Presented to the
UNOLS DEep Submergence Science Committee
May, 1996
by Dudley Foster
Introduction
Since ALVIN was first launched 1964, lead-acid battery technology has provided a source of power that has been reliable, readily available, easily maintained, and cost effective. As in all electrical systems, there has been a continual increase in the need for more power as technology has provided new equipment for science instrumentation and imaging applications. The paramount factors in power consideration are safety and service reliability. The following report discusses several battery technologies considered for use in ALVIN. When factors of serviceability, reliability, cost, payload and space are considered, the higher capacity lead-acid cells still have an overall advantage compared to other relatively mature battery technologies.
Background
Appendix A is a summary of ALVIN's battery and power distribution evolution. The first design philosophy was to separate power for control (instrumentation, communication, life support, etc.), propulsion (thrusters and lights), and science applications. The design insured that if science equipment depleted its power, propulsion and control would still be available. Also, if propulsion power were depleted, control power was still available for critical surfacing, life support and communications functions. Since ALVIN was very much a prototype and development vehicle at that time, this conservative approach was prudent. The fundamental concern for safety as the primary design criteria has not been neglected in all subsequent design evolutions of the power system. However, experience has safely allowed changes resulting m equal power distribution throughout the submersibles systems.
By 1967, having three different battery systems was proving to be unreliable and a major maintenance problem. At this time a two-voltage system was implemented, essentially combining the science and control batteries into a single 30 volt system. The 60 volt propulsion batteries were retained as a separate power source. The resultant change allowed all batteries to be of the same type (six volts each), allowed design of a lower maintenance battery box, allowed simplification of the power distribution system, increased the battery power, but also decrease the payload due to heavier batteries.
In 1986 continued failures of the Hoover propulsion motors at increased operating depths necessitated conversion to DC motors. The hydraulic propulsion system was replaced with individual electric thrusters. These motors required 120vdc and the combined 60/30 volt battery tanks were replaced with two 120 volt battery tanks and one 30 volt battery tank. In the previous configuration, both 30 volts and 60 volts were combined in each battery tank. Either battery tank could therefore provide all necessary submersible voltages in the event of a failure in the other tank. In order to retain the safety features of this power redundancy, the new 30 volt battery tank had three separate 30 volt strings. One string had lower capacity, lighter, batteries that went to an auxiliary 30 volt bus. In the event of a problem with the primary 30 volt bank the auxiliary 30 volt string could be brought on line to allow uneventful return to the surface. The disadvantage of this was that all the 30 volts were in one tank, and if that tank flooded or had to be dropped, only emergency batteries in the sphere were available for surfacing. At this time, the added batteries resulted in a 67% power increase but required filling almost all available space with a total of 1000 pounds of syntactic foam.
In 1988 the source of the 225ah Exide batteries became unreliable and they were replaced with 200ah KW batteries. Although these cells had less power, they were fighter, less expensive, and of better quality.
Even though the sub now had more power than pre-1986, many missions were still limited by the depletion of either the 120 or 30 volt batteries. Dive profiles requiring heavy lighting, hydraulics, or propulsion usage would deplete the 120 volts when the 30 volts still had power available. Conversely, missions with heavy instrumentation would deplete the 30 volt battery with power left in the 120 volt propulsion batteries.
The Current Battery System
In 1989, the 30 volt batteries were completely eliminated by installing 120vdc to 30vdc power converters inside the sphere. This had the advantage of allowing all the power available in the batteries to be consumed regardless of the mission profile. At this time the sub started to carry only two 120 volt battery tanks. This allowed 100% power redundancy because all systems could be run from either battery if one were flooded or dropped. Although there was space for a third 120 volt battery tank, there was insufficient payload to install one. A study was done in 1991 to determine how we could accommodate another battery and increase overall capacity. This was the motivation to develop the pressure tolerant motor controllers (PTA). By adding a large piece of syntactic under the stern removing reserve steel ballast, and removing the motor controller pressure cases, there would be sufficient payload for a third tank. Appendix B shows the results of that design study. Since the pressure tolerant controller program was unsuccessful, we are still limited to two battery tanks using lead acid cells in the 190-260ah range.
