BATTERY TECHNOLOGY AND THE FUTURE FOR UPS

Transcription

1 BATTERY TECHNOLOGY AND THE FUTURE FOR UPS Steven Shapiro, PE Mission Critical Practice Lead 1

2 Data Center World Certified Vendor Neutral Each presenter is required to certify that their presentation will be vendor-neutral. As an attendee you have a right to enforce this policy of having no sales pitch within a session by alerting the speaker if you feel the session is not being presented in a vendor neutral fashion. If the issue continues to be a problem, please alert Data Center World staff after the session is complete. 2

3 Batteries have been the magic power behind supporting critical loads. This session will review the battery technologies available today, in the near future and the distant future relative to UPS applications. What, when, how and why to use each technology supported by performance, physical footprint, initial cost, life cycle cost and disposal data. 3

4 Agenda Review Basic Battery Concepts Review Battery Technologies Old New Future Review of Codes and Room Requirements Total Cost of Ownership UPS Old New Future Questions 4

5 Basic Battery Concepts Battery Charge and Discharge 5

6 Flooded Lead Acid

7 Basic Battery Concepts Lead Acid Flooded Valve Regulated (VRLA) Absorbed Glass Mat (AGM) Gel Cells Pure Lead (Lead-Tin/Thin Plate) Lithium Ion Salt New Tech- Solid State 7

8 Basic Battery Concepts Issue Flooded Wet Cell 20 Year VRLA - AGM 10 Year VRLA Reliability * * (Failure Rate) * 2.67 Times The Flooded Cell Rate 2.67 Times The Flooded Cell Rate Shorted Cell Fails Open Fails Open Failure Mode Results in lowered string voltage. String Makes the string useless. Possible for still provides power in an event. cell to fail violently in thermal runaway. Makes the string useless. Possible for cell to fail violently in thermal runaway. Larger than 20Yr VRLA Larger than 10Yr VRLA Smallest footprint Footprint 72 WX2 D 2 Tier,144 SF 18 Wx5 D 5 Cabinets of 4 Wx3 D=> 20 Wx3 D 48 WX2 D 3 Tier,96SF 90SF/Module 60SF/Module Module Only 7 minutes Max Moderate Minimal Minimal Maintenance Electrolyte level must be maintained above the plates. Quarterly Semiannual Semiannual $5,600/Year/String $2,800/Year/String $2,000/Year/String 20 year design, 20 year design, 10 year design, Life Cycle 12 to 15 year actual 11 to 13 year actual 3 to 5 year actual 20 Year Scope 1 Replacement 1 Replacement 3 Replacements 8

9 Basic Battery Concepts Flooded Wet Cells Positive Plate PbO 2 Negative Plate Pb Acid H 2 SO 4 9

10 Basic Battery Concepts Flooded Major Modes of Failure Positive Grid Corrosion Loss of Active Material Internal Shorts C&D Technologies C&D Technologies Grid Thickness (inches) 10

11 Valve Regulated Lead Acid

12 Basic Battery Concepts VRLA Gel Cell - Gel to Suspend Electrolyte, Often Called Sealed Lead Acid Advantages: 1. No liquid electrolyte to spill or leak 2. Can be Deep Cycled several time without damage 3. Totally corrosion & maintenance free 4. No/Minimal off gassing 5. Typically Shorter Life Disadvantages: 1. As it ages capacity fades 2. Fails Violently 12

13 Basic Battery Concepts VRLA AGM, Absorbed Glass Mat, Glass Mat to Suspend Electrolyte Advantages: 1. No liquid electrolyte to spill or leak 2. Can be Deep Cycled several time without damage 3. Totally corrosion and maintenance free 4. More plate surface and closer plate spacing provides a compact case size 5. Typically Longer Life 6. No/Minimal Off gassing Disadvantages: 1. As it ages capacity fades 2. Fails Violently 13

