
T ComputeBoundaryResidual(Vector<T> inputs, Vector<T> outputs, PDEDerivatives<T> derivatives) |
|
bool IsOnBoundary(Vector<T> inputs) |