Play video to see the complete setup of the installation

Last week i recorded a small video during operation of the system while the motor was operating smoothly. This gives a better impression of the whole installation and setting: the room, desk in the corridor, the system,  tanks, piping, wiring, electrical and electronic setups, tools and devices, and airconditioning!

October shows early winter signs

October is almost midway but it already resembles very much of late november or even December. The weather is not at all typical of previous years when it was mostly sunny with high temperatures until late october.

Especially last week, the temperature was about 16-8 degrees Celcius and mostly cloudy, and with abrupt changes during the day which are very difficult to forecast. There were days that the system couldn't operate at all (like last weekend) because the irradiation never rose above 400 W/m2 when in order to start smooth operation, it needs at least 300 W/m2. In fact the maximum irradiation that was reached was half than the maximum of all the previous days in october and september:

Solar radiation day distribution on 09/10/10

Solar radiation day distribution on 10/10/2010

One typical day of the previous days in October was like on 8/10/10:

Solar radiation on 08/10/10

As it can be seen from the above graph, fluctuations in irradiation were significantly deep and abrupt. System operation without a motor speed regulator but manual control, was extremely difficult and inefficient while frequent start and stops to protect the pump were unavoidable. 

All results have been logged with the datalogger and conclusions/comparisons will be announced later this week.

Observations from the system operation with LCB and without!

The LCB controller or Linear Current booster is used in solar direct pumping applications. It is claimed to achieve 30 - 90 % increase in the water pumped than connecting the motor directly to the solar panels. At times when the sun is not optimum, a Linear Current Booster boosts the current in the pump motor to produce the torque to turn the pump motor. It does this by sacrificing voltage on motor, hence the motor runs slower. This results in a motor that may be pumping slower (there is after all less solar power), but this is better than having a stalled pump pumping nothing which would happen if solar direct.

The LCB that we are using is from the company "Solar Converters Inc".

So, what are the things which have been observed so far? 

Observations from the absence of the LCB:

In the last week of September (when LCB was not still connected), the sky was heavily clouded (leading to a (global) irradiation of <200 W/m2 for most time of the day for 4 days) and system operation was difficult. The motor was abruptly stalling, leading to significant vibrations and the motor starting and stopping almost every 10 minutes. However, that practice is not safe for the motor over extended time scales, and neither does it producing good results from which conclusions can be drawn. The situation is aggravated by the absence of a speed controller for the Pearson pump, the need for which is felt especially in times when there are wide fluctuations in the irradiation levels. 

Observations with the use of the LCB:

On 3/10/2010, 5 days of system operation with LCB were completed.

The LCB controller mounted on the wall

Open without the cover

The PCB of the controller with the LEDs flashing indicating the voltage levels

Opening at the cover case in order to observe the LEDs with the cover on

Experimentation trials:

29/09/2010: First trial with LCB at 13:00

30/09/2010: Full day of experimentation (2nd trial)

01/10/2010: Half day of experimentation. Campus-wide power failure of 3 hours resulted in stoppage of operation and destruction of data from the data logger.

2/10/2010: operation during first half of the day. Heavy clouds were observed after 14:00 until sunset.

3/10/2010: operation during second half of the day due to dark sky.

After connection and operation with the LCB, the following were observed:

- Motor doesn’t stall and noise is smoothened.

- Solar fluctuations smoothened by the LCB. The controller avoids deep and steep changes of motor speed.

- By observing the LCB LEDs which measure the Voltage level supplied every moment, it is easy to predict when the motor speed has to be adjusted (in clear sky conditions).

- Significant Exchange of Voltage with Current occurs most of the time, in order to supply with the maximum instant current possible. However, the limitation of the LCB is 40 A and it cannot give more than that. Luckily, the flow rate corresponding to this limitation coincides with the recommended limit of the membranes for feed flow set by manufacturer at a maximum of 6gpm (gallons per minute). After 5 days experimentation, it was observed that the 40 A were corresponding to feed flow of 6 gpm or higher. As a result, the limitation of the 40A doesn’t effectively limit the system’s capacity.

- The system operates with higher power than before the LCB during low irradiation in the morning and especially in the afternoon.

- Voltage Exchange for one specific irradiance value depends on the type of the irradiation (if it is direct or diffused). In direct irradiation the LCB manages to give more current than when the same amount if irradiation is diffused. For example, in the afternoon the LCB can give more current (corresponding to higher permeate flow) than in the morning, for the same irradiation. It is important to note here that this is attributed to the characteristics of the PV panels.

