Photovoltaic based reverse osmosis desalination systems (PV/RO) present an effective method of water desalination especially in remote areas. The increase of the feed water temperature leads to an amelioration of the plant performances. Photovoltaic Thermal Collector (PV/T) represents an ideal power source as it provides both electric and thermal energies for the reverse osmosis process. Nevertheless, PV/T based RO plants should be controlled in order to solve operation problems related to electrical efficiency, reverse osmosis membrane, produced water and the rejected salts. This paper suggests a fuzzy logic controller for the flow rate of the circulating fluid into the PV/T collectors so as to ameliorate the system performances. The designed controller has improved the PV/T field electrical efficiency and preserved the reverse osmosis membrane which upgrades the system productivity. LABVIEW software is used to simulate the controlled system and validate the effectiveness of the controller.
Electric energy and fresh water supply present the major challenge in many parts of the world [
Reverse osmosis (RO) is considered as the most used desalination process due to its high recovery ratio and low energy consumption [
Many studies show feed water temperature impacts on RO process recovery ratio, the consumed energy per cubic meter and the quantity of salt rejection [
Photovoltaic thermal technology (PV/T) is a hybrid system which combines the functions of solar thermal collector and photovoltaic (PV) panel. As it provides heat production, fluid circulation in the back of PV/T panel improves electrical efficiency of PV cells [
Several researchers studied control of solar powered reverse osmosis systems. Control for reverse osmosis desalination serves to guarantee good quality of the produced water (acceptable salinity) and to protect the RO membrane. In several cases, the control is an energy management optimization which purposes to reduce the consumed energy by the desalination process and to amplify water production. Therefore, techniques such as Dynamic Matrix Programming, PID, Model Predictive Control and Fault Tolerant were applied [
Moreover, the economic aspect is very important especially for renewable energy based systems. In fact, the economic analysis presents a significant clue to valorize the exploitation of renewable energy resources [
In this paper, the FLC is used to control a PV/T based reverse osmosis desalination system. Thermal and electric productions are both controlled in order to optimize the PV/T field productivity. In addition, the used FLC aims to ameliorate the RO unit recovery and also to protect the RO membrane from an overheating. Fuzzy Logic Controller is utilized to control the fluid flow rate in order to improve the system performances. The implemented FLC balances between the improvement of the PV/T field electrical efficiency and the RO unit recovery ratio. The paper will start with a modeling part where each system block is modeled; then the implemented FLC is identified; the final section is reserved to discuss the obtained results concerning the energetic and the economic gains.
The considered plant is shown in
The most important components in the PV/T collector are the PV plate and the solar thermal absorber. PV/T panel modeling is based on heat transfer principle. The sheet and tube PV/T system used in the present study composed of a PV module, a thermal absorber, glazing, working fluid, tube, and insulation (
System modeling is based on the energy balance equations developed by Chow [
1) In all tubes sections, the mass flow rate of the working fluid is invariant.
2) The material properties of all PV/T system components are unchanged.
3) We compute the heat transfer coefficients in real time.
The temperature change of the glass cover (Tg) is defined by [
Mg and Cg are the mass and the specific heat coefficient of the glass cover respectively. T presents the temperature, and the subscripts “g”, “a”, and “p” present the glass cover, the ambient, and the PV plate respectively. G presents the solar radiation (W/m2).
With (A) is the PV/T collector area and [
The absorptance and the transmittance of the glass are defined respectively by:
δg, Rg, and θi are the refractive index, the thickness of the glass cover, and the incident angle respectively.
The convective heat transfer coefficient due to wind is given by:
with ua is the wind velocity.
The heat transfer coefficient between glass cover and PV plate is expressed by:
Θ, ε, and σ present the temperature (˚K), the emissivity, and the Stefan Boltzmann constant respectively.
The radiative heat transfer coefficient between the glass cover and the surroundings is defined by the following equation:
Hollands et al., 1976 determine the Nausselt number by the following equation:
with:
where, the Rayleigh (Ra) number defined as:
βa, g, κa, and νa, are the coefficient of thermal expansion, the gravitational constant, thermal diffusivity, and kinematic viscosity of air respectively.
The temperature change of the PV plate is defined by [
with E is the produced electric power and:
Do and W are the outer diameter of tubes and spacing tubes respectively.
The plate absorptance is calculated by the following relation:
The conductive heat transfer coefficient between the collector and the PV plate is defined by:
with:
hpt presents the conductive heat transfer coefficient between the PV plate and tubes. kad and δad are the thermal conductivity and the adhesive layer thickness, respectively. kp and δp are the thermal conductivity and the PV plate thickness respectively. L is the length of the tube. “τp” and “r” are the transmittance and the reflectance of the glass for diffuse radiation, respectively.
