Using Data Analytics for Standby Battery Replacement and Budgeting Decisions

Introduction

Standby batteries are widely used as reserve power for continuous operation in power utility, telecommunication, data center, cable, broadband, railway, and renewable energy industries.  They provide a highly reliable and secure power source when commercial power outages occur.  Standby batteries are also used in energy storage applications such as solar, wind, and hybrid power systems. Whatever the technology, Lithium-ion, Lead Acid, or NiCad, batteries are widely used in conjunction with Uninterruptible Power Supply (UPS) systems, Substation 125Vdc Supply and 48Vdc Power systems to protect critical IT equipment, utility switchgears or telecom network elements.

Batteries are cells, but they also have electronics around them that can generate some relevant data sampled near real-time.  The battery life can span over 5, 10, or up to 20 years, which means that significant and essential data are generated.

Battery Replacement Strategies and Budgeting

Budgeting allows businesses to align their goals to meet network needs while mitigating the risks related to network reliability.  Budgeting should also account for network growth, modernization programs, and replacement (risk) plans.  Every year, approximately four to six months before the end of the year, the budget related to power equipment must be presented for pertinent justifications. In this regard, for many businesses, a significant portion of the power budget is associated with the standby battery replacement costs. 

Strategies for common standby battery replacements are mostly considered reactive rather than proactive. In critical power systems such as 48Vdc plants, stations’ 125Vdc supply and Uninterruptable Power Supply systems, the literature references that a battery should be replaced when it has reached 80 percent of its rated capacity.  Despite this, most businesses will replace the battery bank when it reaches a certain service life (age) or when the battery fails to perform as expected, which leads to loss of service and revenue. 

However, businesses with good preventive maintenance programs, periodic testing, and a monitoring system will produce a large amount of battery and cell data, which are very useful for condition-based monitoring and for understanding the battery state of health.  Staff with battery competencies can then use these data to identify, track, and plan when the battery needs to be replaced, thereby significantly reducing the risk of losing service. At the same time, this type of visibility allows people in charge of finances to make better budget decisions.

The Power of Data Analytics

As most networks rely on a large number of batteries of various types, ages, capacities and conditions that are used in diverse standby operational applications, the only way to really understand and efficiently manage batteries is to collect the manufacturer’s data along with how they operate and how they perform. This should be done on as many batteries as possible to analyze the data and to gain empirical insight.

It does not matter if the data are available from manual periodic routines, permanent battery monitors, or embedded BMS. The important aspect is that the battery (cell) data are collected correctly and consistently over time.  The data must be aggregated consistently throughout the battery lifecycle and preferably in a standard centralized database for easier transformation, analysis, and reporting.

Multitel’s Atlas is a centralized asset inventory solution for condition-based monitoring with a predictive analytics software tool that enables the data to be crunched faster.

As Multitel is transitioning its solutions from FIRM’s BMM to Atlas, more flexibility will be provided to power engineers, planners, and program managers to adjust or design their own SoH calculations based on current and historical data throughout the entire battery lifecycle.

How Battery Data Analytics Saves Time and Money While Ensuring Network Reliability

One of the most recognized battery parameters associated with the SoH is derived from ohmic testing. A deviation of the impedance, conductance, or resistance measurement from a baseline reference has proven to represent how much life remains in a cell and can then be used with other data analyses to ensure an optimal battery replacement decision.

For example, the following battery data analytics can be generated and trended in Atlas. Intelligent notifications can be created based on business rules or exceptions that are detected.

Key Metric

Purpose

Best Practices Usages

Cell Ohmic Average (%)

The goal is to analyze the measured Ohmic value of each cell against the Ohmic average value of the battery string to identify poor/bad cells within a string.

  • Further investigation should be performed on cell(s) with more than 30% variation from the average.
  • Using above Cell Ohmic Pass/Fail data can indicate which cell or string needs to be replaced or can remain in the network.

Cell Ohmic Change

The goal is to analyze the measured Ohmic value against previous measurements to identify cells about to fail or as a sign of improper testing methods.

