Clean Energy Science and Technology 2025, 3(1), 296.

https://doi.org/10.18686/cest296

Article

Impact of self-consumption and regulatory approaches on the profitability

of a grid-connected PV setup plant for a net-zero-emission villa in Morocco
Amin Bennouna

Faculty of Science Semlalia Marrakech, Cadi Ayyad University, Marrakesh 40000, Morocco; [email protected]

CITATION Abstract: For a bioclimatic villa located in Marrakech, Morocco, electricity and bills were
Bennouna A. Impact of self- monitored for several years in this study. Energy was aggregated into days and months to be
consumption and regulatory reconciled with monthly bills. Demand-side management, namely shifting electricity
approaches on the profitability of a consumption to the daytime, improved the investment payback time to acceptable levels.
grid-connected PV setup plant for a
net-zero-emission villa in Morocco.
Results are used (i) to analyze financial sensitivity to self-consumption and (ii) to compare the
Clean Energy Science and profitability of the actual situation of the villa in self-consumption without redemption of
Technology. 2025; 3(1): 296. surpluses with other possible regulatory approaches (surplus sales and net-metering).
https://doi.org/10.18686/cest296
Keywords: net-zero-emission house; grid-connected photovoltaic plant without storage;
ARTICLE INFO demand-side management; technical monitoring and assessment; financial monitoring and
Received: 16 December 2024 assessment
Accepted: 25 January 2025
Available online: 17 February 2025
1. Introduction
COPYRIGHT
Without energy storage, the difficulty in making a domestic solar photovoltaic
(PV) system profitable lies in the usually poor synchronization of electricity demand
Copyright © 2025 by author(s). with daytime solar electricity production. An hourly load unsuitable for solar can even
Clean Energy Science and cancel out the advantage of being in a location with an excellent solar resource, such
Technology is published by Universe
Scientific Publishing. This work is
as this study’s site in Marrakech, Morocco. For a villa located in the suburbs of
licensed under the Creative Marrakech, its goal was, as written in the first brochures of the villa at the time, to
Commons Attribution (CC BY) offset all greenhouse gas emissions by solar compensation of energy consumption
license.
https://creativecommons.org/licenses/
(electricity and some butane gas). The villa is bioclimatic since its construction and
by/4.0/ has become an energy research laboratory on buildings for researchers from Cadi
Ayyad University, Marrakech [1–4].
Obviously, electric energy storage [5] or thermal energy storage [6] is the best
solution to increase the part of solar energy locally consumed by any photovoltaic
plant (self-consumption) but, at present, with high battery prices in Morocco due to
import duties, an additional investment for storage is not likely to improve the
profitability of installation. Because of this, the author decided to avoid storage in the
present study in order to adapt the capacity to net-zero conditions and to explore the
path of energy demand-side management to improve the profitability of the setup. The
energy demand-side management was enabled by the fact that the villa has a large
garden (4000 m2) requiring irrigation and a swimming pool requiring water filtration,
both of which can be programmed to be done during the daytime.
In Morocco, the initial version of the 2010 Law 13/09 on renewable energies did
not allow the injection of electricity into the low-voltage network, but the 2016
Amendment 58/15 lifted this ban. However, a buy-back tariff for surpluses has not
been defined yet by the National Electricity Regulatory Authority, created in 2016 by
Law 48/15, even though it was supposed to. In the meantime, surpluses that are

1
 Clean Energy Science and Technology 2025, 3(1), 296.

injected into the network are therefore lost, to the benefit of electricity distributors.
The objective of the study was therefore to make as profitable as possible a solar PV
system that is installed without storage, despite this unfavorable legal context forcing
the injection of surpluses at a loss when operating a net-zero-emission house.

2. Photovoltaic system sizing

2.1. System sizing according to existing energy consumption and PV yield

before the project
The aim for the villa since the beginning was to make a net-zero-emission home,
and hence manual consumption readings were performed to size the PV system and
not just to check the bills.
2.1.1. Rated PV power required to compensate greenhouse gas emission from
grid electricity
Figure 1 shows the following:
• The cumulative feed from the grid (kWh) is in blue, as taken periodically from
two installed utility meters for 14 years.
• The two different slopes are line adjustment for periods before and after the end
of 2019 following the PV system’s installation.
The graph in Figure 1 clearly shows the different trends between the periods
before and after the installation of the PV system (end of 2019). It can already be
anticipated that nearly 52% energy savings ((18,706−9088) kWh/18,706 kWh) during
the period from 2020 to 2024 with respect to the period from 2011 to 2019. For the
period before 2019, the adjustment of cumulative consumption by a straight line
(presented as a dotted blue line) shows an average slope of 18,706 kWh/day, or 6832
kWh/year (i.e., 569 kWh/month).

Figure 1. Cumulative electricity consumption before and after installation of PV

system (end of 2019).

Considering a 95%-efficiency inverter fed by 30°-tilt polycrystalline silicon PV

modules mounted facing nearly south (the building faces 17°towards the east), the

2
 Clean Energy Science and Technology 2025, 3(1), 296.

long-term yearly averages of PV yields (YAC in kWh/Wp.day) from three different

sources were obtained:
• From SolarMed Atlas [7]: YAC (SMD) = 1722 Wh/Wp.year.
• From the inverter manufacturer (SMA Solar Technology AG), as collected from
its Sunny Portal website [8]: YAC (SMA) = 1758 Wh/Wp.year.
• From a calculation following a previous work [9]: YAC (PRP) = 1736
Wh/Wp.year
Comparing the three datasets with that of this project is somewhat outside the
scope of the study. However, it can be shown that before this work, at the end of 2019,
the three available data showed an average of YAC = 1739 Wh/Wp.year. The definition
of yearly AC photovoltaic yield links the solar production of AC electricity, Esp, to the
rated power, PR:
YAC = Esp/PR (1)
Hence, the 6832 kWh/year found above would require a theoretical rated power
of PR1 = 3929 Wp of PV modules.
2.1.2. Rated PV power required to compensate greenhouse gas emissions from
butane gas
Electricity carbon intensity (CI) is the mass of greenhouse gases generated per
kWh of supplied electricity:
CI = m(CO2)/Esp (2)
hence,