And:
The electrical power produced by the PV/T module is calculated by the following expression:
where “p” is the ratio of the cell area to the collector area also known as the packing factor.
The PV cells efficiency is determined by:
βr and ηr present the temperature coefficient and the reference cell efficiency respectively.
The heat balance on the absorber plate is determined by the following equation:
The conductive heat transfer coefficient between the absorber plate and insulation is defined by:
δ and k are thickness and thermal conductivity of the insulation layer respectively. The subscripts “c”, “t”, and “i” refer to the absorber plate, tubes, and the insulation, respectively.
The heat transfer coefficient between tubes and the absorber plate is defined by:
With:
The energy balance equation for tubes is given by:
With:
The transferred energy to the circulated fluid is determined by:
Tfo and Tf2 are respectively fluid temperature inlet PV/T tubes and fluid temperature outlet PV/T tube. htf is the heat transfer coefficient between tubes and the circulated fluid. The temperature of the circulated fluid into tubes is defined by:
For the insulation layer, it is considered that:
Di is the inlet tube diameter.
The energy balance equation for the insulation is given by:
With:
The transferred energy from the working fluid to pumped water by the exchanger is determined as follows [
The subscripts “tk,i” and “pump” refer to water temperature inlet the boiler (outlet the exchanger) and pumped feed water, respectively.
Water temperature in the boiler is determined by:
Mtk, Ctk, are the mass, and the specific heat coefficient of water in the boiler, respectively. N,
Reverse osmosis presents an effective filtration process for water purification. In fact, the water enters the cartridge and under the applied pressure, water molecules pass the membrane while the other molecules are rejected with a portion of unfiltered water (
Reverse osmosis modeling aims to identify:
-Mass flow rate of the permeate.
-Permeate salinity.
-Consumed energy.
-Water recovery rate.
Reverse Osmosis modeling is based on the solution diffusion principle. The permeate flow rate (Mp [m3/s]) is determined by the following relation [
with SM is the membrane active area in (m2).
Permeate mass flux through the membrane (Jw) and the salt mass flux through the membrane (Js), are defined respectively by the following equations:
Am, B, and S represents the membrane pure water permeability (kg∙m−2∙s−1), the membrane salts permeability (kg∙m−2∙s−1), and the RO membrane surface area (m2) respectively.
The transmembrane pressure (ΔP) and the transmembrane osmotic pressure (Δπ) are expressed as follows:
Pf and Pp are the feed pressure, and the permeate pressure respectively.
The universal gases constant (R) is equal to 8.314 J∙mol−1∙K−1∙T, ρ, and MNaCl are water temperature, water density, and the molar mass of NaCl respectively.
The pressure drop along the membrane channel (ΔPdrop) is defined by:
And,
With: α = 1.7 and λ = 9.5 × 108.
The membrane water permeability (A) is defined by the following relation:
Aref presents the reference permeability at T0 = 298 K without fouling, TCF is the temperature correction factor, and FF is the fouling factor. The influence of membrane fouling on membrane permeability is expressed by the fouling factor FF which varies between 100% for new membranes and 80% for 4 years old membranes.
The temperature correction factor (TCF) is defined as a function of feed water temperature. It is determined by the following relation [
Even if the considered plant architecture can perform well, several operating problems can appear. In fact, Photovoltaic cells temperature impacts directly the electric efficiency of the PV/T field: the increase of Photovoltaic cells temperature decreases the PV/T field electrical efficiency. In addition, for a reverse osmosis desalination process, an overheat of feed water temperature destruct the reverse osmosis membrane. Therefore, the operating water temperature is limited by the manufacturer to 45˚C [
Hence, a Fuzzy Logic Controller based on controlling the flow rate of the circulated fluid into the PV/T panels is designed in order to ameliorate the system performances. The designed control system aims essentially to improve the PV/T electrical efficiency and preserve the reverse osmosis membrane. In addition, it aims to reduce the rejected salts quantity and, upgrade the system productivity.
Fuzzy Logic Control provides an effective manner to draw definite conclusions from imprecise information or vague ambiguous. It presents a suitable way to control solar based systems such as solar thermal plants. FLC is a control system where a mathematical system analyzes analog inputs related to logical variables ranged between 0 and 1. In fact, FLC is based on a control algorithm of linguistic control strategy Depending on a defined set of rules. The consequent results from a linear input-output relation in a region of the state space. The main blocks in the fuzzy logic controller are (
-Fuzzification.
-Rule base.
-Defuzzification.