14 Basic Battery Concepts VRLA Pure Lead Thin Plate Pure Lead - TPPL Advantages: 1. Longer Shelf Life And Warranty Than Typical VRLA, 5 Year Warranty, 12 Year Design Life, Actual 7-12 years in practice 2. Higher Temperature Operation Than Typical VRLA 3. Shorter Discharge Times Than Typical VRLA 4. Better TCO than other VRLA Disadvantages: 1. Shorter Discharge Times Than Typical VRLA 2. More Expensive than other VRLA 14

15 Basic Battery Concepts VRLA Lead Calcium Vs Pure Lead (TPPL) - CYCLE LIFE 15

16 Basic Battery Concepts VRLA Major Modes of Failure Loss of Element Compression Internal Shorts Thermal Runaway Negative Plate Undercharge 16

17 Basic Battery Concepts VRLA Loss of Element Compression If the gases formed during recharge are allowed to leave the cell, the water in the electrolyte will be consumed In flooded cells, this would require water addition. In VRLA, the water amount is fixed and cannot be replenished. Loss of water will result in physical contraction of the separator glass mat, resulting in loss of electrical contact between plates Undersaturated, shrunken glass mat Severe cases interrupt the electrical path, and will prevent acid replenishment to the plates Fully saturated glass mat C&D Technologies 17

18 Basic Battery Concepts VRLA Internal Shorts VRLA cells have limited acid volume - if discharged to a very low voltage, the acid concentration drops to the point that lead will dissolve into the electrolyte During the next recharge, the dissolved lead will be converted back into lead metal in the glass mat separator causing electrical shorts Glass mat separator C&D Technologies 18

19 Basic Battery Concepts VRLA Thermal Runaway VRLA cells operate hotter than flooded batteries Recombination reaction generates heat (as O2 ==> H2O) There is less thermal mass to absorb the heat There is lower heat transfer to the outside case Particularly at high temperatures and high float voltage, the cell may produce more heat than it can dissipate - at this point, it will go into thermal runaway and can destroy itself. 19

20 Basic Battery Concepts VRLA Negative Plate Undercharge Effective Recombination Reaction: The recombination reaction of oxygen at the negative plate will depress (depolarize) the negative plate voltage Because of this, the float current may not maintain the negative plate at a full state of charge (even though VRLA batteries require a higher float current than flooded) Solutions to the undercharged negative plate situation include: Periodic boost charges - (typically > 2.40 volts/cell) Internal catalyst [An additional benefit of an internal catalyst is to reduce the overall float voltage, which will reduce the positive plate polarization. This will prolong the life of the cell.] + - O 2 H 2 O C&D Technologies 20

21 Basic Battery Concepts 21

22 Basic Battery Concepts Gas Evolution Overcharge Reaction After cells are fully charged, the positive plate evolves oxygen gas and the negative plate produces hydrogen gas Positive plate: 2H2O Û O2 + 4H+ + 4e- Negative plate: 4H+ + 4e- Û 2H2 In flooded cells, these gases must be exhausted to maintain H2 levels below 4% O 2 H C&D Technologies 22

23 Basic Battery Concepts Gas Evolution (Cont.) In VRLA cells, the evolved gases are contained within the cell Oxygen travels through the separator to the negative plate - where it recombines with a hydrogen ion to recombine into water Recombination reaction depresses voltage of negative plate Recombination reaction gives off heat Glass mat separator + - O 2 H 2 O Positive plate 2H 2 O ==> O 2 + H + + 4e - Negative plate 2Pb + O 2 ==> 2PbO 2PbO + 2H 2 SO 4 ==> 2PbSO 4 2PbSO 4 + 4H + + 4e - ==> 2Pb + 2H 2 SO 4 C&D Technologies 23

24 Lithium Ion

25 Basic Battery Concepts Lithium Ion Power Cells (Power ) Vs Energy Cells (Energy Storage) Battery Management System (BMS) required UL listings (Cell 1642, Cabinet/BMS 1973) 25