PR2 = m(CO2)/(YAC × CI) (3)

Due to the dominance of coal-fired electricity in the Moroccan electricity mix, if

all line losses are taken into account, each kWh saved at the exit of the National Office
of Electricity and Drinking Water’s low-voltage network prevented a CI of 820
gCO2eq of greenhouse gas emissions during the year preceding the installation of the
photovoltaic system. In addition, the butane emission factor is around 3033 gCO2eq
per butane kg [10], which means that, to offset the m(CO2) = 218 kgCO2eq emissions
from the annual consumption of six cylinders of butane gas of 12 kg, 266 kWh of
renewable electricity must be yearly injected into the low-voltage output of the
electricity network, which would require a theoretical PR2 = 153 Wp of PV modules.
2.1.3. Final photovoltaic system sizing to compensate emissions from both
electricity and butane
The total theoretical nominal power to be installed of PR = PR1 + PR2 = 4082 Wp
was rounded to 4400 Wp for the installation of 16 Jinko Solar modules rated at 275
Wp each. Two rings of 8 PV modules feed two inputs for a 5kW SMA inverter. With
a yield of 1739 Wh/Wp.year, these were expected to generate 7652 kWh/year of solar
AC electricity, i.e., more than the 6985 kWh of power needed for net-zero emissions.
In 2018, the 6832kWh annual electricity requirement would have generated 5602
kgCO2eq/year. When added to the 218 kgCO2eq from the butane gas emission, this
gave a total of 5820 kgCO2eq/year. Therefore, the elimination of 6275 kgCO2eq of
emissions by the 7652 kWh/year of planned solar electricity consumption should

3
 Clean Energy Science and Technology 2025, 3(1), 296.

achieve the objective of net-zero emissions, while even offering a margin for a
possible increase in energy consumption.
2.1.4. Technical details of installation
Figure 2 shows how the photovoltaic modules act as the input of the synchronous
inverter, which itself feeds the internal electric grid of the villa. An SMA smart meter
was placed between the villa’s general connection box and the distributor’s meter to
measure the energy coming from the electricity network. In Figure 2, the solid-line
arrows show the possible directions of electricity flows, while the dotted lines show
the information network connections. The smart meter not only separately measures
the input or output of electricity but also collects inverter data and provides commands
to reduce solar electricity production to a given level when the option is enabled.

Manual
Internet
survey

16 PV 5kW Smart Utility

modules Inverter
H meter meter
Figure 2. Simplified diagram of the installation, its measuring instruments and information system.

2.2. Data acquisition system

In the data acquisition system, the smart meter collects the measured solar
electricity production (Esp) from the inverter and measures both the feed from the grid
(Efg) and the surplus injection into the grid (Eig) from its included bidirectional meter.
Electric power is measured in 5s increments before being averaged and stored with a
15-min time interval until the end of March 2024, after which it was lowered to a 5-
min interval [11]. All average powers are subsequently integrated to obtain daily,
monthly or yearly energy yields.
To ensure the best possible coincidence with the energy readings of the
distributor’s energy meter located on the street, which were taken as a reference, the
energy data were all adjusted by a multiplicative factor of 99.935%, as obtained via
the least squares method of 55 monthly manual readings of the feed from the grid on
the distributor’s energy meter by counting more than 15000 cumulative kWh.
In Figure 3, each one of the basic data items is characterized by four different
datasets: its designation and symbol followed by its key performance index and the
economic revenue that it could possibly generate.

4
 Clean Energy Science and Technology 2025, 3(1), 296.

Basic system data

Technical data itself
Its related Key Performance Index
Its related economic data

Total demand Solar production R

Etd = Efg + Efs ↓ Esp is measured ↓
Share of daytime demand H PV yield
Monthly bills without P.V. (Levelized Cost of Electricity)

Feed from grid Feed from solar R Injection into grid R

Efg is measured ↓ Efs = Esp - Eig ↓ Eig is measured ↓
(no useful KPI to calculate) H Self consumption = E fs /E td H (no useful KPI to calculate)
Monthly bills with P.V. Savings on monthly bills (Earnings if injection sold)

Figure 3. System basic data, their related key performance index and economic data (arrows show energy flow
directions).

The feed from the solar PV system, Efs, is obtained by subtracting the injection
into the grid, Eig, from solar electricity production, Esp. The total demand, Etd, is
obtained by adding the feeds from the grid, Efg, and the solar PV system, Efs.
The feed from the solar PV system (Efs) appears as a “bridge” shared by the total
demand (Etd) and the local solar electricity production (Esp), and the combination of
Efs (yellow box) results in:
• The total demand, Etd, (orange box) when combined with the feed from the grid
(Efg (red box).
• Solar electricity production, Esp, (green box) when combined with the injection
into the grid, Efg (blue box).
Through careful daily collection of data between 22 October, 2019, and 30
November, 2024, a database was created consisting of approximately 200,000
instantaneous average power data points, which were subsequently integrated for each
of the 1869 days and then aggregated into monthly and annual energy figures.
However, for unidentified reasons, some data from the smart meter were unable to be
collected for a total of 59 days, namely, 18 days between 17 November and 5
December, 2021, where all data were missing, and 41 days between 1 April and 11
May, 2024, where only the demand data were missing. The missing data were replaced
with those of equivalent days from another year. Missing data due to electric grid
failure, however, were not replaced in order to obtain a realistic dataset.