Another factor limiting the use of the rated power of the batteries was the depletion of water in the cells over time. The normal loss of water in the cells during charging eventually exposed the battery plates to compensating oil and reduced the battery capacity until the battery could be removed for servicing. Due to operating demand, this service took place every six months, and experience showed that the last few cruises before a service period usually suffered from reduced battery capacity. To try and maintain battery capacity, and allow more operational days, a battery rotation program was implemented that replaced one of the two batteries every two months. As a result, each battery is serviced every four months and is less likely to suffer from water shortage and exposed plates. This method eliminated the need for a two-week maintenance stand-down every six months and allowed an additional 30 operational days per year. Installing three battery tanks with fighter and lower capacity batteries was evaluated. By using three tanks of 165ah cells, the total capacity could be increased by 24%. This would force us to go back to stretching the service period to six months, or to four months with a two-week stand down three times a year (six weeks out of service). Numerically we could replace one of the three batteries about every port stop, but this becomes logistically impossible. The final evaluation was that more sustained power was available with a two-tank rotation than would be available with a 24% larger battery capacity serviced every six months. Improvements in battery charger control also allowed us to customize the charging profiles to maximize charging and minimize water consumption.
It should be noted that during the battery evolution, the power demands have continually increased, payload demands have continually increased and the average dive time has remained about constant since 1986. The design tradeoffs and efficiency improvements have given more performance with less cost and have maintained the payload capacity. Appendix C illustrates the average dive time over the last three decades for each of the design change periods.
There are two areas which can improve dive duration. The most effective is to maximize the power available and to use it more efficiently. The second is to provide a raw increase in available power. Without doing the first, even the second would not provide satisfactory results. Power efficiency and capacity must consider the whole system, not just the battery. The system also includes personnel performance, battery maintenance and performance monitoring.
Appendix D is a comparison of recent dive statistics for four pilots. This is an average of 30 dives over a similar time period to minimize the effects of different mission profiles between each individual. The average difference between the shortest bottom time and the longest is about 40 minutes, or 15%. Without more data regarding the reason for completion of a dive, it is impossible to determine the reasons for the variability. It is possible that the work was generally completed which suggests the longer bottom time is due to slower work at reduced power, whereas the shorter time is due to faster work at a higher power rate. If the dives were terminated due to lack of power, it would indicate more emphasis should be placed on pilot training to improve their efficiency. In either case, the total health of the battery may have been an issue. To gather better data about pilot efficiency and battery performance, future pilot debrief sheets will include a comment about why the dive was terminated, whether it was lack of power, work completed, weather, equipment failure, etc. If the comment is lack of power, this should give the electrician daily feedback about the condition of the battery and the need for some corrective action.
Maximizing Available Power
Battery maintenance has a major affect on battery performance. Problems with battery chargers not delivering a full charge, insufficient battery equalization, battery age, and electrolyte levels all interact to establish how much power is available on a given dive. When a battery starts to loose capacity, the corrective action usually requires several battery discharge/charge cycles to regain the capacity. Even with the current four month battery rotation, a delicate balance must be maintained between slightly overcharging daily, periodic equalization and battery water consumption. Manufacturers recommend equalizing (periodic overcharging) batteries every 5-7 cycles to maintain their full capacity. These maintenance requirements are not unique to lead-acid batteries. All other vented aqueous type cells need similar care. This also results in the need to add water to the batteries every couple of weeks. Various automatic watering devices have been investigated by the ALVIN group to allow watering without removing the batteries. AR the devices have a large risk involved in the event that one of the devices failed, resulting in an electrolyte overflow within the battery tank, subsequent short circuit, and possible fire or explosion.
