
Many BESS DC systems are designed to operate without either polarity intentionally bonded to earth. That arrangement can allow continued operation after an initial insulation fault, but it also makes early detection, fault location and controlled response essential. A buyer who asks only for “ground-fault protection” may receive very different functions from different suppliers.
A BESS insulation-monitoring requirement should define the monitored electrical boundary, compatible measuring method, alarm thresholds, response time, operating states, location process, data retention and safe return to service. It must also distinguish insulation resistance monitoring from overcurrent protection, residual-current detection and protective bonding. These functions can interact, but they do not prove the same condition.
The IEC description for IEC 61557-8 covers insulation monitoring devices for unearthed AC and DC IT systems within its stated voltage scope. The project should confirm the applicable standards, local rules and equipment ratings with its qualified designer; citing an IMD standard does not establish the complete BESS protection philosophy.
Draw the Monitored Boundary for Every Switching State
Start with the DC single-line diagram and mark battery strings, rack contactors, combiner circuits, pre-charge paths, DC bus, PCS input, service disconnects and any auxiliary DC circuits. Then show the portion visible to each monitoring function when contactors are open, closed or in transition. A device cannot detect a fault in a section electrically disconnected from its measuring path.
Record whether the PCS contains its own insulation measurement and whether multiple battery blocks or PCS units can become galvanically connected. Two active measuring devices on one connected system may interfere unless their compatibility is established. Conversely, opening a contactor can leave a section with no active monitor. The sequence and equipment design determine the answer.
Include maintenance and emergency states. A cabinet may be isolated from the PCS while internal battery voltage remains present. A technician needs to know which monitor remains active, what local indication is available and how an insulation condition is verified before access. Connect this logic to the isolation principles in the battery DC isolation guide.
| Operating state | Monitoring question | Required evidence |
|---|---|---|
| All main contactors open | Which battery and bus sections remain monitored? | State diagram and injected-fault test points |
| Pre-charge | Does the temporary circuit change the measured boundary? | Sequence trace and monitor behavior |
| Normal charge/discharge | Do switching noise and PCS filters affect stability? | Trend across representative power states |
| Multiple blocks connected | Can monitors interfere or mask location? | Compatibility declaration and system test |
| PCS isolated for service | Is the battery-side circuit still supervised? | Local indication and maintenance procedure |
| Auxiliary power lost | Is monitoring retained and is loss visible? | Power-loss and recovery test |
| Insulation alarm active | Which automatic limits or trips apply? | Approved cause-and-effect record |
Define What the Measurement Means
Specify the displayed quantity, range, resolution, update behavior and diagnostic state. Insulation resistance to earth is not the same as fault current, and a site should not convert one into the other without an approved model. Ask whether the device reports a combined system value, estimates polarity or location, and distinguishes an internal diagnostic failure from a low-insulation condition.
Thresholds should follow the selected architecture, applicable requirements and supplier safety analysis. Avoid copying one resistance value from another product. Document warning and trip or inhibit levels, delays, hysteresis, latching, reset and behavior during startup. If the measured value varies normally with contactors or PCS state, define expected transitions so operators do not normalize a real defect as nuisance behavior.
The 2026 IEC TR 63436 describes IMD setting parameters and interpretation examples for its application scope. Buyers can use that as current technical context while requiring the BESS supplier to provide configuration-specific settings, expected curves and diagnostic limitations.
Separate a First Fault from a Safe-to-Continue Decision
An unearthed system may remain energized after an initial insulation fault, but continued operation is an engineering and operating decision, not an automatic entitlement. A second fault on the opposite polarity can create a damaging path. The cause-and-effect should consider measured severity, rate of change, affected boundary, redundancy, environment, service duty and the owner’s ability to investigate promptly.
Define which conditions produce an advisory, maintenance alarm, power restriction, blocked start or immediate shutdown. State who can authorize temporary operation and what compensating controls apply. A remote reset should not clear the obligation to locate and correct the cause. Preserve the initial value, minimum value, transitions and related events.
The BMS, PCS and plant controller may each participate in the response. Use the BMS–PCS–EMS control-loop model to assign detection, enforcement and supervisory roles. If the IMD signal reaches the plant only through a gateway, specify local behavior during communications loss.

Plan Fault Location before the First Alarm
An alarm that cannot be localized can keep a large plant unavailable while technicians divide circuits manually. Ask how the design narrows the affected battery block, rack, branch, PCS or cable. Options may include selective monitoring, switching sequences, portable test equipment and supplier diagnostic data. The method must respect stored energy and should not require unsafe trial-and-error disconnection.
