Doc #73 — Solar Panel and Inverter Maintenance

---

RECOMMENDED ACTIONS (BY URGENCY)

First months (Phase 1)

1. Include solar installations in the national asset census (Doc #8): Establish the number, capacity, type (string inverter, microinverter, hybrid/battery), location, and condition of all NZ solar systems. Lines companies have data on grid-connected systems; off-grid systems are undocumented.
2. Inventory inverter spare parts nationally: Electronic component distributors, solar installation companies, and electricians hold inverter spares — particularly capacitors, fuses, fans, and replacement boards. These are strategic materials.
3. Classify experienced solar installers and power electronics technicians as critical-skills personnel (Doc #1). NZ has an estimated 2,000–5,000 workers in the solar industry, with a smaller subset competent in inverter-level diagnostics.⁴
4. Issue guidance to solar system owners: Keep systems operational. Do not disconnect. Clear shading obstructions. Check for obvious physical damage.

First year (Phase 1–2)

5. Establish regional inverter repair workshops in areas with high solar penetration (Auckland, Canterbury, Bay of Plenty, Waikato). Equip with oscilloscopes, multimeters, soldering stations, and component stocks salvaged from electronic waste.
6. Begin structured knowledge capture from inverter design engineers and power electronics specialists — document repair procedures, common failure modes, and component substitution tables.
7. Develop direct-DC utilisation guidelines: Identify applications where panels can be used without inverters — battery charging (Doc #35), resistive water heating, DC lighting, DC motor drives.
8. Inventory electrolytic capacitors nationally: Capacitors are the most common inverter failure point. All stocks at electronic component suppliers (RS Components NZ, Element14, Jaycar, independent distributors) should be requisitioned and catalogued by type, voltage, and capacitance rating.
9. Assess all battery-equipped solar systems for off-grid conversion capability — these systems already have charge controllers and battery inverters that may outlast grid-tied inverters.

Years 1–3 (Phase 2)

10. Scale inverter repair to handle fleet-wide demand — target repairing 2,000–5,000 inverters per year as the first wave of failures arrives.
11. Begin converting high-priority installations to direct-DC applications where inverters have failed and cannot be repaired.
12. Develop training programme for inverter repair technicians — draw from electricians, electronics hobbyists, and engineering graduates (Doc #157).
13. Establish capacitor reconditioning and testing capability — salvaging electrolytic capacitors from non-essential electronic equipment and testing for remaining useful life.

Years 3–7 (Phase 3)

14. Begin development of simplified square-wave or modified-sine-wave inverters for resistive loads (heating, lighting) using discrete components.
15. Develop low-voltage DC micro-grids in communities where solar panels remain functional but inverters have failed — aggregating DC output for local battery charging and direct-DC applications.
16. Establish systematic panel cleaning and inspection programme as panels age and potential-induced degradation (PID) becomes more prevalent.

Years 7–15 (Phase 4)

17. If discrete semiconductor production becomes available (Doc #115), develop locally manufactured inverter components.
18. Assess whether simplified pure-sine-wave inverter production is feasible using imported or domestically produced power transistors.
19. Transition surviving solar capacity into integrated local energy systems alongside micro-hydro (Doc #72) and the national grid (Doc #67).

---

ECONOMIC JUSTIFICATION

The asset at stake

NZ’s installed solar fleet represents an estimated 400–600 MW of generation capacity that was installed at a pre-event cost of approximately NZ$1–2 billion.⁵ Replacing this capacity with other generation under isolation conditions would require hundreds of micro-hydro installations (Doc #72) or significant expansion of wind capacity — each requiring materials, labour, and years of construction.

Even degraded, the solar fleet is significant. Panels retain roughly 85–95% of original output after 20 years of degradation (at 0.3–0.7% per year, per section 2.2). With inverter losses factored in — assuming a repair programme keeps 50–70% of inverters operational — effective fleet output would be 40–65% of original nameplate capacity, or roughly 160–390 MW. This represents approximately 3–7% of NZ’s total generation capacity.⁶ This is modest in the context of the national grid, but the value is disproportionate to the nameplate capacity because solar generation is distributed — it produces power at the point of consumption, reducing demand on the transmission and distribution network (Doc #67). Every kilowatt-hour generated by a rooftop solar panel is a kilowatt-hour that does not need to be transmitted through aging transformers (Doc #69) and distribution lines.