To better monitor the battery condition and maintain maximum capacity, the electrician needs daily feedback about battery performance. By the time a pilot makes a comment about the battery performance, the battery has deteriorated to the level that mission performance is noticeably affected. Better instrumentation is required to monitor the daily and gradual change in the battery condition. WHOI is pursuing two courses of action to provide the necessary instrumentation. A prototype microprocessor based electronic circuit which resides within the battery tank has had some initial testing on the submersible. This circuit monitors volts and amps on charge and discharge, battery tank temperature, and electrolyte level in four cells. It has a SAIL data interface to log the battery data in real time. There have been several design refinements to the circuit based on the results of the field trials, and continued testing and evaluation is required to demonstrate reliability before the circuit can be considered operational. ALVIN does not presently have a SAIL system, and refinements to the logging system would be required as well as development of data evaluation software. Another device to be evaluated is a microprocessor based commercial product with an RS-232 interface that measures volts and amps during charging and discharging with respect to time. The data output includes volts, amps, number of cycles, deepest discharge, average depth of discharge, coulomb efficiency (power out/power in), etc. Testing and evaluation of this product will start in late July 1996. If it provides useful and accurate information, and a safety evaluation of the necessary wiring is satisfactory, the device may be incorporated in ALVIN during the next overhaul. In addition to potential use on the submersible, this device may also be used to log the performance of our Exide brand battery chargers which do not have a computer interface.
The overall performance of a system is also a function of the power usage efficiency. In addition to the pilot factors and battery maintenance already mentioned, the power efficiency of individual electrical and electronic equipment should be maximized. A design or selection criteria for any new equipment should include an evaluation of energy efficiency, especially for equipment that will be in continual use during a dive. Video cameras are an example of how of efficiency tradeoffs need to be evaluated. One way to get a better picture is to provide more light on the subject. This increases power requirements. An alternative way is to use a camera that needs less light to get a better picture which reduces the power requirements. Another power conservation method is to periodically assess the need for existing power consumers. This is constantly done on ALVIN and is a natural selection process. Those items that are not frequently used are removed from permanent installation. This helps control weight growth and reduces maintenance. Reducing obsolete or underutilized equipment means less power consumption. In designing the next generation of ALVIN data logger and display systems, an evaluation of what information is actually required in real time, and what is not necessary, will be considered. By reducing the complexity and demand for real time information, weight and energy savings may be possible.
Battery Technologies
In addition to striving for maximum performance and efficiency from the current power systems, the ALVIN Group is continuing to evaluate sources that might provide more power with less weight in the existing space available. Appendix E shows performance comparisons of four battery technologies considered for ALVIN. [Table used to develop Appendix E.] This data is based on production cell dimensional and weight information supplied by the manufacturers and should not be confused with specifications relating to laboratory test data. The total Kwh value is based on the possible installation configurations as discussed below. Three of these batteries represent technologies that are mature and are available in capacities suitable for electromotive applications such as ALVIN. The NiMH cells cannot be considered mature since they have not been broadly applied to electromotive applications but may have future applicability.
Nickel-Cadmium batteries are used in the Russian MIR submersibles to replace the Nickel-Iron batteries that were no longer available. The cells were made by SAFT-NIFE of France and are available in the U.S. through their facility in Georgia. The Russians first started to evaluate these cells in 1989. There was a problem with the battery case material and a different cell was evaluated in 1992. In 1995 they replaced their nickel-iron batteries with the present NiCd's. This suggests they had a five year development and evaluation program before committing to this type of cell. Discussions with the MIR program manager and SAFT indicated the cells are performing well but require watering every 20 dives. In a full ALVIN diving schedule the battery would need to be removed and serviced every month, which would be operationally unacceptable. One of the Russian design requirements was for the cells to take a 30-degree roll without spilling the electrolyte. With an appropriate vent cap, it may be possible to increase the electrolyte volume and extend the watering interval. Presuming this shortcoming could be overcome, two new battery tanks containing 285ah cells (a 50% increase over Douglas) would fill the complete battery compartment and result in a 150-pound loss of payload. Alternatively, two tanks with 208ah cells (a 9% increase over Douglas) could be installed with a 700-pound increase in the payload. NiCd cells are not susceptible to damage by over discharging as opposed to lead-acid batteries which may be damaged if discharged below 80%. This suggests the NiCd cells could be discharged closer to their rated capacity as opposed to the 80% discharge limitation of Pd-acid cells. If one were to presume a 90% discharge for the 208ah NiCd, and a 70% discharge for the 190ah Pd-acid, the net useable power gain would be 41 %, not the 9% suggested by just the amp-hr rating of the cells. For a 260ah Trojan lead acid cell, the above analysis results in only a 3% useable gain. Either of the NiCd battery packs would result in loss of the science space in the present third battery position. The French source for these batteries might present a logistics problem. The only other source of high capacity NiCd batteries identified is VHB Industrial Batteries of Canada, a subsidiary of Varta. Their construction methods result in a cell with 47% fewer watt-hrs/liter and a weight 92% more than the SAFT cell. The SAFT battery cost for three tanks (including one rotating for service) would be approximately $210,000. New battery boxes and base plates would cost approximately $50,000. This is more than a 1000% increase in battery cost for a potential 41%, (or 3% with Trojan cells) increase in power.
