Battery condition is not described by one number alone. A battery can be evaluated through capacity, resistance or impedance, voltage behavior, temperature response, self-discharge, charge acceptance, and operating history. These measurements form a multidimensional diagnostic problem: patterns in several signals can reveal whether a battery is healthy, aging, or unsuitable for a particular mission.
Cloud-connected monitoring extends this analysis from one battery to an entire inventory. Measurements from battery-management systems, diagnostic equipment, chargers, and operating assets can be collected for comparison and trend analysis. This can support portable batteries as well as stationary energy-storage installations.
State of Health (SoH) is an estimate of present battery condition relative to a reference condition, usually when the battery was new. Capacity is a major SoH indicator, but it is not the only one. Capacity fade reduces the amount of charge a battery can deliver, while increasing resistance or abnormal temperature behavior can reduce usable power and reliability before a simple capacity test shows a major loss. Remaining Useful Life (RUL) is an estimate of how long a battery can continue to meet a defined requirement under expected conditions.
The source concept known as RUBY, or Remaining Useful Battery Yield, presents these ideas as a battery-yield and analytics framework. Its status and current product availability are not established by the supplied evidence, so it is best treated here as a historical or conceptual example rather than as a verified commercial system.
Energy Allocation: From Rated Capacity to Usable Battery Energy
A battery rating is a starting point, not a guarantee of mission energy. For example, under stated assumptions, a 10 Ah battery delivering a constant 10 A load could provide approximately one hour of discharge. In practice, that simple calculation assumes that the battery can deliver its rated capacity at that discharge rate and over the relevant voltage range. Actual runtime also depends on temperature, discharge rate, conversion efficiency, cutoff voltage, load variation, and the condition of the battery.
Ampere-hours describe charge delivered, while watt-hours are more directly related to energy because they include voltage. A battery’s voltage may decline during discharge, and the connected equipment may stop operating when a cutoff voltage is reached. Consequently, a 10 Ah rating does not mean that every application can extract the same amount of usable energy.
A useful distinction is the difference between nominal capacity and current available capacity. Nominal capacity is the rated value associated with a new battery and a defined test method. Current available capacity is what the battery can deliver now under specified test conditions. Aging, repeated cycling, storage, temperature exposure, and high-current operation can reduce that value. A battery may still appear fully charged while holding substantially less absolute energy than it held when new.
An application also may not use all of the battery’s current capacity. Some energy can be withheld as a reserve to accommodate unexpected load, delays, temperature changes, measurement error, or reliability requirements. This reserve is different from capacity lost through degradation:
- Degradation loss is capacity that the battery can no longer deliver because of aging or damage.
- Reserve capacity is energy intentionally not allocated to the normal mission.
- Usable mission energy is the portion accepted for the application after both effects and operating limits are considered.
End-of-life is therefore application-dependent. An 80% capacity threshold is often used as a reference in battery discussions, but it is not a universal rule. A low-power consumer device, a medical system, an aircraft subsystem, and a stationary storage installation may have different requirements for runtime, power, reliability, and safety. A battery can be unsuitable for one application while remaining serviceable in another less demanding role.
For illustration, suppose a new battery is rated at 100 units of nominal capacity. Aging may reduce its current capacity to 80 units. If the application also withholds a reserve from that remaining capacity, the energy allocated to the normal mission is lower still. A calculation that combines an end-of-life allowance with a reserve margin can produce an illustrative usable value near 60% of the original rating, but that figure is not a general battery rule. The exact result depends on whether the reserve is calculated from nominal capacity, current capacity, expected mission energy, or another operating policy.
Battery diagnostics can screen a battery against a target capacity or health requirement. Smart batteries may report values such as Full Charge Capacity, which represents the system’s estimate of how much charge the battery can currently store or deliver under its reporting definitions. The meaning and accuracy of such data depend on the battery-management system, its calibration, operating history, and communication specification. A reported value should therefore be interpreted together with test conditions and other indicators.