Investment required

Inverter repair programme (Phase 2 onward):

• Facilities: Utilise existing electrical workshops, electronics repair shops, and solar company premises.
• Equipment: Oscilloscopes, soldering stations, component testers, multimeters — most already exist in NZ. Estimated 20–50 equipped repair stations nationally.
• Training: 50–200 inverter repair technicians, drawn from electricians and electronics technicians. Training period: 3–6 months under experienced supervision.
• Materials: Capacitors, MOSFETs, IGBTs, control boards — from existing stocks, salvage, and eventually trade.
• Estimated labour: 10–30 person-years per year of ongoing repair work at scale.

Direct-DC conversion programme (Phase 2 onward):

• Materials: DC-rated switches, fuses, wiring, charge controllers — from existing stocks and salvage from non-repairable inverter systems.
• Labour: Qualified electricians for each conversion — estimated 2–8 person-hours per installation depending on complexity.
• Estimated labour: 5–15 person-years per year if converting 5,000–10,000 installations.

Comparison with alternatives

The alternative to maintaining solar capacity is accepting its progressive loss and compensating with other generation and efficiency measures. Given that the panels themselves last decades, the repair programme’s primary cost is the electronic components and technician labour for inverter maintenance — a modest investment relative to the value of the electricity produced. Even a programme that keeps only 30–50% of the installed fleet operational represents a meaningful contribution to NZ’s energy system for decades.

---

1. NZ’S SOLAR FLEET

1.1 Scale and distribution

NZ’s solar PV installation base has grown rapidly from a negligible base in 2010 to an estimated 50,000–70,000 grid-connected systems by 2025.⁷ The majority are residential rooftop systems in the 3–10 kW range, with a growing number of commercial systems (30–300 kW) and a small number of utility-scale installations. Total installed capacity is estimated at 400–600 MW, though exact figures require verification through the national asset census (Doc #8).

Geographic distribution is uneven. Solar installation rates are highest in the upper North Island (Auckland, Waikato, Bay of Plenty) where solar irradiance is strongest and population is concentrated. The South Island has fewer installations but longer summer daylight hours partially compensate for lower solar intensity. Canterbury, particularly Christchurch and surrounding areas, has a meaningful concentration.

NZ solar irradiance context: NZ receives approximately 1,300–1,700 kWh/m²/year of global horizontal irradiance depending on location, with the upper North Island at the higher end and the deep south at the lower end.⁸ This is lower than Australia (~1,800–2,200 kWh/m²/year) but higher than the UK (~900–1,100 kWh/m²/year) and comparable to Germany (~1,000–1,200 kWh/m²/year), a country that has installed enormous solar capacity. Under nuclear winter conditions with 5–8°C cooling and potential stratospheric aerosol loading, solar irradiance reaching the surface could decrease by 10–30% for 3–7 years, depending on aerosol density and geographic latitude.⁹ This would proportionally reduce solar panel output, though panels continue to function — they produce less power, not zero.

1.2 System types

Grid-tied without battery (majority of NZ systems):

A string inverter converts DC from a series-connected array of panels to 230V 50Hz AC, which is fed into the household distribution board. Surplus is exported to the grid. When the grid goes down, the inverter must shut off — this is a safety requirement called anti-islanding. These systems produce nothing during grid outages unless modified.

Grid-tied with battery (growing minority):

A hybrid inverter manages power flow between the solar array, a battery bank (typically lithium-ion), the household loads, and the grid. These systems can island — they can disconnect from the grid and continue powering the household from solar and battery. This capability is disproportionately valuable under recovery conditions because it provides resilience against local grid failures.

Off-grid (small number):

Stand-alone systems with charge controllers, battery banks, and battery-based inverters. Not grid-connected. Found in remote locations — bach communities, off-grid farms, remote telecommunications sites. These systems are already self-sufficient and represent the most resilient solar installations in NZ.