A Nickel-Metal-Hydride battery in the 90 ah range is manufactured by Ovonic Battery Co. These batteries are being developed for use in electric cars and provide about 60% more power/weight than the best lead acid cells. The cell is proprietary and details about internal construction are not known. The standard cells are a sealed, no maintenance design which relies on recombining the hydrogen or oxygen generated during discharge/charge cycles. They use an aqueous potassium hydroxide electrolyte so there is a possibility that the cells could be applied to deep ocean applications. However, at the present time it is not known if the sealed nature of the cell is required to retain cell fife or if the chemical reactions will effectively take place at 6,000 psi. The Ovonic cell life is only 600 cycles (compared to 1000+ for other types) and mass production quality has not been established since these cells are not yet generally available. General Motors expects to market vehicles with these cells in 1997. Because this technology is in its infancy, these cells are not appropriate for use in ALVIN at this time. With assistance from high level contacts at GM, VMOI plans to look more closely at the possible deep sea application of these cells. This will probably require a meeting with Ovonic engineers in Michigan to gather more information and start a pressure test evaluation program if appropriate. We have recently received word from VHB/Narta that they may have NIMH cells which are of high enough energy density to be of interest to us. We will be working with them to evaluate the potential for application of their product.
Silver-zinc batteries (Yardney) were investigated. Two sets of 750ah batteries (including one rotating spare) would cost approximately $336,000, could provide a 97% increase in power over Douglas, are light weight (a 1000 pound increase in the payload), have a short lifetime (only 70% capacity guaranteed after 12 months), have a poor reputation for reliability, and only one 120-volt battery could fit in ALVIN, eliminating the safety of having a redundant power source. Overall, these were not considered a viable alternative power source.
Lead acid Chloride Canada tubular plate cells have been a candidate for installation in past years. The most likely cell is a 180ah unit which is fight enough that three tanks could be installed resulting in a net power increase of 42% over the present Douglas cells. There would be a 250-pound loss in the payload. This company also offers a 190ah cell that would require completely new battery boxes. With three tanks installed, there would be a 1000-pound payload loss and a 50% increase in power. In both cases, the third battery space would not be available for science applications and would have all the disadvantages of the three-battery configuration discussed above (six month services, electrolyte depletion, etc.).
Exide has gone through some reorganization in recent years, and two of their current 225ah cells have been purchased for evaluation. Initial cycle testing indicates they do deliver the advertised capacity. Assuming that the Japanese purchase of the company has solved the previously experienced delivery and quality control problems, these are an option for installation in the coming overhaul that could provide a power increase of 18%. A major disadvantage is that they will reduce the payload by 200 pounds. With the increase in science payloads since these were last used, this could have a major act on science capabilities. One possibility for offsetting the payload loss is to add syntactic to the third battery space. That space is being used frequently for science equipment such as altimeters, down looking cameras, lights, strobes, and down looking sonar systems. Syntactic in that area may impact those capabilities.
