Entropy Solution

The general form of the systems of conservation laws is

where , .

where and .

Definition

A convex function is called an entropy function of if there exists a function , called the entropy flux, such that the following condition holds

We can left-multiply and obtian an extra conservation law for the entropy function

Definition

A weak solution of is called an entropy solution if for all entropy functions , we have

Integrating the entropy condition in space, we have

That is, the total entropy is non-increasing with respect to time.

Define the entropy variables , i.e.

The entropy condition can be written as

If we assume that U is strictly convex, , the mapping is one-to-one and can be regarded as a change of variables. Setting and , we can rewrite the system according to the entropy variables

The following theorem tells us the symmetry of is equivalent to the existence of entropy function.

Theorem

A strictly convex function serves as an entropy function if and only if

is symmetric positive-definite, and

is symmetric for .

There exist functions and , called potential function and potential fluxes, such that

The derivatives of and are as follows.

The second-order derivatives of and are as follows.

By left-multiplying the entropy variables to and integrating by parts, one can obtain an entropy conservation law of the system on the continuous level satisfying

The second term above is an entropy flux, measuring boundary effects on the total change in entropy inside the domain .

Proof

From the definitions of and , we obtain

Thus,

Examples (scalar)

Consider the conservation law

Definition

A convex function is called an entropy function of if there exists a function , called the entropy flux, such that

By left-multiplying both sides of the equation by , we derive an additional conservation law satisfied by the entropy solution:

Definition

A weak solution of is called an entropy solution if, for all entropy pairs ,

Define the entropy variable . The entropy condition can be written as

Assume that the mapping is 1-1, which can be regarded as a change of variables. Let and , we can rewrite the equation according to the entropy variable as

Example

Consider the linear equation,

Solution

  • If we take and define , we have

  • If we take and define , we have

Example

Consider the burgers equation,

Solution

  • If we take and define , we have

  • If we take and define , we have

Example

Consider the Buckley-Leverett equation,

Solution

  • If we take and define , we have

Examples: 1d Euler Equations

Example

Consider the 1d Euler equations,

Solution

The choice of entropy pair is

where . The entropy variable is