26 Basic Battery Concepts Lithium Ion Battery Reaction : LiMn2O4 + C6 Li1-xMn2O4 + LixC6 Charge Test Description Criteria Pass Short Circuit Short of <100mOhms Max temp <150C, no explosion or fire Short Circuit (55C) Short of <100mOhms, 55C Max temp <150C, no explosion or fire Pass Pass Discharge Abnormal Charge Charge with 3x maximum specified charge current to maximum target charge V (4.5V) No fire or explosion Pass Charge Li + electrolyte Crush Cell crushed between flat plates up to force of 3000lb (13kN) No fire or explosion Pass Impact 15.8mmbar dropped from height of 61cm onto longitudinal axis, with 9.1kg mass No fire or explosion Pass CAN Shock 125 to 175g magnitude shock in each of 3 perpendicular directions No leak, fire or explosion Pass Li + Vibration 10 to 55Hz vibration at 1Hz/min increase and back down across each 3 perpendicular directions No leak, fire or explosion Pass Discharge Cathode + Anode - Al foil Active Material (LMO base) Active MaterialCu foil (Graphite base) Anode - intercalation graphite Cathode - Lithiated Manganese Oxide (LMO) Electrolyte - Lithium hexafluorophosphate, organic carbonates No gassing on charge/discharge Heating Heat at 5C/min to 130C for 10 minutes No fire or explosion Pass Temp Cycling 70C to -40C cycling 4 hours at each temperature, repeat 9 times < 0.1% mass loss Pass Low Pressure 6 hours at 11kPa (1.68 psi) No leak, vent, fire or explosion Projectile Cell placed onscreen above burner according to UL specified setup until explosion or burned out Cannot penetrate mesh screen Pass Pass 26

27 Basic Battery Concepts Handheld electronics mostly use batteries based on lithium cobalt oxide (LCO), which offers high energy density but presents safety risks, especially when damaged. Lithium iron phosphate (LFP), lithium ion manganese oxide battery (LMO) and lithium nickel manganese cobalt oxide (NMC) offer lower energy density, but longer lives and inherent safety Source: Wikipedia 27

28 Basic Battery Concepts Type Cathode Anode Electrolyte Lead Acid Lead Dioxide Lead Sulphuric Acid Li-ion Metal Oxide Carbon Based Lithium Salt 28

29 Basic Battery Concepts Manufacturers for UPS Applications LiiOn First to Deploy 2014 Samsung Rise in Market LG UPS application in Korea and looking to US Saft New to UPS Application Enersys New to Lithium Not available yet 29

30 Basic Battery Concepts Lithium Ion Lowest TCO Smaller Lighter Footprint Lower Maintenance 50% Plus Less Temperature Sensitive Totally Sealed In Multiple Layers No Ventilation Computer Managed (BMS) For Status, Performance And Safety Order Magnitude Better Cycle Life Versus VRLA Up To Four Times Longer Shelf Life Before Recharge Becoming Mainstream As Electric Cars Use Only Lithium 30

31 New Technology

32 Basic Battery Concepts Sodium Based - Not Ready for UPS Applications No Spill Issues Store And Operate In Any Temperature Stored At Any Charge Level No Deep Discharge Issues No Shipping Issues Like Lithium Ion Solid State Batteries - Ceramics/Crystals Not Ready for UPS Applications 32

33 Basic Battery Concepts Battery Monitoring Cell Jar System Loading Impedance Networked 33

34 Codes

35 Codes and Room Requirements NFPA 70E Concept of Safety, Not Highly Detailed Proper Lighting Spill Containment/Control Kit Eye Wash Hydrogen Detection/Ventilation PPE Signage Terminal Barriers International Fire Code/Uniform Fire Code Fire Protection OSHA EPA IEEE 450 and

36 Total Cost Of Ownership (TCO)

37 Total Cost of Ownership (TCO) TCO Incorporates Cost of Energy ($/kwh) and Cost of All Resources Throughout The Life of Facility 37

