Energy storage working group report

Kieron Osmotherly

January 05, 2017 04:43 PM

Lively debate characterised the inaugural energy storage working group at the TowerXchange Meetup Africa 2016, where dozens of experienced buyers and sellers of batteries exchanged insights into how to extend lifecycles in challenging operating conditions in Africa and the Middle East.

Working group members, buyers:

Econet Power, Helios Towers Africa, Vodafone, Leadcom, TowerTech

Working group members, suppliers:

4energy, Aquion Energy, Energic Plus, EnerSys, Fluidic Energy, GNB (Exide), GS Yuasa, NorthStar, Redflow, Saft

Energy storage

Working group key learnings

LEARNING 1: There are a lot more lead acid batteries deployed at SSA cell sites than alternate chemistries – lead acid probably still has over 85% market share

LEARNING 2: Buyers felt that vendors of alternate energy storage chemistry solutions needed to do a better job presenting a “full business case, with complete costs and ROI”

LEARNING 3: At a “legacy” African cell site, a lead acid battery might last 18 months. With modernisation and process optimisation, that can increase to 24-30 months, although some MNOs are still seeking five year lifecycles

LEARNING 4: Cell site owners need to optimise the relationship between the rectifier and the batteries in order to maximise equipment lifecycles

LEARNING 5: The process of battery hybridisation of MNO’s African cell sites is nowhere near complete – there is a substantial addressable market, and a substantial replacement system market

LEARNING 6: Towercos’ upgrade of energy equipment acquired at the towers formerly owned by MNOs doesn’t happen overnight – sites acquired several years ago are not all up to spec, while other sites have only recently been transferred

LEARNING 7: Towercos own >40% of Africa’s cell sites, and provide a full DC power service at the majority of those sites – meaning towercos now buy the batteries for almost half Africa’s cell sites, including the majority of sites occupied by Airtel, MTN and Millicom

LEARNING 8: The typical power load on a cell site in Africa is around 2-3kW per tenant, rising as high as 8-10kW for indoor and hub sites. The load on single tenant, low cost rural sites can be as low as 300W-1.5kW

LEARNING 9: The unpredictability yet severity of grid outages in SSA demand significant autonomy, which in turn means battery banks will be ‘lazy’ for long periods

LEARNING 10: Cell site autonomy requirements vary widely across the continent; the lowest we heard was for 4 hours, the highest for 10-20 hours.

LEARNING 11: There may be a case for mixed battery banks including lead-acid and lithium ion batteries to cater to a greater variety of operating conditions, although one buyer suggested <10% of their sites might suit such a configuration

LEARNING 12: Battery theft varies across Africa: we heard a range between “negligible” and “higher than 5%”

LEARNING 13: Integrating monitoring systems into “intelligent batteries” is a promising development, but MNOs and towercos often struggle to translate the volume of data into actionable intelligence

LEARNING 14: Standard battery warrantees are meaningless in challenging grid conditions; a few alternate chemistry energy storage vendors are offering warrantees based on a guaranteed kWh output

Executive summary

Sub-Saharan Africa may represent the world’s toughest operating conditions for energy storage at cell sites. Poor grid reliability means it can be impossible to maintain disciplined charge/discharge regimes. High operating temperatures, combined with a legacy of indoor site configurations, often require that premium high temperature batteries be installed. Premium batteries, particularly those with domestic or light industrial re-use scenarios, become targets for theft. And maintenance regimes and maintenance skillsets aren’t always optimum. This all combines to mean that the lifecycle of a backup battery bank at an African cell site can be as short as 18 months.

TowerXchange’s inaugural energy storage working group convened in Johannesburg in October 2016 to highlight these operating challenges and to challenge buyers and sellers of batteries to work in closer partnership to devise products and processes to enhance battery performance at African cell sites.

Let’s start our report with snapshots of three of the key battery buyers present at the working group.

Helios Towers Africa

Helios Towers Africa has widely deployed charge discharge (CDC) battery hybridisation, mainly using lead-acid batteries, across their 6,556 site network.

Helios Towers Africa’s portfolio spans four countries and a range of vintages, from older portfolios in Ghana (750 sites), and Tanzania (3,500) to newer acquisitions including 400 sites in Congo Brazzaville and the Airtel DRC sites, bringing their count in the country to around 1,600.

Helios Towers Ghana is shifting from AC only to a full DC power service, starting with around 300-400 Millicom sites, which are being converted from indoor to outdoor sites at the same time.

Grid availability in Ghana has returned closer to norms prior to the devaluation of the local currency, the Cedi, which saw availability slide from and average of 22 hours to 14 hours back in January 2015. Grid conditions in Tanzania are also relatively reliable, at just under 20 hours of good grid per day, while in the DRC availability drops to 14-15 hours, although grid reliability falls dramatically (before it disappears altogether) outside Kinshasa, Lubumbashi and Goma. Helios Towers Africa has around 640 off-grid sites in the DRC, where the company are deploying their first 50 site solar trial, and 180 off-grid sites in Congo Brazzaville. A further 800 Tanzanian sites are off grid.

Vodafone

Vodafone were represented by the Network Site Infrastructure team at Vodafone Procurement Company, which is responsible for procurement across over 50,000 Vodafone, Vodacom and Safaricom sites in Africa.

Vodafone are committed to ‘going green’ and reducing carbon emissions, evaluating solutions on a Total Cost of Ownership (TCO) basis, incorporating up front capex, transportation, maintenance et cetera.

Vodafone typically has two to four hours of battery backup power at sites in good grid markets such as those in Northern Europe. Greater autonomy is required in SSA, where high temperature lead-acid batteries are most commonly used, although Vodafone are looking at future energy storage chemistries.

