IEEE Power & Energy Magazine - May/June 2017 - 32

Hosting capacity is defined as the maximum distributed
generation penetration for which the power system
operates satisfactorily.
indicators to measure the main aspects involved in the
selection process was studied, but finally there is one indicator standing out from all the others: increasing hosting
capacity (HC).

Measuring Hosting Capacity
HC is defined as the maximum distributed generation (DG)
penetration for which the power system operates satisfactorily. According to this definition, numerous methods for
estimating HC have been tested in the past. The most recurrent solution for an effective HC evaluation is based on optimal power flow (OPF), in which the objective function is
designed to maximize the injection of active power within
the selected limits (typically voltage and loading). However,
OPF-based approaches have a strong limitation: the returned
HC value is often referred to as fixed allocation of one or
more generators on the network. Typically, one DRES unit
is simulated and the HC calculation is repeated for all the
network buses, obtaining the results plotted in Figure 1.
From the study of the obtained curves, relevant information about the amount of production that can be connected
to the network can be deduced. This method is traditionally
adopted in network planning due to its simplicity. However,
a single generator is often an unrealistic representation of
the DG expansion since it is more dispersed throughout the
distribution network, with no control from grid operators.
In IGREENGrid, a simulation approach is preferred in
which numerous allocations and dimensions of multiple generation units are considered for each selected network, and
the HC is deduced by comparing the voltage/current limits

Bus HC (MW)

Voltage HC
Current HC
Overall HC

30
20
10
0

4
6
8
10
Electrical Distance (Ω)

figure 1. The traditional approach for network HC estimation (OPF based).
32

ieee power & energy magazine

with the set of resulting levels from simulations (Figure 2).
The returned HC is a series of values describing the realistic
potential of the network for DG integration, including the
dependency between maximum generation and position of
DG units, the best and worst cases of HC, and eventual presence of network bottlenecks.
The extraction of HC values is computationally demanding due to the necessity of simulating several DG scenarios
(Monte Carlo simulation). The process can be faster when
combined with an OPF-based approach. Once a value of
nominal power is randomly assigned to each simulated DG
unit, the optimization problem can be set up to maximize
production by DRESs while respecting network constraints
but maintaining the initial power proportion among generators. At this point, when repeating the OPF for several DG
initial power distributions, the returned values of maximum
generation can be summarized by the probabilistic functions, as shown in Figure 3.
Having checked the equivalence of these results with
the ones plotted in Figure 2, it can be concluded that the
proposed probabilistic method is more realistic than the traditional one (Figure 1) where a single DG unit is normally
considered. According to this, simulating realistic scenarios
(in which the dispersed nature of DG is taken into account)
is a very good basis for studying the distribution network
potential in terms of HC increase in relation to SG technology implementation.

Solutions for Increasing
Grid Hosting Capacity
In a first attempt to compare solutions that were installed
in different countries and operating conditions, it was clear
that centralized or even supervised (centrally monitored but
not real-time controlled) solutions are generally more effective (in terms of HC increase, see Table 1) than decentralized
solutions. The best performing solutions usually include a
coordinated use of an on-load tap changer (OLTC) and other
means, such as static synchronous compensator (STATCOM),
DG-reactive power control, and field measurements. The use
of either a single solution [STATCOM, OLTC, or automatic
voltage regulator (AVR)] or several individual elements working
locally produces inferior results.
Since the simulation provides probability functions of
HC, additional information can be extracted from the results.
Table 1 reports, for each solution investigated, how the HC
increase is influenced by the position of the DRES units on
the network (standard deviation of the probability function).
may/june 2017

Table of Contents for the Digital Edition of IEEE Power & Energy Magazine - May/June 2017