Microinverter systems:

Each panel has its own small inverter (Enphase is the dominant brand in NZ). Advantages: panel-level monitoring, no single point of failure (one microinverter failing loses only one panel’s production). Disadvantages: microinverters are on the roof, harder to access for repair, and generally not designed for field-level component repair — they are sealed units intended for replacement.

1.3 What is in the fleet

The majority of NZ’s installed panels are crystalline silicon (monocrystalline or polycrystalline) — as of 2025, this technology dominates the global market at over 95%.¹⁰ NZ has a very small number of thin-film installations (CdTe or CIGS), primarily on commercial buildings. For practical purposes, this document addresses crystalline silicon panels, which is what NZ has.

Inverter brands in the NZ fleet include Fronius, SMA, Enphase, Huawei, Sungrow, ABB/FIMER, and various Chinese brands. The diversity of brands and models is both a problem (many different component types needed for repair) and an advantage (failure of one model does not take down the entire fleet).

---

2. PANEL DURABILITY AND DEGRADATION

2.1 Why panels last so long

A crystalline silicon solar panel has no moving parts, no fluids, no chemical reactions during operation (the photovoltaic effect is a solid-state quantum process), and minimal thermal stress (operating temperatures rarely exceed 70°C). The active material — silicon wafer cells — is one of the most chemically stable materials known. The panel is essentially a glass-and-aluminium sandwich protecting fragile silicon cells from the environment.

Construction layers (front to back):

• Tempered glass front sheet (~3.2 mm, designed to withstand hail impact)
• Ethylene-vinyl acetate (EVA) encapsulant — transparent polymer that bonds the cells to the glass
• Silicon cells, series-connected with soldered ribbon conductors
• EVA encapsulant (rear)
• Backsheet — polymer film (typically PVF or PET-based) protecting the rear
• Aluminium frame
• Junction box with bypass diodes and cable connections

Under normal conditions, the glass, silicon, and aluminium frame are essentially permanent. The vulnerable elements are the EVA encapsulant (which can yellow and delaminate over decades, reducing light transmission), the backsheet (which can crack, allowing moisture ingress), the solder joints connecting cells, and the junction box seals.

2.2 Degradation mechanisms

Light-induced degradation (LID): A 1–3% drop in output in the first year of operation, caused by boron-oxygen defects in the silicon. This has already occurred in NZ’s existing fleet and is not an ongoing concern.¹¹

Annual degradation: Crystalline silicon panels degrade at approximately 0.3–0.7% per year under normal conditions, primarily from microcracking of solder joints due to thermal cycling, gradual EVA yellowing, and slow increase in cell series resistance.¹² At the 0.3–0.7% range, a panel retains roughly 75–90% of its original output after 30 years and 65–85% after 50 years. This is slow enough that age alone does not render panels useless for decades.

Potential-induced degradation (PID): A voltage-driven degradation mechanism in which leakage current from the cell to the grounded frame causes ion migration in the glass and encapsulant, degrading cell performance. PID can cause 10–30% output loss in affected panels, particularly those at the negative voltage end of a string. It is partially reversible if the voltage stress is removed or reversed.¹³ Under recovery conditions, where inverters may be operating outside their design parameters or panels are reconfigured, PID risk should be monitored.

Hot spots and bypass diode failure: If a cell is shaded or cracked, it becomes a current bottleneck and heats up. Bypass diodes in the junction box route current around the affected cell group. If a bypass diode fails (short or open circuit), the panel either loses a third of its output or develops a persistent hot spot that can melt solder and damage the backsheet. Junction box inspection and bypass diode testing is a low-effort, high-value maintenance activity.

Physical damage: Broken glass from hail, wind-borne debris, or falling trees. A crack in the front glass allows moisture ingress into the EVA and cells, causing corrosion and ground faults. Panels with broken glass should be removed from service or, if the damage is localised, sealed with silicone or resin as a temporary measure — this is not a permanent fix but can extend useful life by years.

Moisture ingress: Through cracked backsheets, damaged junction boxes, or delaminated edges. Causes cell corrosion, increased series resistance, and ground faults. The most common long-term failure mode for panels that are not physically broken.

