Energy storage part #1 - the 24volt system
The goal is to make Artemis self-provisioning on electrical power. Two things are needed. The first are means to harvest wind and solar, the second requirement are means to store it.
There are no photovoltaics nor windgenerator, yet. We begin with the battery system.
But lets explain how the electrical system looks today. Most heavy loads are 24 volts and this is a good thing. It makes motors run faster and it reduces cable sizes and hence losses that would become an issue with 12 volts. Those heavy loads are the bow-thruster, windlass and winches. However, we still need 12 volts. The engine uses this voltage as standard and many of the smaller appliances, lights (LED) and electronics, such as navionics are driven by 12 volts.
Looking closer at the picture above you can see that there are two types of lead-acid batteries in the 24volt bank, one large SLA type and the smaller AGM. This is not good, they have a slightly different nominal voltage and also need different charging strategy. Yet, they are parallelled and thus charged by same sources, the engine generator and the 230 to 24Volt 40 amp charger. This charger may be fed by shore-power or from the 9,5kW Onan generator.
It was soon evident after taking ownership of Artemis that neither 12 or 24 volt service banks managed to hold the voltage more than a day or so. The SLA´s gulped down 4,5 liters of distilled water before the cell levels where correct. However, this is academic as they are all pretty end of life, this happens soon with lead acid. Expect two years. For this reason and a few others, the plan was to replace it all anyway.
We need something better. Something that takes charge faster and has ability to give back all that was put into it. This is not the case with lead-acid, useable capacity is nominal rating divided by two. Hence half of what says on the label. I guess you can see where this is going.
The choice was not to buy drop-in LFP replacements to lead-acid. There are several reasons for this. Cost is one, battery management and system integration is another.
What capacity is required? There are several ways to go about this to work out what you need. Some methods will require educated guesses and estimations. Typically, it is not economically and practical to dimension a system to deal with peaks, instead you back off a bit. With batteries there is also a relationship in-between depth of discharge (DOD) and lifetime, i.e number of cycles. Thus if you have larger capacity for same loadcycles, the larger capacity will live longer.
The choice for Artemis was based on estimates, cost and simply available space. We ended up with 400Ah and at 24volts this amounts to 9,6kWh. Useable energy would be 80-90% of this, in the region of 8-8,6kWh. It is unlikely we will spend this in one day, instead we get at least 2-3 days redundancy if the sun doesn´t shine and no wind. This strikes a balance between what we may maximally produce from 1kW installed solarpanels and wind generator and when we may have to either run the genny or connect shore power. Ideally, we would like to be independent of both.
LFP, if managed, can give way over the rated 2000 cycles (5000 or more), on some conditions. The first to ensure is that neither under nor overcharge can ever occur. The other is to keep the cells balanced. The third is to not charge fully and to not discharge as much as you can. Basically you derate the system and in exchange you get longevity. Herein lies the reason for building the batterypack from scratch, I can set the operating parameters as I wish.
This system has operational consequences, contrary to what many vendors out there say of "drop-in" replacement, the simplicity comes at a cost and this is that you can´t re-design existing or legacy electrical systems for the needs of LFP. This is what is being done here.
LFP´s need, as mentioned above, to be protected from overcharging and overdischarge, meaning in practice that the battery voltages must not exceed limits. Lead-acid batteries have this going for them, they are more forgiving to abuse of this kind. Overcharging simply produces oxygen and hydrogen out of the water in the electrolyte. It may be replaced. Taking out too much, that is going below around 11,5 volts takes a toll on the battery life by sulfation of the electrodes but can, to certain extent, be regained by charging strategies.
In LFP no such mechanism exists, if gassing is allowed to happen it will be preceeded by increase in temperature. Once the heat of the electrolyte reaches around 60 degrees Celsius it begins do decompose. It leads to a pressure increase, in turn resulting in the cell expanding in size, swelling. This changes the electrical properties radically. All of which is inconsequential as the cell is now destroyed.
From this it may be learned that having control over the cell temperature is fundamental in the battery management.
For the LFP system it means that the battery bank should be looked upon as being a resource that decides for itself if going to be available for charging and discharging.
The consequence of this is that both the load and the chargers may be disconnected from the batterybank without prior warning. If this was allowed to happen generators and charges would not be happy resulting in possible damage to attached devices. Thus there has to be something that ensures this situation is dealt with.
As for the argument that we are putting LFP into a lead-acid system this requires som explanation. One way to look at this is from a voltage - current perspective. Legacy electrical systems in boats and cars are voltage driven. Simplified, this means that if you provide a certain constant voltage in the system the lead-acid storage will charge itself and it will do this quite gradually, hence not load the source of the voltage unduly as it is a rather slow process. The higher this voltage is, the faster the energy is taken up by the lead-acid battery. The voltage level can be seen a pressure difference regulating in which direction current may flow. Eventually the voltage at the lead-acid battery rises to become the same as the one from the generator, the charging stops by itself.
New chemistries, such as lithium-ion of which LFP (LiFePo) belongs to, are more current driven.