2.3. Characteristics of actual or hypothetical phases studied

From the beginning of April 2011 to 16 February 2020 (3241 days), the electricity
supply contract for the villa was prepaid. This contract required the use of a digital
meter (called “Noor”), which counts the absolute value of the power passing through
it for billing the electricity injected into the network. This meter (labeled as Meter A)
prevents the construction of a net-zero-energy villa. Hence, for the solar PV system’s
installation, an inverter was required in order to put into a mode of restriction of
electricity injection into the network. A total of 3241 days of this supply contract were
subdivided into three unequal phases:

5
 Clean Energy Science and Technology 2025, 3(1), 296.

• A long preliminary phase (3113 days), named Phase A1, made it possible to
define the existing needs used for sizing the system. In addition, Phase A1 served
as a reference to measure the savings gained from the installation of the PV
system.
• A very short subsequent phase (14 days), named Phase A2, was used to
approximately estimate, with the new digital metering, what happened when a
nighttime irrigation system was programmed following the solar PV system’s
installation.
• A short phase (114 days), named Phase A3, allowed the understanding of the
impact of switching the programming of the irrigation system to a daytime
schedule during the waiting time for the distributor to change the supply contract
and the meter in order for the project to achieve the net-zero-emission target
during Phase B.
During Phase B, electricity injection into the grid was ignored by the newly
installed analog meter because of the meter’s anti-return wheel. Phase B lasted for
1746 days at the writing of this paper.
Referring to Phase A1 (without the installation of the solar PV system), five
scenarios were constructed for the financial study:
• Three scenarios were constructed on the basis of a 20-year continuation of Phases
A2, A3 and B.
• A hypothetical fourth scenario was built on the basis of a 20-year continuation of
the energy data from Phase B but with a future purchase of surplus energy at 0.45
Dh/kWh (1 Dh ≈ 0.1 US$).
• Another hypothetical fifth hypothetical scenario used the same hypotheses but
with a net-metering approach.
Table 1 summarizes all the conditions for each phase described above.

Table 1. Different phases and hypothetical cases treated in this study.

Supply Surplus sales PV Injection Duration
Counter Phase Irrigation Period Date started Date ended
contract (Dh**/kWh) (kW) into grid (days)
A1 Night 0 No PV 01/04/11 09/10/19 3113
“Noor” Prepayment 0.00 A2 Night 4.40 Restricted 10/10/19 24/10/19 14
A3 Day 4.40 Restricted 24/10/19 15/02/20 114
Analog Post-payment 0.00 B Day 4.40 Allowed 16/02/20 27/11/24 1746
Surplus sales* 0.45 C Day 4.40 Allowed With future sales of surplus
Smart
Net-metering* In kind D Day 4.40 Allowed Hypothetical case
* Hypothetical cases; ** 1 Dh = 1 MAD = 1 Moroccan Dirham ≈ 0.1 US$.

Obviously, the injection of electricity into the grid was restricted during Phases
A2 and A3 to avoid having to pay for it.
For the hypothetical case of surplus sales, which still does not yet exist in
Morocco (Phase C), the price was set up at the rounded value of 0.45 Dh, just 0.10
above the Levelized Cost of Energy at 0.35 Dh for an investment of 40,000 Dh, a
2%/year depreciation and a −0.7%/year production loss. Just for reference, distributors
buy daytime electricity at 0.85 Dh/kWh.

6
 Clean Energy Science and Technology 2025, 3(1), 296.

2.4. Details of demand-side management (for phases A3 and B)

For a short period in 2019, preliminary approximate sub-metering measurements
made it possible to determine the following segmentation of electricity consumption:
• 37% was used for pumping well water using a 2700W pump supplying
programmable irrigation via eight different solenoid valves, to which eight half-
hour periods of irrigation were assigned.
• 23% was used filtering the swimming pool water using a 1500W pump, whose
operation was programmed at 15-min periods.
• The remaining 40% was used by the house. In order not to change the habits of
its occupants, it was decided not to manage the demand by appliances. This is
why it was already known in 2019 that reaching a 60% self-consumption level
would be difficult.
To allow some form of synchronization with the direct production of solar
electricity (depicted as the green histogram in Figure 4a), the demand for the garden
(irrigation and filtration) programming is shown by the orange histogram in Figure
4a:
• The half-an-hour activation of the eight solenoid valves triggers the drip irrigation
of the garden (vegetable garden, fruit trees, ornamental plants and lawn) via a
well pump with a 2.7kW power demand.
• The swimming pool filtration pump has a 1.5kW power demand.
• The base of the histogram represents the consumption of the house, assumed to
be constant at 0.25 kW. The total daily consumption therefore approaches the
18.7 kWh defined in Figure 1.

3 000 Planned solar production (W), 3 000 3 000 2021-2023 average of 3 000
20.7 kWh/day Solar production [W]
Planned power demand (W), 20753 Wh/d
2 500 18.7 kWh/day 2 500 2 500 2 500
2021-2023 average of
Total demand [W]
2 000 2 000 2 000 19711 Wh/d 2 000

1 500 1 500 1 500 1 500

1 000 1 000 1 000 1 000

500 500 500 500

0 0 0 0
GMT GMT
0:00 6:00 12:00 18:00 00:00 06:00 12:00 18:00

(a) (b)
Figure 4. (a) Predicted load curve of the villa; (b) three-year average of total demand in reality.