38 TCO - Flooded vs. VRLA Example 100,000 SF white space 150 watts per SF 15 MW UPS load 20 year analysis 7% interest rate 15 minutes backup time Flooded vs. VRLA 20 year vs. VRLA 10 year batteries 5 2N systems Total 10 systems of (5) 750kVA/675kW modules 50 strings of batteries 38

39 TCO - Flooded vs. VRLA Load Systems Modules Day 1 5 MW 4 systems 20 modules Year MW 2 systems 10 modules Year MW 4 systems 20 modules 39

40 TCO - Flooded vs. VRLA Issue Flooded Wet Cell 20 Year VRLA 10 Year VRLA Reliability (Failure Rate) Failure Mode * * * 2.67 Times The Flooded Cell 2.67 Times The Flooded Cell Rate Rate Shorted Cell Fails Open Fails Open Results in lowered string voltage. String still provides power in an event. Makes the string useless. Possible for cell to fail violently in thermal runaway. Makes the string useless. Possible for cell to fail violently in thermal runaway. Larger than 20Yr VRLA Larger than 10Yr VRLA Smallest footprint 5 Cabinets of 4 Wx3 D=> 72 WX2 D 2 Tier,144 SF 18 Wx5 D Footprint 20 Wx3 D 48 WX2 D 3 Tier,96SF 90SF/Module 60SF/Module Module Only 7 minutes Max Moderate Minimal Minimal Maintenance Electrolyte level must be maintained above the plates. Semiannual Semiannual Quarterly $5,600/Year/String $2,800/Year/String $2,000/Year/String Life Cycle 20 year design, 20 year design, 10 year design, 12 to 15 year actual 11 to 13 year actual 3 to 5 year actual 20 Year Scope 1 Replacement 1 Replacement 3 Replacements Budget Cost Material $550,000 $662,000 $366,000 40

41 TCO - Flooded vs. VRLA Maintenance Cost Annual Flooded Cell - $5, Year VRLA - $2,800, 10 Year VRLA - $2,000 $90,000 $80,000 $70,000 $60,000 $50,000 $40,000 $30,000 $20,000 $10,000 $0 Annual Battery Maintenance Costs Flooded Cells 41

42 TCO - Flooded vs. VRLA Space Requirements Space savings generated by the physical size differences between the various systems Flooded cell - 96 SF x 5 modules x 10 systems x 3.33 = 15,984 SF 20 Year VRLA 90 SF x 5 modules x 10 systems x 3.33 = 14,985 SF 10 Year VRLA 60 SF x 5 modules x 10 systems x 3.33 = 9,990 SF There is a 999 square foot savings by utilizing the 20 year VRLA cells over the flooded cells 42

43 TCO - Flooded vs. VRLA Cash Flow Tables Flooded Cell Yearly Cash Flow 20 Year VRLA Yearly Cash Flow 10 Year VRLA Yearly Cash Flow Year Equipment Maintenance Equipment Maintenance Equipment Maintenance Yearly Cost Yearly Cost Yearly Cost Cost Cost Cost Cost Cost Cost 1 $11,400,000 $112,000 $11,512,000 $13,540,000 $56 $13,596,000 $7,470,000 $40,000 $7,510,000 2 $115,360 $115,360 $57,680 $57,680 $41,200 $41,200 3 $118,821 $118,821 $59,410 $59,410 $42,436 $42,436 4 $122,385 $122,365 $61,193 $61,193 $43,709 $43,709 5 $126,057 $126,057 $63,028 $63,028 $8,659,777 $45,020 $8,704,798 6 $129,839 $129,839 $64,919 $64,919 $46,371 $46,371 7 $7,010,281 $133,734 $7,144,015 $8,326,246 $66,867 $8,393,113 $4,593,579 $71,643 $4,665,222 8 $212,817 $212,817 $106,409 $106,409 $73,792 $73,792 9 $219,202 $219,202 $109,601 $109,601 $76,006 $76, $225,778 $225,778 $112,889 $112,889 $10,039,055 $26,095 $10,065, $232,551 $232,551 $116,276 $116,276 $81 $81 12 $16,253,674 $239,528 $16,493,202 $19,304,802 $119,764 $19,424,566 $10,650,434 $55,369 $10,706, $449,318 $449,318 $19,883,946 $41,119 $19,925,065 $5,484,973 $142,576 $5,627, $462,797 $462,797 $231,399 $231,399 $146,853 $146, $17,760,829 $87,246 $17,848,075 $238,341 $238,341 $21,094,870 $90,755 $21,185, $490,982 $490,982 $245,491 $245,491 $155,797 $155, $505,711 $505,711 $252,856 $252,856 $11,205,002 $64,188 $11,269, $520,882 $520,882 $260 $260 $165,285 $165, $536,509 $536,509 $268,254 $268,254 $170,243 $170, $552,604 $552,604 $276,302 $276,302 $175,351 $175,351 43