Less instrumented batteries require estimation from measurements. Coulomb counting integrates current over time to estimate charge entering or leaving the battery. It is useful for tracking charge, but errors can accumulate from sensor offset, uncertain initial state, changing efficiency, and incomplete charge or discharge cycles. Periodic reference points or capacity tests may be needed to correct drift.
Model-based methods use a mathematical battery model together with measured current and voltage. A Kalman-filter approach, including an Extended Kalman Filter for nonlinear models, can estimate internal states from noisy measurements. Such methods can be valuable, but they depend on model quality, parameter identification, sensor accuracy, and the chemistry and operating range being represented. No estimation method should be treated as universally accurate across all battery designs and conditions.
SoH is consequently a multidimensional estimate. Capacity may be combined with resistance or impedance, voltage response, temperature behavior, self-discharge, cycle history, and other evidence. A capacity threshold can help define whether a battery remains suitable, but it does not by itself establish the battery’s runtime, power capability, or safety condition.
Intelligent Target Selection for Application-Specific Battery Screening
An intelligent target selector is an application-specific rule for deciding whether a battery meets a required capacity, reserve, or health threshold. Instead of asking whether every battery has the same percentage of its original capacity, the selector asks whether a particular battery can satisfy a defined mission.
The rule is based on a what-if and risk-management calculation. Relevant inputs can include the expected load, duty cycle, traffic or activity level, temperature, operating duration, reliability requirement, and acceptable probability of failure. A battery may pass a low-demand task but fail a high-demand task even when its measured condition has not changed.
A basic screening system might provide a pass/fail result or a green/red status. Such an indicator is useful for sorting an inventory, but it is only as reliable as the measurements and assumptions behind it. Calibration, test repeatability, battery temperature, cutoff behavior, and uncertainty in the health model all affect the decision. A green indicator should not be interpreted as a guarantee of runtime or freedom from every failure mode.
Reserve energy can be included before the pass/fail decision is made. For example, a communications device may require enough energy for its planned operating period plus an additional reserve for retransmissions, higher traffic, cold conditions, or an unexpected delay. The battery passes only if its estimated usable energy exceeds both the expected demand and the required reserve.
Consider a portable system operating in an area where increased traffic causes its radio and processor to remain active for longer periods. The average load rises, and a battery that previously met the mission threshold may become marginal. The target selector can raise the required capacity or reserve level, causing more aging batteries to fail screening. This does not necessarily mean that the batteries have suddenly degraded; the mission requirement has become more demanding.
The inverse case is also important. If demand is lower, the same battery may remain suitable for a shorter or less intensive assignment. Continued use is acceptable only when the revised requirement remains consistent with safety, reliability, and operational policy. Adaptive screening should not be used to conceal a known defect or bypass a required inspection.
A capacity threshold alone cannot guarantee runtime. Runtime depends on the load profile, current peaks, temperature, voltage limits, conversion losses, battery chemistry, and the equipment’s cutoff behavior. Two devices drawing the same average current can have different runtimes if one has frequent high-current pulses or a more restrictive voltage cutoff.
The RUBY or diagnostic-charger example can be understood as a conceptual implementation of this type of adaptive screening. The supplied evidence does not verify that a current commercial product implements all of these functions, nor does it establish the accuracy of any particular green/red classification. In an engineering deployment, the threshold would need to be validated against actual mission data and periodically reviewed as batteries, loads, and environmental conditions change.
Fleet Supervision for Battery Capacity, Reserve, and Remaining Useful Life
Managing one battery is different from managing hundreds or thousands. A fleet dashboard can compare each battery’s estimated usable capacity, health indicators, reserve category, predicted life, test history, and service status. This allows an organization to distinguish a battery that is ready for a demanding assignment from one that should be tested, restricted, repaired, replaced, or quarantined.
Reserve should be displayed separately from immediately usable energy. A battery can have adequate stored charge for a short task but insufficient reserve for the required reliability policy. Separating these categories prevents a large nominal capacity from hiding a small operational margin.