The difference between the programming and reality (orange histograms in

Figure 4) is explained by several reasons:
• The simplest explanation is that the nighttime consumption of the house turns out
to be higher than expected.
• Opening the irrigation solenoid valves results in the intermittent operation of the
well pump due to the hysteretic control of the pressure tank–pressure switch
system (which cuts at 3 bars and restarts at 2 bars).
• Some or all of the solenoid valves can be closed due to the plants’ irrigation cycle

7
 Clean Energy Science and Technology 2025, 3(1), 296.

or rain, respectively, when some or all parts of the garden no longer need to be
irrigated.
• Finally, various reasons required the short or prolonged shutdown of the pool
water filtration.
Figure 4b allows the calculation of an average self-consumption of 55%
(11,025/(11,025−8983)) during the period from 2021 to 2024. This same self-
consumption value is found later in Subsection 3.1.4.

2.5. Financial hypotheses: Rates, investment and depreciation

There cannot be a realistic financial study adapted to a country (e.g., Morocco)
without actual electricity prices. The low-voltage ones are listed in Table 2. They have
not been changed since August 2017, which was the end of the latest program contract
between the Moroccan government and the National Office of Electricity and Drinking
Water.

Table 2. Low-voltage electricity fares in Morocco since August 2017.

Supply Monthly Price (Dh) per consumption segment (kWh/m)

Fixed fees (Dh*/m)
contract (kWh/m) ≤ 100 ≤ 150 ≤ 210 ≤ 300 ≤ 500 ∞
Prepayment 0.00 >0 1.5407
≤ 100 0.9010 NA
≤ 150 0.9010 1.0732 NA
≤ 210 1.0732 NA
Postpayment 62.65 for 4kW subscription
≤ 300 1.1676 NA
≤ 500 1.3817 NA
> 500 1.5958
* 1 Dh/m = 1 MAD/month = 1 Moroccan Dirham/month ≈ 0.1 US$/month.

Table 2 shows that the prepayment contract (using Meter A) is simple, which is
without fixed fees and with a flat 1.5407 Dh/kWh for a subscribed power of 6 kW. For
postpayment (Phase B and using Meter B), the fixed monthly fee was set at 62.65
Dh/month for the villa’s 4kW subscription. Social and environmental reasons are
behind the tiered pricing system, which is as follows:
• Below 150 kWh/month, customers pay:
• 0.9010 Dh/kWh for the first 100 kWh.
• 1.0752 Dh/kWh for the following 50 kWh.
• Higher consumptions are treated as follows:
• For more than 500 kWh/month: 1.5958 Dh/kWh.
• For consumptions below or equal to 500 kWh/month: 1.3817 Dh/kWh from
0 to 500 kWh.
• For consumptions below or equal to 300 kWh/month: 1.1676 Dh/kWh from
0 to 300 kWh.
• For consumptions below or equal to 210 kWh/month: 1.0732 Dh/kWh from
0 to 210 kWh.
Phase B rates also allow appreciating the billing non-linearity resulting from the
non-flat kWh price.

8
 Clean Energy Science and Technology 2025, 3(1), 296.

For each of the five scenarios, all financial simulations (payback time and net
present value) considered the following:
• An investment of 40,000 Dh (rounded) for the PV plant.
• A project duration of 20 years and a discount rate of 2%/year (tax-free popular
savings for deposits at the Morocco Postal Bank).

3. Results and discussion

3.1. Phase B: Results of technical monitoring and assessment

Sections 3.1 and 3.2 will only refer to Phase B not only because it is the longest
phase and the present phase but also because, as already said, Phase A and its numeric
utility counter charges for electricity injection and makes the net-zero-emission project
non-profitable.
3.1.1. Phase B: Effective achievement of net-zero-emission goal
At the beginning of 2020 and at the end of 2024, the data for the days of a few
weeks missing during Phase B were supplemented with average days of the year in
order to complete the data for all years. Figure 5 shows the evolution of the yearly
technical data of the villa, which revealed the following:
• A total demand of 7304 kWh/year, as fed from the solar PV system (4030
kWh/year) and from the grid (3284 kWh/year).
• Solar electricity production of 7714 kWh/year, among which 3691 kWh/year
were injected into the grid.

9000 Total demand, 7304 kWh/y Solar production, 7714 kWh/y

Feed from solar, 4030 kWh/y Injection into grid, 3691 kWh/y
Feed from grid, 3284 kWh/y Avoided GHG emissions, 6516 kg/y
8000

7000

6000

5000
8 101
7 943
7 927
7 853
7 725
7 709

7 500

4000
7 196
7 053

6 903
6 796

6 672

6 467

6 442
6 098

3000
4 471

4 371
4 309

3 937
3 781

3 767

3 760
3 739

3 571
3 457

3 457

2000
3 401

3 386
3 059
2 744

1000

0
2020 2021 2022 2023 2024

Figure 5. Evolution of yearly cumulative technical data (2020–2024).

As there is still no GHG emissions data available from the Ministry of Energy
after 2021 [12], the carbon intensity of Moroccan electricity delivered in low voltage
was calculated from (i) emission intensities of electricity from fossil sources [13] and
(ii) electricity production data in [14], which also made it possible for calculating the
efficiency of the network. The calculations found that the solar electricity production
offset an average of 6515 kg of CO2 equivalent per year. From Figure 5, it can be

9
 Clean Energy Science and Technology 2025, 3(1), 296.

confirmed that, with more than 6275 kg avoided, the goal of net-zero emissions has
been achieved.
3.1.2. Phase B: Monthly total demand and its fluctuations

20 1000
58 months of Total demand Feed from grid (kWh/m), Decomposition of
(616kWh/m, σ=151) 900 3284 kWh/y TOTAL DEMAND
Number of months

15 800 Feed from solar (kWh/m), monthly average

4030 kWh/y

503
700
10 600

290
257
17

15

254
218
500

272

322
300

208

284
11

214
10

5 400

202
300

445
2

404
394
0

365
0

351
200

337

324

296
293

286
279
[100, 200[

[200, 300[

[300, 400[

[400, 500[

[500, 600[

[600, 700[

[700, 800[

[800, 900[

[900, 1000[

[1000, 1100[

[1100, 1200[

230
100
0

FEB

MAY

SEP
JAN

MAR

NOV

DEC
APR

JUL
JUN

AUG

OCT
Monthly total demand range (kWh/m)

(a) (b)
Figure 6. (a) Distribution of monthly total demands; (b) monthly averages over four full years.