It consists of converting numerical input variables into linguistic variables. In this step we identify which state variables will be considered as input variables in the fuzzy logic controller. Each piece of input data is converted to degrees of membership by a lookup in a number of membership functions.
The input variables in our system are PV cells temperature and water temperature whose are the main impactful factors on PV/T field efficiency and RO unit performance. PV cells temperature is considered hot when it surpasses 40˚C. From this value we start to think about cooling PV/T cells by increasing the fluid flow rate in order to ameliorate the electric efficiency of the PV/T field. Concerning the feed water temperature, from 30˚C, water temperature is considered hot even if the operating temperature limit is 45˚C in order to prevent any risk of exceeding the operating temperature limit.
The chosen Fuzzifier type is triangular (
Rules are presented in “If… then” format. The collect of rules is named “rule base”. The computer executes the rules and determines the control signal corresponding to the considered measured inputs. The following table shows the fuzzy controller rule base (
Tanked water temperature | PV/T cell temperature | ||
---|---|---|---|
Cold | Average | Hot | |
Cold | Very low | Low | Average |
Average | Low | Average | High |
Hot | Average | High | Very high |
In fact, the control balances between cooling the PV/T cells to promote electric energy production and heating feed water temperature to ameliorate the RO system recovery rate. The relationship between the Fuzzy system inputs and output is represented by
Defuzzification consists of combining and converting all the actions into an output signal, this signal presents the control output signal attributed to the system. The considered method of Defuzzification is “center of gravity”. The controlled signal in our system is fluid flow rate. The membership function distribution of controlled signal is presented by the following figure (
Modeled system and the developed fuzzy controller are tested for the considered distribution of solar radiation and ambient temperature shown in
PV/T layers | Parameters | Symbols | Value |
---|---|---|---|
Glass cover | Aperture area (m2) | A | 2 |
Density (kg/m3) | ρg | 2200 | |
Emissivity | εg | 0.88 | |
Extinction coefficient (/m) | Ʌ | 26 | |
Refractive index | Rg | 1.5 | |
Specific heat (J/kg) | Cg | 670 | |
Thickness (m) | δg | 0.004 | |
PV plate | Packing factor | p | 0.804 |
Temperature coefficient (%) | βr | 0.4 | |
Reference cell efficiency (%) | ηr | 15 | |
Absorber plate | Absorptance of collector (%) | αc | 95 |
Density of plate (kg/m3) | ρc | 2702 | |
Emissivity | εc | 7 | |
Thickness of plate (m) | δc | 0.0002 | |
Thermal conductivity (W/mK) | K | 310 | |
Tube | Diameter of tube (m) | Do | 0.00953 |
Density of tube (kg/m3) | ρt | 2702 | |
Length of tube (m) | L | 1.916 | |
Number of tubes | N | 10 | |
Specific heat of tube (J/kgK) | Ct | 903 | |
Spacing of tubes (m) | W | 0.1 | |
Thickness of tube (m) | δt | 0.0009 | |
Insulation | Density of insulation (kg/m3) | ρi | 20 |
Specific heatofinsulation (J/kgK) | Ci | 670 | |
Thermal conductivity (W/mK) | ki | 0.034 |
The controlled system is used to desalinate seawater. The salinity of the considered seawater is 36000 ppm. It is acquired that potable water salinity should be lower than 500 ppm. RO model simulation shows that feed water temperature should be kept lower than 45˚C in order to guarantee an acceptable permeate salinity (
The simulation illustrates the advantages of the developed fuzzy controller. The employed fuzzy controller balances between the improvement of PV/T field efficiency and the amelioration of the reverse osmosis unit performance. Temperature evolution of each system blocks is shown in
PV/T field electric efficiency is shown in
An improvement in the recovery rate of the reverse osmosis unit is also mentioned. The simulation shows that the recovery rate exceeds 47%. This presents a gain of 12% compared to standards conditions where the recovery rate is equal to 35% with a feed water temperature equal to 20˚C (
Fluid mass flow rate varies according to the state space variables and the system states (PV/T cell temperature, and boiler water temperature), in order to optimize the plant yield as shown in
Moreover, the produced permeate salinity is kept lower than the acceptable limit (500 ppm) (
