Time-coordinated optimal power flow for energy service networks - Modelling Distributed Energy Resources in Energy Service Networks - page 93

Equally important, for the aggregator to function properly in the provision of
demand response services it needs to account for every PHEV and CHP unit avail-
able. Under this premise, the aggregator would employ dispatch algorithms and
then communicate its decisions to each DER device, thus optimally coordinating
the serviceable capacity. Although the monetary value for providing ancillary ser-
vices is beyond the boundaries of this work, it is important to learn the technical
value of such load control actions.

5.1.2 Optimisation solver
The TCOPF is a multi-period non-linear optimisation problem with linear and non-
linear equality and inequality constraints which have been coded in the gPROMS TM
software [209]. The objective of the optimisation solver is to guarantee the best
possible solution for the whole energy service system by simultaneously calculating
the operating values of the infrastructures and the embedded devices. This means
the solver is global and unbiased when solving any objective function proposed,
giving no preference to any particular DNO or DER technology.
gPROMS TM provides an environment for modelling the behaviour of highly
complex systems. Although by default, gPROMS TM mainly manages optimisation
problems as if they were of the dynamic type, it is also possible to carry out various
multi-period steady-state optimisations of systems. From the mathematical point of
view, this is equivalent to solving snapshots of an algebraic problem in which a non-
linear objective function is minimised or maximised subject to non-linear constraints
by manipulating a set of optimisation decision variables that may be either continuous
or discrete. Thus, the gPROMS TM software features are well suited to implement the

TCOPF program.
The TCOPF uses the steady-state control vector parameterisation (CVP-SS)
solver approach, available in gPROMS TM , in order to execute the analysis of electric
and natural gas networks with embedded technologies. Although various piecewise
controls exist such as linear and polynomial, in this work the CVP-SS method
assumes that all the time-varying control variables are piecewise constant functions
operating over a specified number of intervals, as the example depicted in Figure 5.3
illustrates.
As their name implies, piecewise constant controls have a fixed value over a
certain part of the time horizon before discretely changing to a different value over
the next interval.
Hence, the piecewise properties of the control variables can be easily imple-
mented and properly adopted to portray on-load tap-changer (OLTC), compressors,
PHEVs and CHPs with thermal stores. The values of these control variables for each
time interval are determined by the optimisation algorithm, while the length of the
time intervals is part of the input data.
Furthermore, since the TCOPF program has a small amount of control variables
when compared to the total number of variables present in a problem, the solver
algorithm has to deal only with a rather small number of decisions. This makes the
CVP-SS approach applicable to solve optimal power flow and dispatch problems