Figure 6 shows that, at 151 kWh/month, the standard deviation of the monthly
total demand for electricity lies around 25% of the average at 616 kWh/month. The
fluctuation is driven by seasonal factors, such as the following:
• In February, irrigation is minimal and there is no need for heating the bioclimatic
house.
• In July, an excess of consumption is caused by air conditioning from three earth-
air heat exchangers [1,2]. The house is usually empty in August.
3.1.3. Phase B: Monthly solar electricity production and its fluctuations

40 1000
58 months of Solar production Injection into grid (kWh/m), Decomposition of
35 900 3691 kWh/y
(634kWh/m, σ=57) SOLAR PRODUCTION
Number of months

30 800 Feed from solar (kWh/m), monthly average

4030 kWh/y
25 700
20 600
204

288

255
328
34

343

269

15 500
375
269

206
253

231
310

10 400
14

5 300
9

445

404
394
0

365

0
351

200
337

324

296
293

286
279
[100, 200[

[200, 300[

[300, 400[

[400, 500[

[500, 600[

[600, 700[

[700, 800[

[800, 900[

[900, 1000[

[1000, 1100[

[1100, 1200[

230

100
0
FEB

MAY

SEP
JAN

MAR

NOV

DEC
APR

JUL
JUN

AUG

OCT

Monthly Solar production range (kWh/m)

(a) (b)
Figure 7. (a) Distribution of monthly solar electricity productions; (b) monthly averages over four full years.

Figure 7 shows that, at 57 kWh/month, the standard deviation of solar electricity

production lies around 9% of the average at 634 kWh/month. The fluctuation is driven
by the main seasonal extremes:
• The maxima for months near the equinoxes, for which the 30°tilt angle is optimal.
• The minima during the five months surrounding the winter solstice between
October and February.

10
 Clean Energy Science and Technology 2025, 3(1), 296.

3.1.4. Phase B: Monthly self-consumption and its fluctuations

Self-consumption is given by the ratio of the feed from the solar PV system to
the total demand, as given in Figure 3. Figure 8 shows that, at 6.5%, the standard
deviation of the self-consumption lies around 12% of the average at 55%. The self-
consumption fluctuation is driven by the main seasonal extremes:
• The maxima where there is more solar electricity.
• The minima for the month of July, when total demand is exceptionally high.

35 70%
58 months of Self consumption Self consumption, 55% yearly
30 (55%, σ=6,5%) 60%
Number of months

25
50%
20
40%
30

15

62%
61%

61%
59%

58%
57%
10 30%

54%
53%
49%
15

48%
48%
13

44%
5
20%
0

0
[30%, 40%[

[40%, 50%[

[50%, 60%[

[60%, 70%[

[70%, 80%[

[80%, 90%[

[90%, 100%[

10%

0%

FEB

MAY

SEP
JAN

MAR

NOV

DEC
APR

JUL
JUN

AUG

OCT
Monthly self consumption range (kWh/m)

(a) (b)
Figure 8. (a) Distribution of self-consumptions; (b) monthly averages over four full years.

3.1.5. Phase B: Monthly daytime total demand share and its fluctuations
From sunrise to sunset each day, the total daytime demand (in kWh/day) was
calculated and divided by the total 24-h demand (in kWh/day) to obtain the total
daytime demand share. This was done because, obviously, the increase in daytime
consumption improves the rate of self-consumption, which in turn has a favorable
impact on the financial performance of the investment.

40 90%
58 months of Daytime demand Daytime demand share, 74% yearly
35 share (73%, σ=6%) 80%
Number of months

30 70%
25
60%
20
35

50%
15
80%

80%
79%
77%

76%
74%

40%
73%

73%
68%

10
67%
65%

59%
12

30%
10

5
0

0 20%
[30%, 40%[

[40%, 50%[

[50%, 60%[

[60%, 70%[

[70%, 80%[

[80%, 90%[

[90%, 100%[

10%

0%
FEB

MAY

SEP
JAN

MAR

NOV

DEC
APR

JUL
JUN

AUG

OCT

Monthly share of daytime demand range (kWh/day)

(a) (b)
Figure 9. (a) Distribution of monthly daytime demand shares; (b) monthly averages over four full years.

Figure 9 shows that, at 6%, the standard deviation of the total daytime demand
share is about 8% of the average at 73%. The self-consumption fluctuation is driven
by the same aforementioned periods.

11
 Clean Energy Science and Technology 2025, 3(1), 296.

3.1.6. Phase B: Self-consumption rate—Nexus of total daytime demand share

The share of daytime electricity demand plays an essential role in improving the
self-consumption rate, which is a key factor for the profitability of a photovoltaic
system for self-consumption and even more so for those that do not use storage. Figure
10 shows the correlation between monthly self-consumption and the related share of
daytime total demand. The bisector, separated by the red line, indicates that self-
consumption is lower than the daytime total demand share simply because it is only a
part of the total demand.
Figure 10 suggests that, for cases such as this villa, monthly self-consumption
reaches 75% of the monthly daytime total demand.