Meaning that they present themselves to the voltage source as a high load and will take all you give them. That is, their properties are to be able to give off and take up energy radically faster than lead-acid. This, desirable, characteristic also makes them capable of loading the voltage source, i.e chargers, more heavily. Most in-situ lead acid batteries live their life in relatively primitive charging situations. The typical automotive/marine engine has a generator that simply tries to maintain a pre-determined voltage (typically in the range of 13,8 to 14,7volts) against the loads of the system and the battery. BTW, this is ways off how the manufacturers of smart lead-acid battery chargers claim these batteries should be charged. They may be right, but there is not much they can do about it when most of the charging in cars and boats is done with simple constant voltage sources. The system offers a constant voltage and this is what we have to work with. Now, if we exchange lead-acid for LFP we are not having the desired charging method as LFP wants constant current for most of the charging period, changing over to constant voltage at the very end and where this voltage is very carefully set (down to one-hunded of a volt per cell) such that it never exceeds this voltage. Then, after a short period, it is turned off as it is not desirable or necessary to trickle charge LFP. As you know, lead-acid needs trickle charging to be healthy as the "leak" a lot.
Let´s return to our system and the issue of the charging side. It would be possible to turn off the field current in the generator if it had this possibility to stop charging. Some generators have this feature (or you can open the "ignition light" circuit cutting off the field supply). On Artemis it cant be done. For the Solar charger it should be instructed to be turned off and for the shore or genset driven charger same would apply. Doing this is added complexity. A simpler method that is going to be used initially is to allow the LFP group to be disconnected but keep a smaller lead-acid battery as load. Probably an SLA type as it is robust against mishandling and allows for the electrolyte to be replenished (water).
All the charging devices will also be instructed to put out one steady voltage, high enough to charge without overloading the chargers but not so low no charging will take place. No other fiddling around by the chargers is desired.
Such a strategy, although not very refined, also works on the discharge side. Should the LFP bank disconnect due to low voltage there will be no loss of power to the onboard system as the lead-acid is still there. Means to notify the crew in this case is prudent as it is a command to connect to shore or run the genny.
What else should be considered? There is a risk that if the LFP bank is at low state of charge and nothing is charging them and you then start the engine. If you then proceed to run it for awhile above idle, then the generator will try its best to put out as much as it can as the LFP´s are taking it all in. Loading these types of generators is not what they are designed for. Most do not even have means to turn them off when they get hot. It is this repeated and extended heating that will toast the windings. So, ideally the engine generator should have an external controller that protects the generator from overloading.
This ends this part of the blog, will return with pack design, the DC-DC conversion and the actual work to get it all done. All is unchartered territory on Artemis.
There are no photovoltaics nor windgenerator, yet. We begin with the battery system.
 |
| This is what the current storage looks like when the floor boards are removed. The four at the top are all for the 24 volt service bank. The 12 volt ones are the ones lowest in the picture. Unseen, beneath them is the engines own 12volts battery. |
 |
| Shown here is the arrangement of the service banks for 12 and 24 volts. Not shown is the starter battery for the genset and the engine itself. Also not shown is the 230volts has two sources, shorepower and genset. |
Looking closer at the picture above you can see that there are two types of lead-acid batteries in the 24volt bank, one large SLA type and the smaller AGM. This is not good, they have a slightly different nominal voltage and also need different charging strategy. Yet, they are parallelled and thus charged by same sources, the engine generator and the 230 to 24Volt 40 amp charger. This charger may be fed by shore-power or from the 9,5kW Onan generator.
It was soon evident after taking ownership of Artemis that neither 12 or 24 volt service banks managed to hold the voltage more than a day or so. The SLA´s gulped down 4,5 liters of distilled water before the cell levels where correct. However, this is academic as they are all pretty end of life, this happens soon with lead acid. Expect two years. For this reason and a few others, the plan was to replace it all anyway.
We need something better. Something that takes charge faster and has ability to give back all that was put into it. This is not the case with lead-acid, useable capacity is nominal rating divided by two. Hence half of what says on the label. I guess you can see where this is going.
The choice was not to buy drop-in LFP replacements to lead-acid. There are several reasons for this. Cost is one, battery management and system integration is another.
What capacity is required? There are several ways to go about this to work out what you need. Some methods will require educated guesses and estimations. Typically, it is not economically and practical to dimension a system to deal with peaks, instead you back off a bit. With batteries there is also a relationship in-between depth of discharge (DOD) and lifetime, i.e number of cycles. Thus if you have larger capacity for same loadcycles, the larger capacity will live longer.
