| IPP89 Batteries and PV Systems As many of us who are in the business or live with offgrid PV systems know, battery management and maintenance is an area of concern. Many of the user problems associated with offgrid PV systems can be traced to improper treatment and miss understanding of battery performance. From my experience, the most common battery problem is undercharging, leading to sulfation, loss of storage capacity and shortened service life. Sandia National Laboratories recently published “PV Hybrid Battery Tests on L-16 Batteries”. See Access. These tests represent several years of systematic testing of a PV- hybrid system. The Sandia report is very thorough. Four different brand batteries were tested. Tests were repeated so that the data represents good averages. The conclusions in the report are based on good data and methodology. There are four conclusions of the study. They are: 1) The finish voltage (sometimes called the bulk voltage) for a 12 volt battery should be about 15.3 volts (2.55 per cell) rather than the customary 14.4 volts; 2) Finish charge time should be at least 3 hours and often longer; 3) The maximum interval between finish-charges should be about five days; 4) Not all batteries are the same. (Though the reports names no names) The general conclusions of the Sandia report are consistent with the number one problem experienced in fielded offgrid PV systems; under-charged batteries. Richard Perez has for many years advocated higher finish voltages for PV-Gen hybrids. As he says, "I like to run them hot." Finnish Charging is Inefficient There is one downside to the battery management strategy presented in the Sandia report. Due to battery charging characteristics, efficiency is very low during the finish charge phase. This fact resulted in very long generator run times being reported, sometimes from 6-20 hours. These long run times were required to completely refill the batteries to the manufacturers stated ampere-hour capacity. The state of charge (SOC) of a battery is measured with a hydrometer and is indicated as specific gravity (SG). During the Sandia tests full batteries had a SG in the range of 1.290. The long generator run times needed to achieve this SG translate into dollars and pollution (both audio and atmospheric). Perhaps there is a "middle way". A way that preserves the lifetime of the batteries while reducing the time and cost of generator finish charging. Revisit the Assumptions The batteries tested at Sandia were discharged by 60% of capacity and then charged back to rated capacity. In these tests the rated capacities were determined empirically and in most cases were close to the manufacturer's stated value ( in the range of 350 AH for an L-16). These two points require comment. First, this depth of discharge is not typical of most well designed standalone PV systems. This point is clearly stated by the author. Most standalone PV systems, by design, cycle batteries by about 25% daily, not 60%. Secondly, the manufacturer's rated battery capacity and the way it is determined should be understood. All manufacturers recharge batteries on the grid. Using the grid they can finish-charge the batteries for long times (on the order of 8 to 12 hours), cramming maximum ampere-hours (AH) of juice into them. For a manufacturer this method makes sense. The capacity of a battery, determined by this method, gives the greatest AH capacity. The long generator run-times required by PV hybrid systems must mimic the finish-charge conditions the manufacturers use to rate the battery's capacity. Perhaps batteries should be rated based on their application. For instance, a battery used in a stand-by application, (such as backup power system with grid recharging), might specify a full charge SG of 1.290. The same battery used in a cycling application, (such as a PV system with generator backup), might have a recommended SG of 1.250 to be considered full. It is true that a battery with a SG of 1.290 holds more charge than the same battery with a SG of 1.250. However, the shorter finish charge time required to achieve the lower SG reduces the generator run-time. Keep in mind that almost 100% of the energy consumed by the generator is wasted during the latter stages of finish charging. A Thought Experiment Consider two systems. System A would be designed and operated along the lines of Sandia's test. It might be characterized as an undersized PV array with batteries that are cycled deeply (50%) and recharged to the manufacturer's recommended SG of 1.290. System B would provide the same functional capacity (daily AH consumed, in this case 350 ah) but be designed so that the batteries were only cycled by 25% and the generator finish charging takes the batteries to a SG of 1.250. According to Sandia’s findings, system A would need about 6 hours of finish charge every five days due primarily to the deeper cycling. Because system B is only discharged by 25% and the target SG is lower, it would require less finish charge time of about 3 hours every five days. Over a ten year period the difference between the two generator run times is about 2160 (A-B= 4320-2160) hours. A fair assumption (I think conservative) is that it costs a dollar per hour to operate a generator. The ten-year saving of system B over A ($2160) is reduced, however, by the fact that B’s battery set is twice as expensive. If both system, A and B, used L16 batteries, the respective capacities in this comparison would be 700 AH (A) and 1400 AH (B). System B would have an initial battery cost about $800 more than A (based on $200 per L16). Subtracting the increased battery cost from the generator savings ($2160-$800) gives a net savings by system B of $1360. Modified Conclusion Finish (bulk) voltage needs to be about 15.3 volts (for a nominal 12 volt system). Batteries need to be fully charged (finished) about once a week. If the depth of discharge is moderate and a modest SG of 1.250 is chosen, the finish charge time can be reduced from 6 hours to 3 hours. These choices will reduce the generator run costs. On a life cycle cost basis, the reduced generator runtime more than pays for the cost of a larger battery bank. Other benefits include reduced local air pollution, longer generator life, reduced noise pollution, reduced battery watering and maintenance. Don't Undersize the Array Though Sandia's tests were specifically done on L16 batteries, the results are generally true for all lead antimony flooded cell batteries based on my field experience. In very general terms, we can say that the finish charge time is inversely related to the average state of charge. Finally, one of the easiest ways to increase battery life, in addition to limiting the depth of battery discharge, is to add more PV to an existing array. Doing so increases the