The method of the life cycle cost (LCC) is used to evaluate the financial viability of the system compared to a similar plant supplied by photovoltaic panels. It is considered that the two compared plants serve to cover a daily water need equal to 227 m3. The total cost throughout its life including planning, design, acquisition and support costs and any other costs directly attributable to owning or using the asset. Life Cycle Costing adds all the costs of alternatives over their life period and enables an evaluation on a common basis for the period of interest (usually using discounted costs). This enables decisions on acquisition, maintenance, refurbishment or disposal to be made in the light of full cost implications (
Product code | Description | Amount | Unit price ($) | Total ($) | |
---|---|---|---|---|---|
PVT-01 | PV/T panel (200 w) | 440 | 455 | 200200 | |
Inverter | 10000 VA, 48 V inverter charger | 7 | 6550 | 45850 | |
Solar battery | Solar batteries 12 V 100 ah | 196 | 84.32 | 16526.72 | |
DC-AC | DC-AC cables and accessories | 110 | 300 | 33000 | |
BS | boiler single serpentine | 1 | 48598 | 48598 | |
PMP-01 | Circulating pump | 55 | 274 | 15070 | |
PMP-02 | water pumping pump | 1 | 3359 | 3359 | |
FT | fluid tank | 55 | 25 | 1375 | |
RO membrane | reverse osmosis unit | 1 | 234972 | 234972 | |
Plant components cost | 598950.72 | ||||
Installation cost | 149737.68 | ||||
Initial cost | 748688.4 | ||||
Maintenance cost | 190430 | ||||
Replacement cost | |||||
Inverter | 28794 | ||||
Water pumping pump | 2109.5 | ||||
Batteries | 10379 | ||||
RO unit | 450710 | ||||
Total of replacement cost ($) | 491992.5 | ||||
TOTAL PLANT COST ($) | 1431110.9 |
Product code | Description | Amount | Unit price ($) | Total ($) | |
---|---|---|---|---|---|
PVT-01 | PV panel (220 w) | 522 | 528.42 | 275835.24 | |
SMA inverter | 10000 VA, 48V inverter charger | 9 | 6550 | 58950 | |
Solar battery | Batteries 12 V 100 ah for Solar Power System | 244 | 84.32 | 20574.08 | |
DC-AC | DC-AC cables and accessories | 136 | 300 | 40800 | |
PMP-02 | Water pumping pump | 1 | 3359 | 3359 | |
RO membrane | Reverse osmosis unit | 1 | 234972 | 234972 | |
Plant components cost | Plant components cost | 634490 | |||
Installation cost | 158623 | ||||
Initial cost | 793113 | ||||
Maintenance cost | 193670 | ||||
Replacement cost | |||||
Inverter | 37021 | ||||
Water pumping pump | 2109.5 | ||||
Batteries | 12921 | ||||
RO unit | 450710 | ||||
Total of replacement cost ($) | 502761.5 | ||||
TOTAL PLANT COST ($) | 1489544 |
Compared criteria | PV/T | PV |
---|---|---|
Daily water production (m3) | 227 | 227 |
Field surface (m2) | 600 | 860 |
Plant cost ($) | 1431110.9 | 1457839 |
Cells efficiency (%) | 13.71% | 11.96% |
Energy consumption (KWh/day) | 705 | 871.5 |
Compared to the technology using PV based plant, the proposed concept and the developed control strategy reduce the reserved space for the implementation next to the initial cost which will be reduced (
Architecture of a PV/T based reverse osmosis desalination plant presented and system functioning is detailed. The system description shows how to make use of PV/T technology to improve reverse osmosis desalination unit profitability. Dynamic model for the conceived system is developed in order to evaluate the system performances. Indeed, the system is subdivided into different subsystems and a state space model presentation is considered to perform the simulation using LABVIEW.
A Fuzzy Logic Controller is designed to improve the system productivity. The simulation shows an improve- ment of the plant performances (PV/T field efficiency and RO unit recovery). Indeed, the designed Fuzzy Logic Controller ensures an energy saving by reducing the consumed energy in the desalination process per cubic meter, and improving the PV/T field electric efficiency. Furthermore, RO membrane is protected by maintaining the feed water temperature lower than feed water limit temperature of reverse osmosis membrane, and consequently, the quantity of rejected salts is reduced which presents an ecologic advantage.
Additionally, the developed economic analysis shows that an important benefit is afforded by using PV/T technology to power reverse osmosis desalination systems. In fact, the comparison shows that for the same system capacity, the plant initial cost for PV/T based system is lower than for PV based system. Consequently, water production cost is reduced.
We express thanks to the School of Natural Resources Engineering and Management at German Jordanian University for the support.
MahmoudAmmous,SanaCharfi,AhmadHarb,MaherChaabene, (2016) Improvement of PV/T Based Reverse Osmosis Desalination Plant Performances Using Fuzzy Logic Controller. International Journal of Modern Nonlinear Theory and Application,05,11-27. doi: 10.4236/ijmnta.2016.51002