2.3 Nuclear winter effects on panels

Reduced irradiance: Stratospheric aerosol loading from nuclear detonations reduces sunlight reaching the surface. Estimates vary widely — 10–30% reduction for the Southern Hemisphere over a period of 3–7 years, less severe than the Northern Hemisphere due to distance from likely detonation sites.¹⁴ Panel output decreases roughly proportionally — a 10–30% irradiance reduction translates to approximately 10–30% less electricity production. At the midpoint of this range, this is significant but not disabling.

Increased UV: Ozone depletion from nuclear firestorms increases UV-B radiation at the surface for several years (Doc #41). UV-B accelerates polymer degradation — specifically EVA yellowing and backsheet embrittlement. This could modestly accelerate panel degradation rates, though the magnitude is uncertain. NZ already has high UV levels due to its proximity to the Antarctic ozone hole, so NZ-installed panels are exposed to higher baseline UV than panels in most markets.

Temperature: Colder ambient temperatures actually improve solar panel efficiency — crystalline silicon output increases approximately 0.3–0.5% per degree Celsius below 25°C.¹⁵ Under nuclear winter conditions with 5–8°C cooling, panel efficiency would increase by roughly 1.5–4%, partially offsetting the irradiance reduction. This is a small effect but worth noting.

Snow and ice: NZ’s solar panels are predominantly in locations where snow is uncommon. Under nuclear winter conditions, snow cover could become more frequent, particularly in the South Island and higher elevations of the North Island. Snow-covered panels produce nothing until cleared. Panels on tilted rooftops shed snow naturally to some degree; ground-mounted arrays and flat-roof installations may need manual clearing.

---

3. INVERTERS: THE WEAK LINK

3.1 What an inverter does

A grid-tied solar inverter performs several functions simultaneously:

DC to AC conversion: The core function. Power transistors (IGBTs or MOSFETs) switch at high frequency (typically 10–50 kHz) to chop the DC input into a pulse-width-modulated (PWM) approximation of a sine wave. Output filters (inductors and capacitors) smooth this into clean 230V 50Hz AC that meets grid power quality standards.¹⁶

Maximum power point tracking (MPPT): The inverter continuously adjusts the operating voltage of the solar array to extract maximum power under changing irradiance and temperature conditions. This is a software-controlled feedback loop that runs continuously. Without MPPT, a directly connected load operates the panel at whatever voltage the load demands, which may be far from optimal — potentially extracting only 50–70% of available power.¹⁷

Anti-islanding: For safety, grid-tied inverters must detect when the grid has failed and shut down within milliseconds. This prevents the inverter from energising a grid circuit that linesmen assume is de-energised. The anti-islanding requirement means that standard grid-tied inverters produce nothing during grid outages — a significant limitation under recovery conditions.

Grid synchronisation: The inverter matches its output frequency and phase to the grid, and manages reactive power and power factor as required by the grid code. This requires precise voltage and frequency measurement and real-time control.

Monitoring and communication: Most modern inverters include WiFi or cellular communication for monitoring, firmware updates, and remote diagnostics. This functionality becomes irrelevant under isolation but the underlying microprocessor hardware may still be needed for the control algorithms.

3.2 Why inverters fail

Electrolytic capacitors (most common failure): Electrolytic capacitors in the DC bus and output filter dry out over time. The electrolyte — a liquid chemical solution — gradually evaporates through the seal, reducing capacitance and increasing equivalent series resistance (ESR). High operating temperatures accelerate this process dramatically — capacitor life roughly halves for every 10°C above rated temperature.¹⁸ A capacitor rated for 10,000 hours at 105°C may last 20,000 hours at 95°C or 40,000 hours at 85°C. Inverters mounted in hot, poorly ventilated locations fail faster than those in shaded, well-ventilated positions.

Capacitor failure manifests as increased ripple on the DC bus (causing control instability), reduced output power, nuisance tripping, or complete shutdown. It is the most common repairable inverter failure and the highest-priority component to stockpile.

