Management of a PEM Electrolyzer in Hybrid Renewable Energy Systems - Fuzzy Modeling and Control: Theory and Applications - page 212

to this converter a control signal bymeans of an analogue voltage output. This voltage
level is obtained from the FLC output.
From this point, the fuzzy controller developed for the electrolyzer operation is
described in detail. The designed FLC has six inputs and one output. The FLC input
variables are: State Of Charge (SOC) of the battery, solar irradiance, temperature
of PV panel, compromise current, pressure of metal-hydride system and difference
between compromise and electrolyzer currents. The latter one is considered as error
signal because it represents the deviation between the surplus available current and
the one delivered to the electrolyzer. This error signal is calculated in Simulink before
entering the FLC block.
Compromise current is defined by the authors as the possible surplus current that
would be produced in the installation if the PVgenerator was providing themaximum
possible current. It plays the role of threshold to decide if the energy surplus is enough
for the electrolyzer operation. It is calculated as the difference between the maximum
current from modules, Ipmax , and the load current. Ipmax depends on the panels'
voltage, Vpan , and the irradiance, G , according to ( 12.1 ) that has been obtained from
experimental data:

Ipmax = 2 ∗ G ∗ ( 0 . 0049 − 0 . 0002 ∗ Vpan ) (12.1)

The battery SOC is estimated in the PLC with the Ampere-counting method
(González et al. 2012 ) from values of current and capacity of the battery. Incident
irradiance in the PV modules plane is used. It is averaged each 5min in order to
reduce the transitory fluctuations due to clouds.
The PV panel temperature is measured with a Pt-100 probe on the backsurface.
This variable is included in the control process because the generation capacity
and the performance of the modules depend on their temperature. The lower the
temperature, the higher the performance.
Once the battery is charged enough and the load demand is being satisfied, the PV
modules provide a surplus current used for hydrogen generation. These conditions
are evaluated by means of the incident irradiance, the compromise current and the
battery SOC. Furthermore, technological factors must be taken into account such as
the deactivation of the fuel cell and the available capacity for storing in the metal
hydride bottles, i.e., their pressure has to be under the maximum level.
When such conditions are fulfilled, the voltage control signal generated by the
FLC is applied to the DC-DC converter that feeds the electrolyzer from PV modules.
This voltage varies with meteorological and technological changes according to the
rules defined for the controller. So, the current drawn by the electrolyzer and, hence,
the flow of hydrogen produced are both adapted to the availability of energy.
The structure of the FLC has beenmade as simple as possible. The fuzzy controller
is of Mamdani type, the And method is Min , the implication operator is Min ,the
Aggregation is Max and the defuzzification strategy is the Centroid of area. The
membership functions and the fuzzy rules have been defined based on the experience