The choice for Artemis was based on estimates, cost and simply available space. We ended up with 400Ah and at 24volts this amounts to 9,6kWh. Useable energy would be 80-90% of this, in the region of 8-8,6kWh. It is unlikely we will spend this in one day, instead we get at least 2-3 days redundancy if the sun doesn´t shine and no wind. This strikes a balance between what we may maximally produce from 1kW installed solarpanels and wind generator and when we may have to either run the genny or connect shore power. Ideally, we would like to be independent of both.
 |
| The first stage of the reconfigured electrical system. Note the 24 to 12 volt DC/DC conversion. This will be explained in separate blog. |
LFP, if managed, can give way over the rated 2000 cycles (5000 or more), on some conditions. The first to ensure is that neither under nor overcharge can ever occur. The other is to keep the cells balanced. The third is to not charge fully and to not discharge as much as you can. Basically you derate the system and in exchange you get longevity. Herein lies the reason for building the batterypack from scratch, I can set the operating parameters as I wish.
This system has operational consequences, contrary to what many vendors out there say of "drop-in" replacement, the simplicity comes at a cost and this is that you can´t re-design existing or legacy electrical systems for the needs of LFP. This is what is being done here.
LFP´s need, as mentioned above, to be protected from overcharging and overdischarge, meaning in practice that the battery voltages must not exceed limits. Lead-acid batteries have this going for them, they are more forgiving to abuse of this kind. Overcharging simply produces oxygen and hydrogen out of the water in the electrolyte. It may be replaced. Taking out too much, that is going below around 11,5 volts takes a toll on the battery life by sulfation of the electrodes but can, to certain extent, be regained by charging strategies.
In LFP no such mechanism exists, if gassing is allowed to happen it will be preceeded by increase in temperature. Once the heat of the electrolyte reaches around 60 degrees Celsius it begins do decompose. It leads to a pressure increase, in turn resulting in the cell expanding in size, swelling. This changes the electrical properties radically. All of which is inconsequential as the cell is now destroyed.
From this it may be learned that having control over the cell temperature is fundamental in the battery management.
For the LFP system it means that the battery bank should be looked upon as being a resource that decides for itself if going to be available for charging and discharging.
The consequence of this is that both the load and the chargers may be disconnected from the batterybank without prior warning. If this was allowed to happen generators and charges would not be happy resulting in possible damage to attached devices. Thus there has to be something that ensures this situation is dealt with.
As for the argument that we are putting LFP into a lead-acid system this requires som explanation. One way to look at this is from a voltage - current perspective. Legacy electrical systems in boats and cars are voltage driven. Simplified, this means that if you provide a certain constant voltage in the system the lead-acid storage will charge itself and it will do this quite gradually, hence not load the source of the voltage unduly as it is a rather slow process. The higher this voltage is, the faster the energy is taken up by the lead-acid battery. The voltage level can be seen a pressure difference regulating in which direction current may flow. Eventually the voltage at the lead-acid battery rises to become the same as the one from the generator, the charging stops by itself.
New chemistries, such as lithium-ion of which LFP (LiFePo) belongs to, are more current driven.
Meaning that they present themselves to the voltage source as a high load and will take all you give them. That is, their properties are to be able to give off and take up energy radically faster than lead-acid. This, desirable, characteristic also makes them capable of loading the voltage source, i.e chargers, more heavily. Most in-situ lead acid batteries live their life in relatively primitive charging situations. The typical automotive/marine engine has a generator that simply tries to maintain a pre-determined voltage (typically in the range of 13,8 to 14,7volts) against the loads of the system and the battery. BTW, this is ways off how the manufacturers of smart lead-acid battery chargers claim these batteries should be charged. They may be right, but there is not much they can do about it when most of the charging in cars and boats is done with simple constant voltage sources. The system offers a constant voltage and this is what we have to work with. Now, if we exchange lead-acid for LFP we are not having the desired charging method as LFP wants constant current for most of the charging period, changing over to constant voltage at the very end and where this voltage is very carefully set (down to one-hunded of a volt per cell) such that it never exceeds this voltage. Then, after a short period, it is turned off as it is not desirable or necessary to trickle charge LFP. As you know, lead-acid needs trickle charging to be healthy as the "leak" a lot.
Let´s return to our system and the issue of the charging side. It would be possible to turn off the field current in the generator if it had this possibility to stop charging. Some generators have this feature (or you can open the "ignition light" circuit cutting off the field supply). On Artemis it cant be done. For the Solar charger it should be instructed to be turned off and for the shore or genset driven charger same would apply. Doing this is added complexity. A simpler method that is going to be used initially is to allow the LFP group to be disconnected but keep a smaller lead-acid battery as load. Probably an SLA type as it is robust against mishandling and allows for the electrolyte to be replenished (water).
All the charging devices will also be instructed to put out one steady voltage, high enough to charge without overloading the chargers but not so low no charging will take place. No other fiddling around by the chargers is desired.
Such a strategy, although not very refined, also works on the discharge side. Should the LFP bank disconnect due to low voltage there will be no loss of power to the onboard system as the lead-acid is still there. Means to notify the crew in this case is prudent as it is a command to connect to shore or run the genny.
 |
| Block digram of the idealised and cleaned up 24 and 12 volt system using 10kWh LFP (LiFePo) storage. This system can handle charge and discharge separately by way of high power low drop diodes. |
This ends this part of the blog, will return with pack design, the DC-DC conversion and the actual work to get it all done. All is unchartered territory on Artemis.
Comments
Post a Comment