44 TCO - Flooded vs. VRLA Over 20 year period, the flooded cell is the most cost effective system due to fewer replacements though higher initial and higher maintenance costs impact yearly cash flows making VRLA systems more attractive in 1st and 12th year Annual Total Costs for Each Battery Type TOTAL Flooded Cell 20 Year VRLA 10 Year VRLA 44

45 TCO - Flooded vs. VRLA Recommendation Best TCO Size And Space For Flooded Cells Install VRLA 10 Year Systems On Day 1 Replace VLRA With Flooded Cells At End of Useful Life 45

46 TCO - Lithium Ion-VLRA 46

47 UPS UPS with Transformer or Transformer Free Efficiency Ground current issues UPS Modules with internal redundancy Internal redundancy but same single input/output UPS System Static Switch Module level System level UPS Technology Double Conversion Offline Line interactive Delta Conversion Rotary DRUPS Diesel Rotary Hybrid Low Voltage Medium Voltage 47

48 3 Key Things You Have Learned During this Session 1. Lead Acid Battery History 2. Application of Lithium Ion Batteries 3. Complications of Various Battery Technologies when Applied to UPS 48

49 The Data Center Data Center Incident Reporting Network dcirn.org A Non-Profit Organization that is Free to Join for the Data Center Industry 49

50 The Data Center Data Center Incident Reporting Network Make Data Centers More Reliable and Safer by Sharing Knowledge Membership is free. Funded by Industry Sponsors Unbiased, Confidential and Secure Reporting Sharing Knowledge with the Data Center Community 50

51 The Data Center Data Center Incident Reporting Network Root Cause Investigation Findings Of Data Center Failures Are Secret And Bound By NDA Consequently The Data Center Industry Is Not Learning How To Avoid Recurrent Failure Modes Data Center Failures Are Generally Non-fatal, Unlike Other Industries Such As Aviation This Will Probably Change As Human Dependency On IT, Particularly AI Increases Currently The Data Center Industry Is Not Effectively Regulated 51

52 Confidential Reporting and Report Production Process 2 Stage Encrypted Dis-Identification 1. Incident Report sent to Secretariat 2. Secretariat Authenticates 3. Secretariat reviews report for anomalies and dis-identifies 4. Dis-Identified report passed to Advisory Council subject matter experts Technical Analysis Final Review 5. Advisory Council analyze report and determine if suitable 6. Advisory Council prepares Draft Incident Bulletin 7. Secretariat reviews Draft Bulletin for Anomalies 8. Secretariat publishes Bulletin to DCIRN members 52

53 The Data Center Data Center Incident Reporting Network What Failed? Why Did It Fail? How Can it be Prevented From Happening Again? dcirn.org 53

54 Thank You Steven Shapiro, P.E., ATD Mission Critical Practice Lead (914) Veriv, Schneider, C&D Technologies and Enersys images and information are included in this presentation 54