Create a fault-location flow with authorized isolation steps, required waiting or discharge conditions, test points, instrument ratings and stop rules. Identify which operations can be completed remotely and which require qualified personnel onsite. Preserve the original system condition before opening contactors that may remove the fault from the monitored boundary.
Map likely causes without assuming one. Cable damage, moisture, condensation, contaminated connectors, coolant, pinched harnesses, failed surge protection, heater circuits, filters or electronic components can lower insulation. The BESS condensation-control guide is relevant where humidity and cold surfaces create intermittent paths that disappear after the enclosure is opened.
Stock the instruments, adapters, drawings and competent support needed for the location method. If only the OEM can access detailed measurements, define response, remote-access approval, onsite escalation and data ownership in the service contract.
Verify Compatibility with PCS and Other Monitoring Functions
Power converters can include EMC filters, surge devices and measurement functions that influence the apparent path to earth. Ask the PCS and battery suppliers to confirm compatibility for the final connected configuration, not their packages in isolation. Record allowed numbers of parallel units, supported operating modes and any setting changes required when blocks are added or isolated.
Check behavior during grid-connected, islanded, standby, black-start and maintenance states used by the project. Auxiliary transformers, chargers or temporary test equipment can introduce a bond or measurement path. Commissioning procedures should control what is connected when baseline insulation values are recorded.
The DOE’s 2026 Electrical Safety Handbook notes hazards associated with accidental grounding of a battery polarity and recommends ground detection in its federal-work context. The project must translate such safety principles into its current codes, architecture and qualified work practices.
Design FAT and SAT around Known Test Points
Define safe fault-injection points and approved resistance or simulator methods during design. Do not improvise connections on energized battery circuits during witness testing. The procedure should state isolation, instrument ratings, connection sequence, expected values, removal and restoration. The supplier should identify conditions that cannot be simulated on the complete system.
At FAT, verify tags, range, thresholds, delays, controller actions, local and remote alarms, communications loss, diagnostic failure, event retention and reset. Exercise contactor states to show which sections remain visible. Use the BESS FAT protocol to bind results to firmware and settings.
At SAT, first establish a baseline with the final cables, PCS, surge devices and auxiliaries connected. Repeat approved injections at accessible boundaries, verify SCADA mapping and demonstrate the fault-location process. Record ambient condition, plant state and configuration because intermittent moisture or switching effects may not appear in a single static test.
| Acceptance test | Required record | Decision |
|---|---|---|
| Measurement baseline | Configuration, contactors, auxiliaries, PCS state and value | Normal reference is reproducible |
| Warning threshold | Injected condition, delay, local/remote alarm and trend | Early response matches approved logic |
| Trip or inhibit threshold | Automatic actions, interlocks, event order and reset | Unsafe operation is prevented as specified |
| Disconnected section | Visibility with main paths open | No unrecognized monitoring gap |
| Device diagnostic failure | Alarm, fallback and maintenance instruction | Monitor failure is not confused with healthy insulation |
| Communications loss | Local protection, retained data and restored indication | Safe behavior does not depend on SCADA |
| Fault-location drill | Steps, time, test equipment and evidence | Team can identify the affected boundary safely |
| Return to service | Cause, repair, tests, approval and monitoring period | Restart is evidence-based |
Preserve Data before Reset and Repair
Build a Baseline across Environmental and Operating States
One insulation reading at dry, de-energized commissioning does not describe the plant. Establish reference values with the approved configuration during standby, pre-charge, charge, discharge and relevant contactor states. Record temperature, humidity, recent rain or washing, HVAC state and which auxiliaries are energized. The objective is not to define a universal normal value but to understand repeatable behavior for this installation.
Review trends after the first seasonal or wet-period operation where the climate makes that relevant. A value that falls only during a specific transition can point toward condensation, a heater, pump, fan, filter or PCS state. The analysis should preserve the time relationship and avoid averaging away a short but significant deterioration.
Set operating review bands separately from protection settings. A maintenance trend can trigger inspection well before the safety function reaches its approved warning threshold. Define who reviews the trend, what comparison is used and how a slowly deteriorating condition becomes a work order. This turns the monitor into preventive evidence without weakening the protective response.