100%
Self consumption, 55% yearly
Straight line adjustment
80%

60%

40%
y = 0,7511x
R2 = 0,595
20%

0%
0% 20% 40% 60% 80% 100%
Monthly share of daytime total demand

Figure 10. Self-consumption rate against share of daytime total demand.

3.2. Phase B: Results of financial monitoring and assessment

3.2.1. Phase B: Monthly savings and fluctuations

15 900
58 months of Monthly savings Monthly savings (Dh/m)
(610Dh/m, σ=174) 800
Number of months

700
10
600

500
13

13

821

5
772

400
9

712
8

685
605
6

300
524
500

497
483
477
467
3

3
2

385
1
0

0 200
[, 100[

[100, 200[

[200, 300[

[300, 400[

[400, 500[

[500, 600[

[600, 700[

[700, 800[

[800, 900[

[900, 1000[

[1000, 1100[

100

0
FEB

MAY

SEP
JAN

MAR

NOV

DEC
APR

JUL
JUN

AUG

OCT

Monthly Solar production range (kWh/m)

(a) (b)
Figure 11. (a) Distribution of monthly savings; (b) monthly averages over four full years.

Figure 11 shows that, at 174 Dh/month, the standard deviation of the monthly
savings is about 29% of the average at 610 kWh/month. The fluctuation is driven by
seasonal factors, as follows:
• The five “summer” months (May to September).

12
 Clean Energy Science and Technology 2025, 3(1), 296.

• The remaining “winter” months.

3.2.2. Phase B: Evolution of cumulative savings towards payback time

40 000
Cumulative savings (Dh)
35 000
Straight line adjusment
30 000

25 000

20 000
y = 19,633x - 860860
15 000
R2 = 0,9989
10 000

5 000

0
31/12/19 30/12/20 30/12/21 30/12/22 31/12/23 30/12/24

Figure 12. Evolution of cumulative savings towards investment value (40,000 Dh).

Figure 12 shows the growth, at the end of each month during Phase B, of
cumulative savings and their evolution towards the payback date (31 August 2025),
which should occur 67 months after operation (almost 5.6 years). This figure will serve
as a reference to keep a critical eye on future values obtained by the extrapolation of
monthly averages and not by using the real values of each month.
3.2.3. Phase B: Payback time and net present value versus self-consumption
Keeping the investment at 40,000 Dh and savings of 610 Dh per month due to
the PV system, as obtained from Figure 11a, and with fares from Table 2 for the self-
consumption rate to reach 100%, Figure 13 shows the following:
• 20-year net present values (NPVs) at a 2% yearly depreciation rate (green
histogram and left scale).
• Payback time (red line and right scale).

160 000 160

Investment 20 years NPV @ 2%/year (Dh)
134

140 000 Payback time (months) 140

117

Polyomial adjustment
Payback time (months)

120 000 120

104
94

100 000 100

86
NPV (Dh)

78

80 000 80
65
61
58
55
51
49

60 000 60
47
45
44
42
26 708

41
18 471

40 000 40
101 534
111 870
118 268
124 665
134 142
139 513
144 884
150 255
34 944
43 181
51 417
59 654
80 653
87 613
94 573

20 000 20

0 0
20%
25%
30%
35%
40%
45%
50%
55%
60%
65%
70%
75%
80%
85%
90%
95%
100%

Figure 13. Sensitivity of NPV (left scale) and payback time (right scale) to self-
consumption in Phase B.

13
 Clean Energy Science and Technology 2025, 3(1), 296.

The 61 months of payback time obtained at 55% of self-consumption is not the

same as the 67 months obtained in Figure 12, where the non-linearity of the bills was
accounted for, while the calculations for Figure 13 did not.
This graph is valid for almost every country, provided that all financial
hypotheses are fulfilled (investment, depreciation, and electric energy prices), which
tacitly means it is valid for Morocco. In this case, the payback time remains
transposable as it is but the NPVs must be adapted to the installed power (4.4 kW).

3.3. Long-term financial projections of all phases (A1, A2, A3, B, C and
D)
3.3.1. Hypotheses for simulations of long-term financial projection
For the simulations of the long-term financial projection of each phase, monthly
averages were used and assumed to be maintained beyond their term for 20 years.
Figure 14 shows the different hypotheses that were considered in the simulations
of the long-term financial projections of the four real-world phases (A1, A2, A3 and
B) and the two hypothetical situations (Phase C for the sale of surpluses and Phase D
for using the net-metering approach):
• Figure 14a shows the monthly averages of total demands, broken down into the
following:
• The feed from the grid (red);
• The feed from solar electricity production (yellow);
• The amount of solar electricity production injected into the grid (blue).
• Figure 14b shows the resulting monthly averages of the following:
• Self-consumption (yellow histogram and left scale), as deduced from
Figure 14a.
• Net savings (black line and right scale), which included the sales for Phase
C and calculated using the energy data from Figure 14a and fares from
Table 2.

Monthly average of injection 1 041

1 000 into grid (kWh/m) 100% 1 000
Monthly average of self consumption
792

Monthly average of feed from Monthly average of net savings (Dh/m)

800 solar (kWh/m) 80% 800
299

299

658

Monthly average of feed from

600 grid (kWh/m) 16
60% 600
436
339

339
283

13
400 40% 400
99

613
569

152

200 20% 200

295
291

275

275

100%
25%

49%

55%

55%

0 0% 0
"Noor", A1

"Noor", A2

"Noor", A3

@0,45Dh, C

metering, D
Standard, B
"Noor", A1

"Noor", A2

"Noor", A3

Standard, B

Surplus sales

Net-metering,
@0,45Dh, C

Surplus
sales

Net-
D

(a) (b)
Figure 14. Hypotheses considered in simulations of long-term financial projection for each case: (a) initial monthly
technical hypotheses; (b) monthly savings deduced from each self-consumption.