- LCB works better when there are no rapid fluctuations. When irradiation reduces smoothly, then the voltage levels also reduce smoothly (due to LCB), and current increases considerably.

Electrical switch and LCB get installed

Simplifications of systems sometime cause a lot of inconvenience and insecurity. This was the case also to our system which in the beginning had as simple electrical connections as possible:

any time there was a need to change the power source, a procedure of at least 10 min had to be followed with a lot of screwing and unscrewing in fuses, raw cable manipulation, and tapes..

Luckily this was solved this week, with some great family help, a switch got installed between the PVs and the rectifier. We also installed the LCB between the PVs and the switch as well as some extra fuses before the motor as suggested by the LCB manufacturer:

Electrical connections diagram

Some photos of this amateur installation:

Putting a wood plate as a base in order to avoid drilling too many holes on the wall

The switch

Installing the switch

The switch between the LCB and the fusebox

 

One interesting visit from the stuff of the Laboratory of Electric Circuits and Renewable Energy Sources from the Electronic Engineering Department

Last week the Laboratory of Electric Circuits and Renewable Energy Sources from the Electronic Engineering department visited our installation, together with the professor dr.Kalaitzakis and 5 master students (http://www.ece.tuc.gr/Controller?actionClass=loadFirstPage&action=loadFirstPage).

They were very interested in our project, since they are also involved in desalination (of course from the power and control systems point of view) and we discussed many topics like the introduction of the automatic controller so that I free my hands and also get more accurate results. Furthermore, they expressed their worries about the motor lifespan under such fluctuating operations. They are afraid that it won't last for long without damages and that we should take care for its protection. They suggested to install a supercapacitor (small), which will only serve purposes of smoothing these deep current fluctuations in order to provide stable current to the motor. 

The LCB may help the situation with the fluctuations but for sure it will not replace the manual regulation of the motor speed by me during day operation. 

We also discussed about the Voltage and current losses which i observe in datalogger and they think that they are remarkable. However, it is not very obvious what is the reason, it may be the sensors construction or some resistance occuring via the connections. 

Another suggestion was to introduce a W-meter at the PV panels, so that i know also the power from the PV panels and estimate better how much losses i get while measuring with the datalogger and account for them.

Kostas and Ariadni, master students at the Laboratory, visited again the installation trying to understand better how the motor works and find out the safe operating range. After contact with Spectra (the provider of our system), we were informed that there is no operating range as such and no minimum voltage requirement. Ariadni and Kostas found this strange, since all motors have an operating range  in which they operate constant and safe. 

Maybe this has to do with this specific motor which is permanent magnet dc current? 

this is something still to answer..

Manual control of the high pressure pump.. is not a good practice?

While being in Delft, we were thinking that the more simplified the system the better would be. For this reason, we decided not to put any electronics or controllers to control the high pressure pump (Pearson pump).

However, after 4 days of full operation, it is obvious that the manual control of the pump is not the best practice.

The Pink line is the feed flow over time and the blue line is the irradiation over time, for two different days

The Pearson pump is powered by the PVs, which means that during the day, the pump has to be regulated manually depending on the fluctuations (clouds passing). Initially i thought that this wouldn't be very difficult, however because of the season's high sun fluctuations, its very difficult to  achieve the maximum flowrate for each instant irradiance. As a result, the operation of the motor is not optimised, and the total permeate production over the day is lower.

Maybe it would be a consideration to insert an automation system for the pearson pump, which will allow it to follow the fluctuations of the sun more accurately?

A Mobile Air-Conditioner for the Pearson Pump Motor - A Cool(ing) Solution!

Overheating of the Pearson Pump motor is a problem that we have faced since the beginning of our experimentation in Crete (the problem being aggravated due to lack of ventilation within the room that the facility has been placed in). Initially we had used a small table fan to supplement the cooling effect of the motor fan...

Table Fan Cooling System

However, it seemed that this didn't stop the motor of the Pearson pump from overheating. Prof. Diamandopoulos was kind enough to arrange a mobile air-conditioner unit (Cooling Effect = 2200W) for our system. It just arrived and I am very curious to see if indeed it makes a significant difference!