Power semiconductor failure (IGBTs/MOSFETs): Power transistors degrade through thermal cycling — repeated heating and cooling causes solder joint fatigue between the semiconductor die and its substrate. Sudden failure (short circuit or open circuit) destroys the output stage. In lower-quality inverters, undersized transistors running near their thermal limits fail faster. IGBT and MOSFET replacement requires soldering skills and correctly rated replacement components — the wrong component rating causes immediate failure or unreliable operation.¹⁹

Cooling fan failure: Inverters above approximately 3–5 kW typically include cooling fans. Fans have bearings that wear out, typically after 5–10 years. A failed fan causes overheating, which accelerates capacitor and semiconductor degradation. Fan replacement is a straightforward repair — typically 30 minutes — if a compatible fan is available.

Microprocessor and firmware issues: The control microprocessor and its firmware manage all inverter functions. Hardware failures (dead processor, corrupted flash memory) are not field-repairable — the entire control board must be replaced. Firmware bugs can sometimes be addressed by resetting to factory defaults, but this requires access to the manufacturer’s service procedures.

Surge damage: Lightning strikes, switching surges from the grid, or voltage spikes from load switching can damage input and output protection circuits (varistors, TVS diodes), or — if protection fails — can destroy power semiconductors and control electronics. NZ has moderate lightning activity, less than tropical regions but sufficient to cause regular damage.²⁰

Moisture and corrosion: Inverters are rated IP65 or similar for outdoor installation, but seals degrade over time. Moisture ingress causes corrosion on circuit boards, particularly in coastal NZ environments where salt spray is present. Corrosion of solder joints and connector pins causes intermittent faults.

3.3 Inverter lifespan under NZ conditions

Manufacturer warranties for string inverters are typically 5–10 years, with extended warranties available to 15–20 years.²¹ Actual field life depends on installation quality, environmental conditions, and component sourcing. A reasonable estimate for NZ conditions:

• Well-installed, well-ventilated string inverters from reputable manufacturers: 12–20 years before major failure.
• Poorly ventilated or budget-quality inverters: 7–12 years.
• Microinverters: 15–25 years (simpler design, lower power per unit, often potted/sealed electronics — but inaccessible for repair).
• Battery/hybrid inverters: 10–15 years (more complex, higher component count).

Fleet-wide implication: Without repair capability, NZ can expect to lose approximately 30–50% of its inverter fleet within the first 10 years of isolation, and 60–80% within 15 years. With an active repair programme focused on capacitor and fan replacement, many of these units can be returned to service, potentially extending the average fleet life to 15–25 years.

3.4 Repair vs. replacement

Under pre-event conditions, a failed inverter is replaced — the economics of labour-intensive board-level repair do not compete with a new imported unit. Under isolation, the economics reverse completely. A dead inverter renders a functional solar array — worth thousands of dollars of installation labour and decades of remaining panel life — useless. Spending 4–20 hours of technician time to diagnose and repair a failed inverter is overwhelmingly justified.

Component-level repair hierarchy (from most to least feasible):

1. Cooling fan replacement: 30 minutes. Any compatible 12V DC fan. Highest success rate.
2. Electrolytic capacitor replacement: 1–4 hours. Requires correctly rated capacitors (voltage, capacitance, temperature rating, physical size), soldering equipment, and the ability to identify failed capacitors (visual inspection for bulging/leaking, ESR testing with a capacitance meter). Success rate: 60–80% if the correct capacitor is available.²²
3. Fuse and surge protection replacement: 30 minutes–2 hours. Varistors, TVS diodes, fuses. Requires identification of correct ratings.
4. Power semiconductor replacement: 2–8 hours. Requires correctly rated IGBT or MOSFET modules, thermal paste, and careful soldering. Higher risk — incorrect installation causes immediate failure. Success rate: 40–60% for experienced technicians.²³
5. Control board replacement: If the same model board is available from a donor unit, 1–4 hours. If not available, the inverter is unrepairable.

---

4. DIRECT-DC UTILISATION

4.1 The case for bypassing the inverter

When an inverter fails and cannot be repaired, the panels remain functional. Rather than leaving them idle, they can be connected directly to DC loads or battery-charging systems without an inverter. This extracts useful work from the panels even though it does not produce grid-compatible AC power.

Direct DC use is not a complete substitute for an inverter. It is lower-efficiency (without MPPT), limited to specific applications, and requires rewiring. But it preserves the value of the panel asset — a panel producing DC power for water heating is better than a panel producing nothing.