14
 Clean Energy Science and Technology 2025, 3(1), 296.

3.3.2. Simulations of long-term financial projection of all six cases (A1, A2, A3,
B, C and D)
Based on the hypotheses shown in Figure 14 for each of the four real-world
phases (A1, A2, A3 and B) and the two hypothetical situations (C and D), Figure 15
shows the NPV (green histogram) of the 40,000Dh investment for 20 years, taking
into account a 2% yearly depreciation rate and the payback time (red line).

180 000 270

Investment 20 years NPV @

263
160 000 2%/year (Dh) 240

163 520

Payback time (months)

140 000 Pay back time on 40'000Dh 210
investment (months)
120 000 180

114 889
NPV (Dh)

100 000 150

61 88 623
80 000 120
60 000 92 90

-10 240
40 000 60

45 286

50
20 000 30

38
0 0
-20 000 -30
(Phase A1)

(Phase A2)

(Phase A3)

Net-metering,
Surplus sales
(Phase B)
Standard,

("Phase" C)

("Phase" D)
@0,45Dh,
"Noor",

"Noor",

"Noor",

Figure 15. NPV (left scale) and payback time (right scale) for Phases A1, A2, A3, B,
C and D.

By increasing the daytime consumption rate, the use of demand-side management

on its own allowed the following:
• Making the NPV positive.
• Reducing the return-on-investment time to less than 10 years.
Even if the same self-consumptions are found in Figure 13, the results of the
financial analysis may not be the same because the rates are not the same in
prepayment (A) as in postpayment (B). It is the case, for instance, for the very short
Phase A3, for which Figure 14b shows a self-consumption rate of 49% and Figure
15 shows a 92-month payback, while the payback time for a 50% self-consumption
rate obtained from Figure 13 is 65 months.
Even though the values are near each other, the payback time obtained in Figure
12 (67 months) is not exactly the same as the one obtained here for Phase B (61 months)
because the former takes into account the month-by-month non-linearity of the bills
instead of the average. Despite all this, the general trends in the results remain
consistent with each other.
3.3.3. Hypothetical Phase C (surplus sales): Payback time and NPV versus self-
consumption
In this subsection, as presented in Figure 16, for the hypothetical Phase C, the
same sensitivity of NPV and payback time to self-consumption as in Subsection 3.2.3
(Figure 13) but with two different scenarios:
• A distributor will buy all the energy surpluses from a self-producer (as supposed
above for C).
• The distributor limits its purchases to 20% of the self-producer (as provided for

15
 Clean Energy Science and Technology 2025, 3(1), 296.

in the new Self-Production Law, with which the author fully disagrees).

160 000 160 160 000 160

Investment 20 years NPV @ 2%/year (Dh) Investment 20 years NPV @ 2%/year (Dh)
140 000 Payback time (months), no limited sales 140 140 000 140
Payback time (months), 20% limited

113
Polyomial adjustment sales

Payback time (months)

Payback time (months)

120 000 120 120 000 Polyomial adjustment 120

101
100 000 100 100 000 91 100
NPV (Dh)

NPV (Dh)

83
77
75
72

71
80 000 68 80 80 000 80
65
62

60
59

57
54
52

51
60 000 60 000
51
49

60 60

48
48

46
46
45

45
44
42

42
42

42
41

41
41

41
40 000 40 40 000 40
109 813
114 090
118 368
122 646
130 300
134 015
137 730
144 524
147 213
149 901
152 590

105 092
112 053
122 389
128 787
135 184
144 524
147 213
149 901
152 590
63 726
69 280
74 834
80 388
85 942
91 496

28 990
37 227
45 463
53 700
61 936
70 173
91 172
98 132
20 000 20 20 000 20

0 0 0 0
20%
25%
30%
35%
40%
45%
50%
55%
60%
65%
70%
75%
80%
85%
90%
95%
100%

20%
25%
30%
35%
40%
45%
50%
55%
60%
65%
70%
75%
80%
85%
90%
95%
100%
(a) (b)
Figure 16. Sensitivity of NPV (left scale) and payback time (right scale) to self-consumption in two scenarios of
Phase C: (a) without limit of surplus sales; (b) with limit of surplus sales at 20%.

At 100% self-consumption, the two scenarios match perfectly but, because of the
slower decrease, the following observations are revealed:
• To reach 5 years of payback time, the first scenario, as shown in Figure 16a,
needs only 40%–45% of self-consumption, while the second, as shown in Figure
16b, needs 50%–55% of self-consumption,
• It is just hopeless to have profitable investments with less than 30% of self-
consumption.
In the second scenario, as shown in Figure 16b, a project that does have a zero-
emission goal and strives to avoid system oversizing (excess of investment) has to set
a target of 80% self-consumption (and 20% injection), which very difficult to achieve
in a house without storage and without demand-side management flexibilities.

4. Discussion
All theoretical financial studies were carried out with real-world data and
working hypotheses that are actually valid for Morocco. During Phase B, only 59 days
among 1746 of the data were lost (3.3%), which granted overall reliability to the
study’s technical figures and their calculated fluctuations and averages. Extraction and
comments of the technical data and results were limited to whatever was useful for the
financial analysis.
More than 70 scientific articles on grid-connected PV systems in Morocco were
reviewed and even if all of them collected the same type of data, all were based on
100% self-consumption and, moreover, none made the same technical-economic
calculations that were made here. In addition, the attempt to make a habitat or a
professional building with net-zero emissions was not addressed in these works, much
less making it profitable.
In a future study by the author, the maps of solar PV payback time for different
daytime-demand shares in Morocco will be drawn.