Compare similar cabinets carefully. A fleet outlier can be useful, but only if equipment configuration, connected boundary and conditions are comparable. Preserve accepted design differences and service history. Do not reset an alert threshold merely because several units show the same decline; repeated behavior may indicate a shared environmental or production issue.
The DOE Energy Storage Safety Strategic Plan places validation and safety decisions across the full project lifecycle. Use the commissioning baseline as the start of an operating record, then update it after approved configuration changes and confirmed repairs.
Assign Supplier, EPC and Owner Duties for the Alarm
The equipment supplier should define the measuring function, compatible architecture, settings basis, diagnostic states, known limitations and approved repair checks. The EPC should install and verify the connected boundary, earthing, cables, communications and alarm routing. The owner should control response, access, evidence retention, maintenance and changes after handover.
Name who interprets an ambiguous event and how support is obtained outside normal hours. Remote support should receive the minimum necessary data through an approved access path; it should not bypass the owner’s switching or safe-work authority. Record recommendations, configuration changes and the person who approves return to service.
Define warranty notification and investigation duties before the first alarm. The owner should know which logs and measurements must be preserved, whether opening an enclosure requires supplier authorization and which emergency actions take precedence. A warranty clause must not delay work required to make an unsafe condition secure.
Include the monitor and its test method in training. Operators need to distinguish warning, protection action, device failure and loss of communications. Electrical maintenance staff need the monitored-boundary diagram, rated instruments, fault-location procedure and authority limits. A generic SCADA course is not adequate for this function.
Trend insulation value, device status, battery voltage, contactor states, PCS mode, humidity, temperatures and related alarms with synchronized timestamps. Capture a window before and after the event. The sequence may reveal whether the condition followed rain, cooling startup, maintenance, a switching transition or a specific operating block.
Use the evidence-preservation process in the BESS alarm-triage article. A reset may change contactors or clear diagnostic context. Record the value and plant state before action, then document every isolation and measurement used to locate the fault.
After repair, compare the restored value with the baseline under a comparable configuration. Check adjacent or similar equipment if the cause could be systemic. A replaced component closes the work order only after the system passes approved electrical checks and the alarm does not recur during the defined observation period.

Control Changes That Can Alter the Insulation Baseline
Added battery blocks, replacement PCS units, surge devices, filters, heaters, coolant equipment, cabling and firmware can change the measured system. Require an insulation-monitoring impact screen for relevant modifications. Update the single line, monitored-boundary diagram, settings and baseline after accepted changes.
Temporary commissioning and maintenance connections require the same review. Test supplies, portable instruments, laptop interfaces, grounding leads and auxiliary generators can introduce a path that changes the reading or defeats the intended unearthed arrangement. The work procedure should identify approved connections, expected indication and the restoration check after temporary equipment is removed.
When replacing the IMD or a controller that contains the function, retain the old model, firmware, settings, calibration or verification history and reason for replacement. Confirm compatibility with the connected capacitance and system behavior, then repeat baseline, alarm, communication-loss and fault-location checks. Similar terminal labels do not prove equivalent performance.
Sandia’s energy storage codes and standards inventory shows the breadth of safety and system standards that can apply. Buyers should maintain a project-specific requirements register rather than treating one monitor certificate as proof of overall electrical safety.
Put the Insulation-Monitoring Package into the RFQ
| RFQ item | Supplier response | Acceptance gate |
|---|---|---|
| System grounding basis | Architecture, intentional bonds and applicable requirements | Approved before detailed design |
| Monitored-boundary diagrams | Coverage for every contactor and service state | No critical blind section |
| Device and method | Ratings, compatibility, diagnostics and limitations | Technical submittal accepted |
| Threshold philosophy | Levels, delays, hysteresis, actions and authority | Cause-and-effect approved |
| Fault-location method | Steps, instruments, access, skills and support | Tabletop and site drill completed |
| FAT/SAT procedure | Safe injections, states, records and pass criteria | Witness tests passed |
| Data package | Signals, resolution, time, retention and export | Owner can reconstruct an event |
| Change control | Impact triggers, revised settings and retest | Active throughout service |
The Acceptance Decision
Accept BESS insulation monitoring when the owner knows what is monitored in every relevant state, understands the measurement and limitations, can observe the approved automatic response, can locate a fault safely and can verify repair before return to service.
The useful deliverable is not an alarm named “insulation low.” It is a controlled electrical-safety function with a defined boundary, traceable configuration and an operating procedure that turns early warning into timely corrective action.