4.2 Applications suitable for direct DC

Battery charging: Solar panels can charge lead-acid batteries (Doc #35) through a charge controller — a simpler device than an inverter, containing power transistors, a microprocessor, and sensing circuitry. Though simpler, charge controllers still require semiconductor components that NZ cannot manufacture; existing stocks from solar, marine, and camping suppliers are the primary source, supplemented by salvage from non-repairable solar systems. Basic PWM charge controllers (lower efficiency, 70–80% power utilisation) are more widely available than MPPT controllers (higher efficiency, 90–95% utilisation but more complex electronics).²⁴ Battery-stored energy can then be used for lighting, communications, and small loads via a separate battery inverter or directly as DC.

Resistive water heating: Connecting panels directly to a DC-rated water heating element. The panel’s output varies with sunlight, but water temperature is inherently storage — the water gets hotter when the sun shines and stays warm when it does not. No power electronics are required. A 2–4 kW array connected directly to a hot water cylinder element provides meaningful water heating. The absence of MPPT means some power is lost (the panel operates at whatever voltage the heating element demands), but for a purely resistive load this loss is typically 10–25% depending on the match between element resistance and panel characteristics.²⁵ Compared to a grid-tied inverter system heating via a standard AC element, direct-DC heating also loses the ability to divert surplus energy to the grid — the water cylinder becomes the sole energy sink, and any generation beyond what the cylinder can absorb is wasted.

DC lighting: 12V or 24V DC LED lighting systems, powered from panels via a charge controller and small battery bank. Common in marine, camping, and off-grid applications. LED drivers for DC input are simpler than AC drivers.

DC motor drives: Some motors (particularly permanent-magnet DC motors used in pumps, fans, and small tools) can run directly from solar DC. Variable-speed operation with varying irradiance may be acceptable for applications like water pumping (pump faster when the sun is brighter). However, only DC motors are suitable — the majority of NZ’s installed motor base is AC induction motors, which cannot run on DC. Suitable DC motors are found primarily in automotive components (windscreen wipers, fuel pumps, seat motors), battery-powered tools, and marine equipment.

Telecommunications and radio: HF radio equipment (Doc #128) and other communications gear typically operates on 12V or 24V DC. Solar panels with a charge controller and battery bank provide an ideal power source.

4.3 What direct DC cannot do

Most NZ household appliances and industrial equipment run on 230V AC. Refrigerators, washing machines, power tools, milking machines, welding equipment, and the national grid itself all require AC power. Direct DC from solar panels cannot power these loads without an inverter. The direct-DC strategy is a partial solution — it preserves some value from panels whose inverters have failed, but it does not replace grid-tied solar generation.

---

5. SIMPLIFIED INVERTER DEVELOPMENT

5.1 The design spectrum

Modern grid-tied inverters are sophisticated machines. They are not the only way to convert DC to AC. There is a spectrum of inverter complexity:

Square-wave inverter (simplest): Four switches (transistors or even mechanical relays) in an H-bridge configuration, switching at 50 Hz. Output is a crude 50 Hz square wave — far from a sine wave. Harmless for resistive loads (heaters, incandescent lighting). Damaging or unusable for motors, transformers, and sensitive electronics. Can be built from salvaged power transistors, a timing circuit (555 timer IC or discrete oscillator), gate driver components, a heat sink (aluminium, readily fabricated), and a DC bus capacitor — all sourced from existing NZ electronic component stocks or salvaged from failed inverters, UPS systems, and industrial motor drives. The transistors themselves (rated for the target voltage and current) are the binding constraint, as NZ cannot manufacture semiconductors. Feasibility: [A] — achievable now from NZ-available components and salvage, contingent on transistor stocks.²⁶

Modified sine wave (quasi-sine): An H-bridge with a more complex switching pattern that approximates a sine wave using stepped voltage levels. Better than square wave — most motors and transformers tolerate it, though with 10–20% increased heat losses in motors and audible buzzing in transformers. Some sensitive electronics (medical devices, precision instruments, modern refrigerator compressor controllers) may not function correctly or may be damaged. Can be built from discrete components with a more sophisticated control circuit (requiring additional ICs or a simple microcontroller). Feasibility: [B] — achievable in Phase 3–4 with development effort.