The Pearson Pump Motor with the Mobile Air-Conditioning Unit

Electrical Cables and New Voltage Sensor

Last week, while Andres was still in Chania, we noticed that it was difficult to keep changing between PV Power and Grid Power (in case we wanted to carry out some tests in the late afternoon when the PV Power was too low).


Basically, the two sources were equipped with different kinds of cables: the PV cables were bare, while the grid power cables had copper ring endings. 

We asked the electrician who had connected the PVs to help us find an easy solution for this and what he did was that he cut the copper ring endings off and put the bare cables in both the power sources. This week I learnt of a switch which can be installed between the two power sources and with easier way interchange between PV and grid power source. However, I need to speak to Sander (technician from TU Delft) about this to confirm that this will not lead to a loss in calibration.

An unfortunate incident which happened when the electrician was here is that in his effort to fix the cable he accidentally broke the sensor for the Voltage measurement. It took us two days to find where we can get supply of new resistors of the same type, and a tool to connect them. Consequently, for 2 days I was recording the voltage manually (on paper) after every 10 minutes.

My friend, Charalampos who studies Electronics here in the TUC helped me by making a new circuit.

The Soldering-Iron with Charalampos

Charalampos in Action!

There is still a small deviation in the voltage readings we measure and those shown by the multimeter in the fusebox (coming from PVs). A Professor from the Electronics Engineering department is interested to see the facility and is coming tomorrow together with all the people from the laboratory. Hope they can give some good advice about it.

Water heating issues!

Since yesterday we are trying to tackle down the problem of the water which gets too hot during the day. Yesterday the water in the tank started from 28.8 oC and after 8 hours the water temperature rose to 31.8 oC. Initially we tried to put a dozen of icepacks inside the black 1m3 tank. However, as expected this showed no difference to the temperature, some icepacks are not enough to cool 1 m3 of water!

 putting icepacks in the tank

Today we tried another solution: We made the pipe of the concentrate stream to pass from a bucket full of icecubes and icepacks before ending in the 1m3 tank. we hoped that the ice could cool down the concentrate stream before it returned to the tank.However, the material of the pipe proved to be a good isolator, and heat exchange with the ice was not achieved.

concentrate stream (in a spiral configuration) inside a bucket full of ice

we have to think about a heat exchanger solution probably using a spiral coil, placed inside the 1m3 tank.

One week in Creta - day devoted to PV power

Today we close one week since the system reached the laboratory where we are placed. It was a very intensive week, full of hard labour work. Andres may disagree with this, but this is the experience of an unfit person like me. We managed in one week to do lots of things concerning installation, setting up things, cleaning rooms, bringing things into place, modify tanks and pipes. Right now everything seems to be in perfect place, datalogger was installed and calibrated yesterday, solving also some problems in electronics connections.

It is 9:00 and we are waiting for the sun to go up since its still somewhat below 200 W/m2. Today the day is devoted to desalination powered by the sun and datalogging the results.

Still there are issues which are not resolved:

- For the 3rd time of operation the permeate quality we get is not satisfactory (around 680 mg/l when the maximum limit for drinking water is 500 mg/l, assuming the same relation as Brett Ibbotson was using for conductivity and TDS). As a result yesterday night we didnt flush the system with very pure water, and as a consequence the concentrate of the membranes after flushing had a concentration of 1210 mg/l. Note that the membranes are considered clean when the concentrate current is not more than 1000 mg/l.Of course the assumption for the relationship between the TDS-electrical conductivity has to be changed and be calculated for the specific seawater sample we get in chania).

- Second issue is that the power cable of the pearson motor has a copper ring end which doesnt fit from the opening of the fusebox (when powered by the PVs). For the time being, Andres has used a piece of a copper raw cable to extend the original power cable, in order to manage to connect it in the fuse box. However, we are not sure if this can give more resistance, and influence the calibration to the datalogger (when measuring the voltage and the current of the motor).

Sun is rising higher, i will post more at the end of the day :)

1st operation with the PV system for one hour!

On Saturday after the long "running" procedure powered by grid, we connected the system with the PVs! It was around 17:00 and it worked perfect, with no a lot of fluctuations until 18:00. 

desalination with PV!

no grid, only sun!

First water from the sun!



Feed flow = 5gpm
Permeate flow = 1 gpm
Membrane pressure= 55 bar
ΔP at prefilter = 1 bar
Permeate conductivity = 1200 μS/cm corresponding to 780.96 mg/l.


