= Tensor divergence theorem
{wiki}
For a continuously differentiable matrix-valued <tensor field> $T$ on a bounded regular <domain> $\Omega\subset\mathbb R^n$, use row-wise <divergence>, $(\operatorname{div}T)_i=\sum_j\partial_jT_{ij}$. Then the <tensor divergence theorem> is
$$
\int_\Omega\operatorname{div}T\,dx=\int_{\partial\Omega}Tn\,dS,
$$
where $n$ is the outward <unit normal vector>. Apply the ordinary <divergence theorem> to row $i$ and collect the component identities. The side on which $n$ multiplies $T$ depends on this index convention; transposing one convention without the other changes the formula.
For a <Cauchy stress tensor> with the convention that <traction> is $Tn$, the theorem turns total surface force into the volume integral of the stress <divergence>. For $T=pI$, it reduces to $\int_\Omega\nabla p=\int_{\partial\Omega}pn$. The related <integration by parts for tensor fields> allows a varying test <vector field> rather than a constant component selector.
Back to article page