Testing of some Trojan Battery Co. lead acid cells rated at 260ah is currently in progress. Each of these cells is 0.5 pounds lighter than the existing 190ah cells and has a potential to increase power 37% while increasing payload 60 pounds. The manufacturer claims the cells require 40-50 cycles to achieve maximum power. However, after more than 40 cycles, they are delivering 83%-88% of their advertised capacity, or approximately 215-228ah. When sufficient cycle testing is complete, the test data will be discussed with the manufacturer. If these cells eventually deliver at least the capacity of the Exide cells (225ah), these would be an excellent candidate to replace the existing Douglas cells during the next overhaul. This company is the country's largest supplier of deep cycle industrial batteries for fork trucks, golf carts, commercial floor scrubbers, etc. Being a domestic company, there should be a reduced risk of supply problems in the future.
Summary
As can be seen in Appendix C, there has only been a 3% variation in average dive times since 1986, yet there has been a continual increase in power demand and the total power available has decreased 33%. This efficiency improvement can be attributed to improved power distribution, better battery maintenance through charging improvements and battery rotation, constant efforts to improve equipment efficiency, pilot training to improve efficiency of power utilization, and constant awareness of all involved in the operation and use of the facility to be conscious of power consumption. There is still room for improvement in power utilization and sustaining maximum output from the current lead acid battery technology. Daily charging records indicate the batteries are seldom delivering rated capacity. Additional work is needed to determine the threshold of heavier battery charging without prematurely depleting battery water. This effort requires better logging and record keeping over many operational months. Only incremental changes can be made to determine the point of diminishing returns and not adversely affect our primary mission of supporting science diving requirements. Our continued effort to install suitable instrumentation and adding daily commentary of battery performance should help us improve battery performance in the future.
Of the alternate battery technologies presented, the NiCd cells probably have the highest possibility of successful implementation in the future. Although many arguments based on the life cycle cost of the cells (including dollars/minute on the bottom) can be made, it is very difficult to justify the high initial expense to change to the NiCd cells for a theoretical 3% increase in power (@90% discharge) over the potential of the Trojan lead acid cells (@ 70% discharge). NiCd's are not widely used in the proposed application, and data for long term performance is not known. Contact with manufacturers of these cells indicates they know very little about deep sea applications of their products. For WHOI to commit to the usage of this cell, extensive evaluation and testing would need to be done to establish low temperature, high pressure performance, water consumption rates, reliability, quality, and extended service requirements as well as literature search of historical data, and possible direct consultation with manufacturers and MIR operators. This would be a labor intensive and time-consuming effort, require battery procurement, chargers, test instrumentation, and thus would need support from a funded proposal. There would be no guarantee that the study would result in positive recommendation for use of NiCd batteries.
In the immediate future, Trojan, Exide or Chloride Canada lead acid cells will likely be installed during the next overhaul. With the efficiency improvements since the last use of the Exide cells, an 18% increase in power might be achieved over the existing cells without the need to redesign the battery tanks. Part of the payload penalty may be compensated for by adding a partial block of syntactic in the third battery hole and still leave some room for science equipment installations in that area. If the Trojan test program is satisfactory (>225ah capacity), these cells would be an excellent choice for installation. They have the potential to increase power by 37% without redesigning the battery tank and there should be no need to add syntactic in the third battery bay. If the overhaul indicates there is a substantive loss in the payload due to either science equipment additions (HMI's, four cameras, two VCR's, etc.) or gradual deterioration of syntactic foam the lower capacity (180 ah), fighter, Chloride Canada cells may need to be seriously considered.
The power sources elaborated on in this report are not the only technologies that have been reviewed. Carbon pile hot air turbines, kinetic energy flywheel storage, and numerous specialized battery chemistries and laboratory curiosities have also been reviewed. When all factors of reliability, maintenance, payload, mass production quality, cost and implementation impacts are considered, most of these alternatives have been deemed impractical for current applications.
To better monitor the development of possible future power sources, WH0I will try to improve the dialog with Ovonic and VHB to determine the applicability of the NiMH cells for deep ocean applications. We also plan to attend the annual Power Source Symposium each June to closely monitor power technology developments.