The permeate quality is an issue we have to deal with today.

At 18:00 the irradiation fell abruptly because of heavy clouds, and then we had to stop the system. 

Today we are starting operation with PVs and see how it will go!

1st long running to observe the system operation - still by grid!

Using the configuration which is already described, we removed 150 l of permeate to a tank above the 1m3 tank, and then permeate was returning to the 1 m3 by overflowing. as a result the system reached steady state not at 37000 mg/l (which is the real salinity of the seawater of the region), but at a higher salinity of around 44775 mg/l.

On Saturday, we ran for around 4 hours the system (powered by the grid), to observe how it operates.
We observed the following in the beginning:
Feed conductivity= 68.8 mS/cm which corresponds to: 44775 mg/l                               at 27.4 C
permeate conductivity= 1149 μS/cm which corresponds to 747 mg/l                              at 27.5 C
concentrate conductivity= 75 mS/cm which corresponds to 48810 mg/l.
membrane pressure= 56 bar
ΔP of prefilter = 1.2 bar
Qfeed = 5 gpm
Qpermeate = 0.95 gpm

flux= 14.6 lit / h*m2   (Membrane total area = 14.8 m2)



After 4 hours we observed:
Feed conductivity=67.5 mS/cm which corresponds to: 43929 mg/l                               at 28.8  C
permeate conductivity= 1187 μS/cm which corresponds to 772 mg/l                              at 28.9  C
membrane pressure= 55 bar
ΔP of prefilter = 1.2 bar
Qfeed = 4.8 gpm
Qpermeate = 0.95 gpm

flux= 14.6 lit/h * m2  


in the course of 4 hours the temperature rose 1.4 C, and also the permeate conductivity rose instead of falling. Its considerably high compared to the results in the lab.

At the end of the day we flushed with 200 l of demineralised water, making sure that the membranes were totally clean. We will see today if we have any differences in permeate quality.

We decided also to dilute the seawater in the tank with demineralised water,  so that after filling the 140l permeate in the small tank, the steady state of the system would reach the salinity of the  real seawater in chania (around 37000 mg/l).

configuration for flushing and increased salinity in the tank

the salinity of the seawater we transported from the sea is around 37000 mg/l.
finally yesterday we completed the whole setup,tanks and modifications in tanks and placed a 200 L tank (for permeate storage) above the 1m3 tank. The plan was that while running the system, the permeate would fill the 200 L tank and the concentrate would return back to the 1m3 tank. When the 200 L tank was full, permeate would overflow back to the 1m3 tank.

We ran the system yesterday in this configuration. As a result, the salinity in the 1 m3 tank increases during the whole "running" (since around 150 L permeate didn't return in the 1m3 tank but was stored in the 200 L tank).
So the fact now is that after 150 L permeate is stored in the 200 L tank, the salinity in the 1m3 rises to 43000 mg/l, and this is the salinity that our membranes are required to desalinate during the whole "running".

We thought that after we flush the system, we should return the 2 streams back to 1m3 tank to decrease the salinity back to around 37000 mg/l.
We are not really sure if this is the right way to run the system.

The question is: Should we continue in this configuration, even though we run the system in extremely high salinity, which doesn't correspond to the real salinity of the seawater used? (The system can handle maximum 45000 mg/l).

One alternative would be to dilute the seawater contained in the 1 m3 tank, so that when the system reaches steady state after filling up the "permeate tank", the final concentration is 37000 mg/l as the real seawater of the region.

One extra problem we face right now is the high salinity of the permeate.
We ran the system yesterday for 2 hours using the above configuration, but we found out that the permeate quality was really high (1230 mg/l). If we run it for more time, we hope that the permeate quality will decrease. But if this is not the case, then we have to consider other issues.

Filling up the tanks!

It was both fun and complicated to bring seawater from the sea. Mostly because we had to transfer the water with my small car, which could hardly fit a 200 l barrel.

Andres enjoyed some nice moments in the sea while keeping the pump at the right position :)

In order to fill the 1 m3 tank, we had to repeat some times the whole procedure. If we consider that one complete trip (from pumping seawater from the sea, to transporting, pretreating and storing it in the 1m3 tank) was taking a little bit more than 1 hour, it is understandable why we finished late at night!

Photovoltaics look good on the roof!

Our Photovoltaics got installed yesterday :) Yeahh!