Pure sine wave (grid-quality): PWM switching at high frequency with output filtering to produce a clean sine wave. Required for grid connection and for sensitive loads. Requires high-performance power semiconductors, a microprocessor running control algorithms, and quality output filters. This is what modern inverters produce. Feasibility: [C] — the power electronics are the hard part. NZ cannot produce IGBTs or power MOSFETs. Locally manufactured pure-sine-wave inverters would depend on stockpiled or imported semiconductors.

5.2 What NZ could build

A realistic development pathway:

Phase 2–3 (Years 1–7): Square-wave and modified-sine-wave inverters for stand-alone (off-grid) applications, built from salvaged power transistors and discrete components. These would power resistive loads, basic lighting, and robust AC motors (water pumps, fans, workshop equipment). Output power: 500 W to 5 kW. Not grid-tied — these would serve individual installations or small local networks.

Phase 3–4 (Years 3–15): Improved modified-sine-wave inverters with better output filtering. Potentially capable of powering a wider range of loads. Still not grid-compatible — the frequency stability and power quality required for grid connection exceed what is achievable without precision control electronics.

Phase 5+ (Years 15+): If semiconductor manufacturing or import capability develops (Doc #115), pure-sine-wave inverters using locally assembled power stages and imported/domestic semiconductor devices. Grid connection might become feasible at this stage.

Grid connection is the hard requirement. A stand-alone inverter powering a local load needs to produce adequate voltage and frequency, but tolerances are relatively loose — a 5% frequency variation matters little for a water pump. A grid-tied inverter must synchronise precisely to the grid’s 50 Hz and inject current in phase, meeting strict harmonic distortion limits and anti-islanding requirements. This level of control requires the microprocessor-based systems that NZ cannot manufacture. Locally built inverters will serve off-grid and stand-alone applications; grid connection will require repair and preservation of existing inverters or importation of control electronics.

5.3 The component bottleneck

Even simplified inverter designs require power transistors — devices rated to switch significant current (10–50+ amps) at moderate voltage (200–600V) without excessive losses. NZ cannot produce these. The available sources:

• Existing stock: Electronic component distributors, solar companies, and electrical suppliers hold some power semiconductor inventory. Unlikely to be large.
• Salvage: Failed inverters, variable-speed drives (VSDs) in industrial equipment, uninterruptible power supplies (UPS systems), and electric vehicle chargers all contain power semiconductors. A failed inverter may have one or two dead transistors and several good ones — salvaging the good components for use in other repairs or new builds extends the useful life of the component pool.
• Other electronic equipment: Power semiconductors of varying specifications exist in welding machines, induction cooktops, motor drives, and many industrial control systems. Not all are suitable for inverter use, but the pool of available devices is larger than the solar industry alone.
• Trade: Power semiconductors are small, lightweight, and high-value — ideal trade goods. If maritime trade develops (Doc #140), electronic components should be a priority import.

---

6. BATTERY-EQUIPPED SYSTEMS: A SPECIAL CASE

6.1 Why these matter disproportionately

Solar systems with battery storage (typically lithium-ion with a hybrid inverter) have capabilities that grid-only systems lack:

• Islanding: They can disconnect from the grid and continue operating autonomously. When the local grid fails (transformer failure, line damage), a battery-equipped solar home retains power.
• Self-consumption: They store daytime solar generation for evening use, reducing grid dependence.
• Backup power: In an outage, they provide uninterrupted power to designated circuits.

Under recovery conditions, these attributes have outsized practical value. A community with several battery-equipped solar homes has a resilient power source independent of the grid. The hybrid inverters in these systems are more complex than simple grid-tied inverters, but they also include battery management, charge control, and islanding functionality — a more complete energy system.

6.2 Battery lifespan considerations

Lithium-ion batteries in residential solar installations are typically warranted for 10 years or 60–80% retained capacity.²⁷ Actual lifespan depends on depth of discharge, temperature, and cycling rate. Under recovery conditions — where the battery may be cycled more aggressively due to grid instability — degradation may be faster than warranted.