average state of charge and reduces the need for long generator assisted finish charges. The Aftermath Last November Underwriters Laboratories (UL) withdrew its listing for the Xantrex SW series inverters and posted a public safety alert and press release on its web site. Though the SW inverters are now re-listed and Xantrex is taking measures to upgrade all affected units, UL’s unprecedented, heavy handed and punitive behavior leaves many questions. Why would a company (UL) treat one of its clients (Xantrex) so poorly? Why didn’t UL post a public notice stating that Xantrex SWs were re-listed once re-testing demonstrated compliance? I scoured the UL web site looking for a statement of re-compliance and found none. Was there ever a real safety issue involved? Was this public “shaming” of Xantrex and the damage to their reputation commensurate with the severity of the problem? I do not have detailed answers to these questions. However, UL’s pattern of behavior and anecdotal comments from those who have worked closely with UL, suggest an attitude of arrogance and tyranny prevails there. A Short History UL is not the only US testing agency, though it is the oldest. Founded in 1894, UL touts itself as holding "the undisputed reputation as the leader in U.S. product safety and certification." UL was able to attain that reputation, in large part, by maintaining a near monopoly of the certification business. Prior to 1983, only two testing organizations were authorized by the Occupational Safety and Health Administration (OSHA) to certify electrical products nationally for safety. They were UL and Factory Mutual Research Corporation (FMRC). In 1983 a private testing company, Electrical Testing Laboratories (ETL- now ETL Semko) sued OSHA, since under federal law OSHA enforced safety regulations and technical standards. As a key element of that lawsuit's settlement, OSHA set up the Nationally Recognized Testing Laboratory (NRTL) program, breaking UL’s near century long monopoly. Since 1988, (it apparently took five years to establish NRTL), over 20 companies have been "recognized". OSHA recognizes a company based on an evaluation of the company's ability to perform a specific test. OSHA does not set the standards for testing. Rather, OSHA determines whether or not a company has the technical, staffing and administrative resources to conduct a specific test. If this is the case, that company becomes an NRTL for that test. Quoting from OSHA's web site; "The NRTL determines that specific equipment and materials ("products") meet consensus-based standards of safety to provide the assurance, required by OSHA, that these products are safe for use in the U.S. workplace. Given that each NRTL has met the same requirements for recognition, OSHA considers NRTLs recognized for the same product safety test standard to be equivalent in their capability to certify to that standard." There are Choices Today OSHA recognizes several electrical testing organizations, including the well-known UL, ETL and the Canadian Standards Association (CSA). All three are NRTLs and legally equivalent. In large part, UL's continued dominance of the testing market today is based on almost a century’s momentum gained from their near monopoly of the electrical testing and certifying business. Consumers have grown to accept UL almost as a quasi-governmental agency. It is this expectation by the consumer that UL is the only “official” certification mark that influences many manufacturers to go with UL rather than ETL or CSA. Though engineers working in RE have indicated to me they prefer working with ETL and CSA, they end up working with UL for the sole reason that their marketing departments feel constrained to go with UL, fearing consumer rejection if a lesser known but legally equivalent testing agency were used. Clearly, customer attitude (or at least perceived attitude) is the point of inflection here. Once customers understand and accept that other testing agencies exist and that they can provide legally equivalent testing services, manufacturers may choose to have their testing done by a competitive agency other than UL. National Net-metering A national net metering law is part of the national energy legislation currently in the United States Senate. The current federal net-metering proposal has an upper size limit of 500 kW and stipulates a simplified interconnection process making it very similar to California’s latest net-metering law. However, the outcome is very uncertain. Though there is wide popular support for net metering, the expected opposition from utilities is strong. The fact is that there will be considerable horse trading before anything is passed. Net metering in California is fairing well for the moment. Several initially separate efforts have been woven together. Past articles have discussed the Independent Clean Energy Tariff (ICE-T). This proposal, put before the California Public Utilities Commission, sought to eliminate stand-by and other utility interconnection charges for PV systems up to 1 mW that were not net-metered. Subsequently AB29, California’s 1 mW net-metering law was passed last year, though it sunsets at the end of this year if not re-legislated. Most recently, August 2001, Ken Adelman filed a complaint with the CPUC challenging PG&E’s refusal to interconnect his 30 kW system in spite of the expanded net-metering law. Overall things look good but there are still challenges. Though ICE-T has been approved and AB29 is the law, the utilities are continuing to create barriers in the form of foot dragging and CPUC challenges. Adelman’s case is a good example. He was charged a $7,500 interconnection study fee followed by a $605,000 distribution system upgrade fee. The disposition of Adelman's case, ICE-T and AB29, all three having favorable initial legal rulings, now depend on a final ruling by the CPUC regarding interconnection fees and process. A draft ruling is currently receiving comments and a final ruling is scheduled for late March, 2002. But, all these successes are moot if the California legislature does not pass legislation extending AB29 and its one-megawatt system size cap this year. This may be a struggle since the perception of some is that California's energy crisis has been solved. Not! Access: Background information on NRTLs, Christopher Freitas, OutBack Power Systems, Inc, cfreitas@outbackpower.com, http://www.outbackpower.com OSHA, http://www.osha.gov/dts/otpca/nrtl UL, http://www.ul.com CSA, http://www.csa.ca/ ETL, http://www.etlsemko.com Quarterly Highlights of Sandia’s Photovoltaics Program, “PV Hybrid Battery Tests on L-16 Batteries”, Photovoltaic Systems Assistance Center MS0753, Sandia National Laboratories, PO Box 5800, Albuquerque, NM 87185-0753, 505-844-3698, 505-844-6541 Fax, pvsac@sandia.gov, |
||