A practical fleet view might include:
- current or estimated usable capacity;
- SoH indicators and recent trend;
- required mission threshold;
- reserve-energy status;
- charge, discharge, and temperature history;
- last diagnostic test and measurement confidence;
- service, replacement, or quarantine status; and
- an RUL estimate with an uncertainty range where possible.
Batteries that fail the selected application requirement can be marked for action. The next step may be a confirmatory capacity test, a resistance measurement, inspection for physical damage, reassignment to a lower-demand application, controlled recycling, or replacement. Ranking the inventory by risk and mission importance helps technicians focus on assets that could affect operations most seriously.
RUL can be expressed in time units only when the system has sufficient historical and operating data. The estimate may need information about load profile, depth of discharge, temperature, storage conditions, charging practices, maintenance, and degradation history. If future use is unknown or the data are sparse, reporting a precise number of months or years can create false confidence. A range, confidence level, or condition category may be more appropriate.
Cloud analytics can support reliability-centered maintenance by connecting measured condition with operational consequence. Reliability-centered maintenance generally asks what functions an asset must perform, how it can fail, what causes those failures, and what inspection or maintenance task reduces the associated risk. For battery fleets, this means that a replacement decision should consider both battery condition and the consequence of failure in its assigned role.
This approach is relevant wherever many batteries support critical or distributed equipment, including aviation and military systems, transportation, portable field equipment, communications assets, and stationary energy storage. The specific maintenance practices and acceptance limits vary by sector; the common requirement is to relate battery condition to the function the asset must perform.
Cloud monitoring is not a substitute for safe handling, electrical protection, inspection, or manufacturer limits. It also cannot recover missing information. Poor sensors, inconsistent test procedures, incorrect battery identification, and incomplete usage records can undermine the resulting analysis. Data governance, calibration, model validation, and clear escalation procedures are as important as the dashboard itself.
Organizations operating battery-intensive fleets may therefore need specialized diagnostics, structured data systems, and lifecycle expertise. When capacity, reserve, degradation trend, and mission risk are evaluated together, fleet analytics can support better testing priorities, replacement planning, inventory control, and total-cost decisions without reducing battery health to a single percentage.
References
- BU-907c: Cloud Analytics in Batteries. (n.d.). http://www.batteryuniversity.com/article/bu-907c-cloud-analytics-in-batteries
- How to Measure the Remaining Useful Life of a Battery. (n.d.). https://www.large-battery.com/blog/how-to-measure-the-remaining-useful-life-of-a-battery-guide
- A Closer Look at State of Charge (SOC) and .... (n.d.). https://www.analog.com/en/resources/technical-articles/a-closer-look-at-state-of-charge-and-state-health-estimation-tech.html
- What Are SOH and SOC in a Battery? Key Differences and How to Measure Them.. (n.d.). https://www.maxwellenergy.co/blog/soh-and-soc-in-battery
- How Battery Management Systems Estimate State-of-Charge and State-of-Health in Lithium-Ion Batteries. (n.d.). https://www.wonderfulpcb.com/blog/battery-management-system-soc-soh-estimation-lithium-ion
- State of Charge Estimation Using Extended Kalman Filters .... (n.d.). https://pangea.stanford.edu/ERE/pdf/OnoriPDF/Conferences/47.pdf
- Battery State of Charge vs State of Health: Complete Guide. (n.d.). https://www.evlithium.com/Blog/battery-soc-vs-soh-guide.html
- Lead Acid、Lithium & LiFePO4 Battery Run Time Calculator. (n.d.). https://manlybattery.com/battery-tools-and-resources/lead-acid-lithium-lifepo4-battery-run-time-calculator
- Advanced SOH Estimation BESS: Kalman & ML Methods. (n.d.). https://sunlithenergy.com/advanced-soh-estimation-bess
- How to Calculate Battery Run Time for Different Battery Types. (n.d.). https://www.large-battery.com/blog/calculate-battery-run-time-different-types