At the end of the afternoon the last installations of the electrical connections finished, and the photovoltaics were giving 32 Voc at 16:00 in the afternoon :) 

 

We are so looking forward to start testing the panels with our system!

We didn't manage to run the system yesterday, because we had underestimated the time needed to prepare the room (was not used the last 10 years, and we were cleaning for more than 2 hours), and to install the system in the room. The bulky 150 kgs "wardrobe" proved to be too tall to fit from the door! The only solution was to remove the wheels from it!We used jacks to lift it, and Andres and Nikos proved to be really strong!

Our small room, looked even smaller after the tank, the feedpump and all tubes and equipment were placed inside. 

We have the feeling that the pearson pump will get too hot in the room, since there is no ventilator or window. We have to think how to tackle this.

The day today will start with flushing the system with demineralised water produced in our lab :) 

then finally we will visit the beach to start bringing the seawater. Pyranometer and datalogger have also to be connected, cables do not seem long enough, we have to think where to put them.

nothing happens when you are expecting it to happen!

First two days on site passed mostly expecting for things :P

Expecting the desalination system and the rest boxes which came only today afternoon instead of yesterday morning, expecting the Photovoltaics company to confirm that no further delays will happen and expecting also for the rector of the university to give authorisation to use a recently -vacant room to set up the system. (The rector had to be involved, since the room was claimed by two different departments and prof.Diamantopoulos had to fight for it).

Monday was mostly the welcoming day and the day to settle down. Lots of administration work, meeting the people of the lab, planning the next steps and procedures, but also settle down in the office, configure computers, netIDs and doing lots of phone calls to trace the membranes-pumps and photovoltaics! 

The good news came today when we finally received our system!  Following Murphy's law, news cannot be only good; We noticed that the boxes didn't arrive in good order, at least on surface. One wheel of the main-system-box was broken, and all the oil from the pump crankcase had leaked out to the wooden support frame.. from down to up! It seems that the box had been turned upside-down! Unfortunately, we didn't have time to test it today since it arrived late, we have to wait until tomorrow to be sure that there is no further damage.

We got also our 1 m3 tank today to store the seawater, cleaned it and modified it with required fittings and valves. We will fill it in tomorrow with real seawater brought from a perfect location. The beach of Agios Onoufrios is 10 min away from the university and is also awarded with the "blue flag" (www.blueflag.org). This may be very useful with respect to fouling of the filters and membranes, since it is one of the cleanest beaches  in crete! 

For pumping the seawater tomorrow, we will use a submersible pump which will be powered using the facilities of a friendly owner of a cantine just 40 metres away from the coast  :)

It is interesting to note that the seawater here has 39000 ppm of TDS (highest number in the interconnected system of the earth's oceanic waters) which makes our desalination effort quite challenging and of course energy costly.

Whats next for tomorrow?

-The PV company will start the installation of the panels early in the morning, and promised to finish in one day (never trust the greeks!).

-Fortunately, we will get the keys of the room where we will set up the system, and we have to start by cleaning the place (it looks like it has not been used the last 10 years), then connect  electronics, tubes and pumps, and then flush the system with Demiwater. Hopefully, there is no damage from the transport (i try to win my pessimism)..

-If everything goes well, we will pump and bring the seawater in the afternoon and fill our tank.

Andres and me have to note that we get a lot of help and advice from the people at the lab people and prof.Diamantopoulos. Especially we are very grateful to Nikos Vakondios, a phD student who is greatly involved in our project and gives great help in practical things.

PS: Next days will be busier and more tiring and i will not have the time to write long posts like this, luckily for you :)

Tomorrow seems to be a very busy day ..

Goodnight from Crete,

Valia

Drinking with the sun on field!

Hoi!

The drinking with the sun project is about to begin the field testing!

This blog-book is created in order to log the daily activities, achievements and challenges we are facing during the field testing of our solar-powered desalination plant.  Via this blog-book me and andres will be able to keep everyone updated back in Netherlands.

The pilot is expected to reach Chania, (Crete, Greece) on monday 6th September 2010. 

First day activities will be to set up the location of the system, filling up the tank with seawater, make all fittings and pipe connections, install the datalogger and do all electronic cable connections. Final first step will be to test the good operation of the system powered by the grid.

The photovoltaics will be connected around the first days of the next week, waiting for the company confirmation.

curious to see how long will be the procedures of installation. hope for the best :) 

Valia:)