OpenModelica/Pandapower
Power Flow and Grid Model
In this assignment Open Modelica and Pandapower simulation tools where utilised to compute the static power flow
of a four bus network as the one given in Figure 1.1. Power flow computation algorithms are commonly based
on iteratively solving Equation 1.1 and 1.2, for respectively the active and reactive power flow, for all nodes of the
network. Here, N represents the number of nodes in the network and the subscripts represent parameters of any
node n or parameters from node i in reference to any node n.
∑
N ∑
N
Pi = |Vi Vn Yin | · cos(θin + δn − δi ) (1.1) Qi = |Vi Vn Yin | · sin(θin + δn − δi ) (1.2)
n=1 n=1
It is evident from these equations that the power in one node is a function of the parameters in all other nodes of
the network. For a real network, where a large number of nodes exist, this calculation is computationally expensive.
Thus iterative methods such as the Gauss-Seidel, Newton–Raphson, Decoupled or DC power flow algorithms are
preferred [1]. These algorithms, however, require that the network topology and parameters are specified (i.e. line
admittance and nodes interconnection). Moreover, to fully describe the network electrically, two from the following
four parameters should be specified for each node [2]: voltage phasor magnitude, |V | [pu]; voltage phasor angle, δ
[deg]; injected active power, P [M W ]; injected reactive power, Q [M var]. Depending on the specified parameters
each bus is classified as a slack, generator or load bus. One node where the voltage magnitude and angle are
specified is taken as a reference and named slack node. In the other nodes, the active power and voltage magnitude
(generator bus), or the active and reactive power (load bus) are specified.
Pandapower
Pandapower, supported by Python, is used to perform
power systems’ analysis and optimisation [4]. Figure 1.1
shows the single-line model of the pandapower network
case4gs [3], which consists of four busses - connected
by four lines -, four loads, one generator, and one ex-
ternal grid. The input parameters were as given in Ta-
ble 1.1, with the maximum generation capacity of bus
4 is P4,gen = 318 M W . Using these values, a static
power flow simulation was performed. The results are
presented in Table 1.2, where P = Pload − Pgen − Pgrid
and Q = Qload − Qgen − Qgrid . Since P and Q are neg-
ative for busses 1 and 4, these busses supply power
to the network. On the other hand, the positive sign of Figure 1.1: Single-line diagram of the Pandapower case4gs network [3]
P and Q for busses 2 and 3 shows that they consume
power from the network, equal to their respective load.
Table 1.1: Load power of each bus
Bus Type |V | [pu] δ [deg] Pload [M W ] Qload [M var]
1 Slack - Load 1 0 50.00 30.99
2 Load - - 170.00 105.35
3 Load - - 200.00 123.94
4 Generator - Load 1.02 - 80.00 49.58
The load’s active power of bus 4 was then increased from 80 M W to 250 M W . However, bus 4 cannot generate
more power than it did before since in the first simulation this bus was already operating at its maximum generation
capacity. Consequently, bus 4 now supplies less active power to the network. Bus 1 compensates for this, injecting
1
more active power into the system. Besides, increasing P4,load also impacts the reactive power at busses 1 and 4,
which gets redistributed among both. The results of this analysis are presented in Table 1.3.
Table 1.2: Static system analysis when P4,load = 80M W Table 1.3: Static system analysis when P4,load = 250M W
Bus |V | [pu] δ [deg] P [M W ] Q [M var] Bus |V | [pu] δ [deg] P [M W ] Q [M var]
1 1.0000 0.0000 −136.81 −83.51 1 1.0000 0.0000 −307.88 −58.65
2 0.9824 −0.9761 170.00 105.35 2 0.9821 −3.7613 170.00 105.35
3 0.9690 −1.8722 200.00 123.94 3 0.9687 −3.6666 200.00 123.94
4 1.0200 1.5231 −238.00 −131.85 4 1.0200 −3.2753 −68.00 −162.10
OpenModelica
OpenModelica is an open source software for modelling
and simulation. In this report, it was used for the purpose
of dynamic simulation of a power system with four busses.
Three simulations were conducted to understand the effect
of changing input parameters. For a step height of 0.3, the
load of bus 3 increases by 0.3M W after t = 5s by means
of a step function; see Figure 1.2. The active power of
busses 1, 2, and 4 decreases to compensate for bus 3
being stepped up. From Figure 1.3, it is noted that at a
step height of −0.3M W , the active power sees a decrease
of 0.3M W at t = 5s. This decrease is accompanied with
an increase of active power at busses 1,2, and 4. Finally,
Figure 1.4 shows the result of using a ramp function. The
height was set at 0.2, with a duration of 3s. The start of the
ramp was set at t = 5s. As illustrated, the ramping occurs
from t = 5s to t = 8s. This is accompanied with a ramp
Figure 1.2: Step function used to increase the power at down of power at busses 1, 2, and 4 to maintain power
Bus 3 balance since the power at bus 3 decreases.
Figure 1.3: Step function used to decrease the power at Bus 3 Figure 1.4: Ramp Function Used to increase power at Bus 3
Bibliography
[1] P. Schavemaker and L. Van der Sluis, Electrical power system essentials. John Wiley & Sons, 2017.
[2] J. J. Grainger, Power system analysis. McGraw-Hill, 1999.
[3] pandapower, “Case 4gs,” 2023. [Online]. Available: https://pandapower.readthedocs.io/en/latest/networks/power_system_
test_cases.html#case-4gs
[4] pandapower, “pandapower,” 2023. [Online]. Available: https://pandapower.readthedocs.io/en/latest/index.html#