16
 Clean Energy Science and Technology 2025, 3(1), 296.

5. Conclusion
The achievements of the experimental part are as follows:
• Sizing the system with the use of existing yearly AC productivity was sufficient
to obtain results compatible with expectations.
• The daytime programming of irrigation and swimming pool water filtration made
it possible to increase the share of daytime demand from 36% to 74%, which in
turn increased the self-consumption coefficient from 25% to 55%, opening the
possibility of an acceptable payback time between 61 (calculated) and 67 months
(actual) and potential 20-year NPVs bigger than twice the initial investment.
The conclusions from the simulations of the long-term financial projections are
as follows:
• On one hand, the self-consumption sensitivity study gave a minimum payback
time of 41 months, which allows the NPV to be multiplied by 8 when self-
consumption increases from 20% to 100%.
• On the other hand, the simulation of the six cases showed that the sale of
surpluses at 0.45 Dh/kWh should reduce the payback time by around 20% (from
61 to 50 months), while the net-metering approach reduces it to 38 months—very
close to that of the simulated self-consumption at 100%, which gave 41 months.
To achieve buildings, offices or factories with net-zero GHG emissions, it is easy
from the technical point of view to calculate and implement grid-connected PV plants
without storage. However, if their economic performance is excellent under a
hypothetical net-metering regime, under other regimes, profitability strongly depends
on the self-consumption ratio. Designing net-zero-emission buildings with a grid-
connected PV setup alone is shown to be almost impossible when there is no flexibility
to increase daytime consumption.
Despite this, it can be concluded that acceptable profitability is allocated to the
net-zero-emission villa in Morocco by reaching 55% self-consumption through better
demand-side management. However, while this is possible in the villa with its very
large garden, it is very likely that increasing the share of daytime consumption is not
easy to extrapolate for other types of housing. Consequently, making a net-zero-
emission home profitable cannot be done either by refraining from buying back excess
self-produced electricity, as is currently the case, or by limiting the shares bought back,
as provided for in the new Self-Production Law. The next law on self-production
should provide for a repurchase of injected electricity limited to 20% of production
but the price is not yet unavailable.

Conflict of interest: The author declares no conflict of interest.

References
1. Bouhess H, Hamdi H, Benhamou B, et al. Dynamic Simulation of an Earth-to-Air Heat Exchanger Connected to a Villa
Type House in Marrakech. In: Proceedings of BS 2013: 13th Conference of the International Building Performance
Simulation Association; 2013; Chambery, France. pp. 1063-1070.
2. Khabbaz M, Benhamou B, Limam K, et al. Experimental and numerical study of an earth-to-air heat exchanger for air
cooling in a residential building in hot semi-arid climate. Energy and Buildings. 2016; 125: 109-121. doi:
10.1016/j.enbuild.2016.04.071

17
 Clean Energy Science and Technology 2025, 3(1), 296.

3. Lafqir FE, Sobhy I, Benhamou B, et al. Thermal performance of passive techniques integrated to a house and the concept of
passive house in the six climates of Morocco. Science and Technology for the Built Environment. 2020; 26(10): 1490-1508.
doi: 10.1080/23744731.2020.1805983
4. Mahach H, Lafqir FE, Benhamou B, et al. Impact of an Earth-Air Heat Exchanger Combined with Nighttime Ventilation on
Indoor Thermal Comfort of a Residential Building in Hot Semi-Arid Climate. Journal of Solar Energy Engineering. 2022;
145(1). doi: 10.1115/1.4054662
5. Fachrizal R, Munkhammar J. Improved Photovoltaic Self-Consumption in Residential Buildings with Distributed and
Centralized Smart Charging of Electric Vehicles. Energies. 2020; 13(5): 1153. doi: 10.3390/en13051153
6. Draou M, Brakez A, Bennouna A. Techno-economic feasibility assessment of a photovoltaic water heating storage system
for self-consumption improvement purposes. Journal of Energy Storage. 2024; 76: 109545. doi: 10.1016/j.est.2023.109545
7. Global Solar Atlas (formerly SolarMed Atlas). Global solar atlas. Available online: https://globalsolaratlas.info/map
(accessed on 30 November 2024).
8. SMA Solar Technology AG. Sunny Portal. Available online:
https://ennexos.sunnyportal.com/login?next=%2Fanalyze%2Fview-analyze-tool-pro (accessed on 30 November 2024).
9. Aarich N, Erraissi N, Akhsassi M, et al. Photovoltaic DC yield maps for all Morocco validated with ground measurements.
Energy for Sustainable Development. 2018; 47: 158-169. doi: 10.1016/j.esd.2018.10.003
10. Emission factor of butane. Climatiq website. Available online: https://www.climatiq.io/data/emission-factor/74e3d3ba-657e-
48c2-adfd-456898139a77 (accessed on 30 November 2024).
11. Present SMA Sunny Portal powered by ennexOS. Available online: https://ennexos.sunnyportal.com/login (accessed on 30
November 2024).
12. Ministère de l’Energie et de la Transition Durable. Secteur de l’énergie - Chiffres clés. Edition 2022. Version provisoire.
https://www.mem.gov.ma/Lists/Lst_rapports/Attachments/39/Chiffres clés annuels de l'énergie.pdf (accessed on 7 December
2024).
13. Red Eléctrica de España. Metodologí a emisiones CO2 generación España - Emisiones de CO2 asociadas a la generación de
electricidad en España. Page 3. https://api.esios.ree.es/documents/580/download?locale=es (accessed on 7 December 2024).
14. Haut-commissariat au Plan. Bulletins statistiques trimestriels. https://www.hcp.ma/downloads/?tag=Bulletins+statistiques
(accessed on 7 December 2024).

18