Econet Power

COO Neil Taylor represented in-house ESCO Econet Power, recently carved out from Econet Zimbabwe. While parallel carve-out Econet Towers manages only the towers, Econet Power manages the energy systems at 1,380 sites.

While Zimbabwe’s grid is characterised as “bad everywhere”, downtime had reduced from an average of around ten to just 2-4 hours per day at the time of the Meetup, although the impact of the subsequent currency crisis remains unclear. Just 3.5% of Econet’s sites were completely off grid, and over 90% of those sites had solar installed. Around a quarter of their total site network was hybridised, primarily with lead-acid batteries.

What is the typical power load on a site?

Of course it depends on the configuration of the site. An indoor site could have a load as high as 8-10kW, a hub site even higher, but most sites have a power load around 2-3kW per tenant. Energy efficiency initiatives can bring that down to 1.5-2kW.

Single tenant, rural sites might typically be in the 1-1.5kW range, although we’ve seen as low as 300W.

Outage durations and implications for Service Level Agreements (SLAs)

There is huge variation in outage from country to country and from month to month. “We might have a Nation wide outage for six to eight hours, then have no outages for months,” said Econet Power, commenting on Zimbabwe. While seldom as severe, Tanzania was similarly unpredictable, at least outside of Dar es Salaam.

Rather than try to quantify outages, it is perhaps more pertinent to focus on SLAs and their implications for required cell site autonomy: one MNO typically required 10-20 hours autonomy in their RFPs, while a towerco sought four hours at urban sites and six hours at rural sites.

“A battery hybrid site might need eight hours autonomy and discharge to around 50% during a typical outage,” offered one participant as an example.

If you’re sizing your battery bank for 10, 12 or even 20 hours of autonomy, meaning it’s ‘lazy’ most of the time, does that mean there might be a business case for a mix of lead-acid and lithium-ion batteries in the bank? Helios Towers Africa suggested there might be around 200 of their sites in Tanzania where such an approach could prove viable.

How to extend battery lifecycles?

How long are batteries lasting? “At one of our low cost 500W rural sites, operating in 20-35°C, we’ll schedule the swap of the batteries after three years,” said a representative of TowerTech, a managed service provider with experience of building off-grid sites.

Vodafone typical sought five year lifecycles from their battery banks, while Helios Towers Africa reported that a battery might typically last two and a half years at an air cooled outdoor site operating up to 35°C, although batteries would be replaced sooner if autonomy dropped below a certain threshold, or if low voltage alarms were being triggered early.

“Previously rectifiers were killing batteries in as little as 18 months; energy efficiency isn’t just about batteries,” said one buyer. “In my experience, it is often the rectifier that causes under or over charging, so it’s about how the DG, rectifier and battery work together.”

How bad is the problem of battery theft?

Battery theft is an acute problem in some markets, it can be almost non-existent in others. “We have had no battery thefts since I started with the company,” suggested one participant.

“We have 3-4% battery theft per year,” said another.

“It’s higher than 5% in South Africa,” offered a third participant, suggesting that TCO increased as much as 20% over a five year period due to theft.

It was generally agreed that battery theft was being reduced, and that the problem was not as acute as perceived, although it was acknowledged that the resultant downtime could be disproportionately harmful to reputations.

Helios Towers Africa reported having significant success reducing battery theft originating within the supply chain. Their holistic approach involves splitting O&M contracts between suppliers to create competitive benchmarks, eradicating “bad apples”, investing in access control solutions, process optimisation and training, and simply making sure security guards are fairly paid to reduce temptation to steal. “It’s all about creating accountability. While these steps will initially increase opex, it drops in the long term,” concluded Helios.

What buyers want from their battery suppliers

There is an inherent contradiction in two key forces; buyers have strict guidelines for recycling and re-use of equipment, yet they increasingly want to see re-use scenarios reduced and scrap value minimised to disincentivise battery theft.

Integrating GPS has had minimal effect; one buyer said it “over-complicates matters”, while another pointed out that law enforcement authorities seldom have the capacity to track stolen batteries in an effort to apprehend thieves.

Intelligent batteries which can monitor and protect themselves, and other equipment, may be an important development, but beware creating data overload as MNOs and towercos can struggle to integrate, analyse and respond to thousands of data points from thousands of sites.

Alternate chemistries

“Lithium ion and flow batteries won’t replace lead acid over night,” commented one buyer, asking that vendors focus on best fit use cases initially to enable limited trials (vendors suggested 100 site trials would be better than 10 sites!) The buyer continued: “we need to see a full business case, with complete costs and ROI.”

Vodafone Procurement Company highlighted their process; an initial business case from which TCO can be derived. Next follows an in-house or third party trial. Asked target return on investment, Vodafone simply said “ASAP”, and depended on the size of the investment, although they admitted more than four years would be too long. “There is no magic number for battery ROI,” they concluded.

Fixing battery warrantees

Buyers complained that standard lead acid battery warrantees were meaningless in emerging market operating conditions . “There has to be some limitation with lead acid battery warrantees,” responded one vendor. “Cycles kill lead acid more than temperature.”

A flow battery vendor countered “we’ll offer a guaranteed energy throughput, localised warrantees based on operating temperatures up to 60°C, regardless of depth of discharge.”

A lithium ion battery vendor offered similar terms: a guaranteed kWh output with a maximum operating temperature as high as 60°C.

The lead acid battery vendor countered “alternate chemistry batteries have to be as easy to deploy as lead acid before these factors come into play.”

“What’s needed is a warrantee with a margin for error,” suggested another vendor.

“What matters more than the warrantee is the level and quality of training your product teams provide my service teams, and ultimately our charging behaviour,” concluded a buyer.

Download TowerXchange Meetup Africa & ME 2016 report