Unit Commitment

In this section, we give a standard formulation for the Unit Commitment problem. We normally solve this problem as a mixed integer linear programme. However it should be noted that there are nonlinearities in the objective function in form of the binary decision variables.

Note

You can use this formulation to formulate and solve your own example in AIMMS. We will publish an accompanying simple example based on this formulation soon.

Objective Function

As stated above, the aim of the UC problem is to ensure there is enough generation committed at each point in time within a set planning timescale. The objective function for the UC problem is defined in its most general form as below:

(7)\[\min_{\Phi} \sum_{t \in T}\sum_{g \in G} (C_{g}^{SU}(u_{g}^{t})+C_{g}^{SD}(u_{g}^{t})+f_g(P_{g}^{t})+C_{g}^{0}(u_{g}^{t})+\delta_{g} ^{+}(P_{g}^{t})+\delta_{g}^{-}(P_{g}^{t}))\]

In equation (7) the following are defined:

• \(\Phi=(P_{g,t},u_{g,t},V_{a}^{n})^T\) is the set of all state variables in which \(P_{g,t}\) is the active power output of generator g at time t, \(u_{g,t}\) is the binary decision variable for starting up or shutting down of generator g at time t, and \(V_{a}^{n}\) is the voltage angle for node n

• \(G\) and \(T\) are the sets of all generators and time steps respectively.

• \(C_{g}^{SU}\) and \(C_{g}^{SD}\) are the cost functions for starting up or shutting down for generator unit g

• \(f_g\) is the fuel cost function for generator unit g

• \(C_{g}^{0}\) is the running cost function for standby generators

• \(\delta_{g}^{+}\) and \(\delta_{g}^{-}\) are the ramping up and ramping down cost functions of generator unit g

Constraints

The following constraints apply for a DC formulation of the UC problem:

(8)\[u_{g}^{t}P_{g}^{\min} \leq P_{g}^{t} \leq u_{g}^{t}P_{g}^{\max}\]

(9)\[P_{g,n}^{t} - P_{d,n}^{t} + \sum_{n \in N} B_{in}(\theta_{i}^{t} - \theta_{n}^{t}) = 0\]

(10)\[B_{in}(\theta_{i}^{t} - \theta_{n}^{t}) \leq L_{in}^{\max}\]

(11)\[\delta_{g}^{\min} \leq P_{g}^{t} - P_{g}^{t-1} \leq \delta_{g}^{\max}\]

The following are defined:

• Equation (8) is the maximum and minimum allowable active power limit for each generator.

• Equation (9) is the Active nodal power balance for each node in the system. Set \(N\) is defined as the set of all the nodes in the system.

• \(P_{d,n}\), whereas \(P_{g,n}\) is the total generation at node \(n\).

• \(B_{jk}\) is the susceptance of the transmission line connecting nodes \(j\) and \(k\).

• Equation (10) is the maximum allowable thermal (capacity) limit for each line in the system in units of MW. \(L_{jk}\) is the active power transmission limit for the line connecting nodes \(j\) and \(k\).

• All transmission line elements (susceptance and line limits) are contained in the Set \(L\).

• Equation (11) is the ramping up and down limits on each generator in units of MW/h.

Note

In equations above only the most fundamental equations are given and some details have been omitted (for example provisions of primary frequency response which is required in some regions, countries.). In addition, if we want to run the UC as a multi-period problem then there should be appropriate inter-temporal constraints between each time-step (i.e. generators’ status from the previous timestep must carry over to the next timestep).

References

[1] Zimmerman, R. D., and Murillo-Sanchez, C. E., 2020, MATPOWER Optimal Scheduling Tool - MOST 1.1 User’s Manual