  • Cells with more than a 5% to 10% change from its previous measurement can indicate that the cells may fail soon when tested at quarterly intervals.
  • Cells near end-of-life will tend to vary at a much higher rate than brand-new cells.

Cell SoH% (Ohmic Delta)

The goal is to analyze Ohmic value of each cell against the Ohmic baseline reference to identify the aging or state of health for each cell within a string.

  • Further investigation should be performed on cells with more than 30%-50% variation from the baseline reference value.
  • Consider replacing defective cells or the whole string within one year if more than 20% of cells have failed.

String SoHAvg or SoHMin

Based on the Cell SoH% a SoH value can be calculated by averaging all the Cell SoH% or by using the lowest Cell SoH% value from the string.

  • Consider using the (SoHAvg) when you have more than four cells or jars and use the lowest SoHMin when you have four jars per string.

Battery Reserve Time Estimation

The goal is to determine if the installed battery capacity meets the required reserve time for the application.

Track and use the DC System load current to estimate the battery usage level (%) and reserve time (hh:mm) in near real-time.

Cell Float Voltage

System float voltage should be spread equally between all cells within a string and maintained within the manufacturer’s recommended values. Significant variation from previous measurements, should trigger verification and corrective measures.

  • Consider providing remedies for cell(s) that have float voltage variation of 1% and more than average cell float voltage.
  • Higher or lower system float voltage settings will accelerate battery degradation.

Strap  resistance (Inter-cell)

Highly resistive strap connections may induce charging issues resulting in accelerated sulfating and aging.

  • A variation of more than 20% in strap resistance should trigger a clean-up and retorquing of the strap.

Cell temperature

High ambient temperature is the most important factor that influences battery aging and can cause premature battery failure.

Higher temperatures mean a faster chemical reaction inside the battery, which increases water loss and corrosion and can lead to thermal runaway.

  • For every 10°C constant increase in temperature above this recommendation, it is generally accepted that battery service life will halve (reduce by 50%).
  • 15oF difference between the cell and ambient temperature should require immediate attention.

Battery Float Current

Higher float current value than normal can be an indication of a float voltage set too high, operating at higher temps than usual, presence of a ground fault on a floating battery system or it may indicate failing cells. The internal battery problem could be dry outs, shorted cells or others.

  • An upward of more than five times the float charging current or more than 1A of float current should trigger immediate attention (Thermal runaway).
  • Extremely low or absence of float current can indicate an open-circuit between the charging system and the battery string (TPL-001-5).

Battery Age (life time)

Battery manufacturers tend to provide design-life value within a given set of parameters. This can be used to set a maximum replacement service age.

  • It is not uncommon to establish mandatory battery replacements based on the battery service life. Example: (X) years after the design life.

Battery warranty period

Using SoH% can identify initial cell failures and cells about to fail to benefit from warranty claims.

  • Consider replacing defective cell or the whole string within one year if more than 20% of cells have failed.

Conclusion

When managing batteries, everyone would like to move away from Excel.

Therefore, involving the right people and engaging in a respected preventive maintenance program is crucial. By using Atlas, Multitel’s data analytics software, users can have a global visibility, in a matter of seconds, of the critical network assets including the batteries’ SoH conditions. This information can then be used to better decide what corrective actions should be implemented to ensure the network reliability. 

Fundamentally, Atlas is beneficial not only for managing battery asset inventory, collecting battery data and generating data analytics, but also for budgeting replacement programs, monitoring available power capacity, managing long-term provisioning forecasts, identifying and locating failed cells that are still under warranty, and helping to extend the battery service life.

References:
Effects of AC Ripple Current on VRLA Battery Life  | Emerson Network Power 2009
188-2005 – IEEE Recommended Practice for Maintenance, Testing, and Replacement of Valve-Regulated Lead-Acid (VRLA) Batteries for Stationary Applications | IEEE Standard
The complete guide to battery monitoring | BTECH V2.0
What is the value of a number? | Battcon 2023 – George Pedersen,  Principle Associate and Carrann Associates, LLC Hagerstown, Maryland

Share this post