When the lithium-ion battery reaches end of life, the system can potentially be converted to use lead-acid batteries (Doc #35) with a compatible charge controller. Lead-acid is heavier, lower energy density, and shorter-lived, but it is domestically producible. The hybrid inverter continues to function — the battery chemistry is managed by the charge controller, not the inverter itself.

6.3 Prioritisation

Battery-equipped solar systems should receive priority for inverter repair and maintenance, ahead of grid-only systems, because their grid-independent capability provides more recovery value per installation. Marae with solar installations also warrant priority maintenance: under recovery conditions, marae function as community coordination hubs (Doc #150), and electrical power at these facilities supports food preparation, communications, lighting, and community gathering. The national asset census should specifically identify and locate all battery-equipped solar systems and solar-equipped marae.

---

7. WORKFORCE AND SKILLS

7.1 Existing NZ capability

NZ’s solar industry workforce includes:

• Solar installers: 2,000–5,000 workers involved in system design, installation, and commissioning. Most are qualified electricians with solar-specific training. Their skills are primarily in DC and AC wiring, panel mounting, and basic system commissioning — not inverter electronics.
• Electrical engineers and power electronics specialists: A smaller pool of engineers who understand inverter design, control theory, and power semiconductor applications. Found at universities, in the solar industry, and in the broader electrical manufacturing and drives industry.
• Electronics technicians and hobbyists: A significant community of people with component-level electronic repair skills — from professional repair technicians to amateur radio operators (Doc #128) and electronics hobbyists. These individuals have soldering skills, test equipment, and the diagnostic mindset needed for inverter repair.
• Inverter manufacturer representatives: Some inverter manufacturers (Fronius, Enphase, SMA) have NZ-based service and support staff. These individuals hold manufacturer-specific repair knowledge that is irreplaceable once they leave the workforce.

7.2 Training requirements

Inverter repair technicians (priority training target):

Training draws from electricians and electronics technicians. Curriculum:

• Power electronics fundamentals: how switch-mode converters and H-bridges work
• Inverter architecture: DC bus, power stage, output filter, control board
• Diagnostic procedures: oscilloscope use, capacitor ESR testing, thermal imaging, fault code interpretation
• Soldering and rework: surface-mount and through-hole component replacement
• Safety: DC arc flash hazards, stored energy in capacitors (lethal voltages persist after shutdown), working at height (roof-mounted equipment)

Training period: 3–6 months for technicians with existing electrical or electronic background. Apprenticeship model under experienced power electronics practitioners.

Direct-DC system installers:

Simpler training — existing electricians can be trained in DC system design and installation in 1–2 weeks. Key topics: DC wiring standards, charge controller selection and configuration, battery safety (Doc #35), overcurrent protection for DC circuits (DC arcs are harder to extinguish than AC arcs). This training should extend to community members in rural and Māori communities, where basic solar maintenance and direct-DC utilisation skills reduce dependence on external technicians and centralised support.

---

8. CRITICAL UNCERTAINTIES

Uncertainty	Impact	Mitigation
Total NZ solar fleet size and condition	Determines scale of maintenance programme	National asset census (Doc #8)
Inverter failure rate under NZ conditions	Determines repair demand timeline	Monitor fleet, establish repair database
Electrolytic capacitor stocks in NZ	Determines how many inverters can be repaired	National inventory, salvage programme
Nuclear winter irradiance reduction	Determines solar output for 3–7 years	Monitor with pyranometers at meteorological stations
UV effect on panel degradation rate	Determines long-term panel output trajectory	Monitor panel output over time, inspect for EVA yellowing
Feasibility of simplified inverter manufacture	Determines whether lost capacity can be partially recovered	Development programme beginning Phase 3
Power semiconductor availability from salvage and trade	Determines scale of any new inverter production	Inventory, salvage programme, trade prioritisation
Battery system conversion to lead-acid	Determines whether battery-solar systems outlast lithium-ion batteries	Testing programme (Doc #35)

---

CROSS-REFERENCES

• Doc #1 — National Emergency Stockpile Strategy (solar components as strategic assets; critical personnel classification)
• Doc #8 — National Asset and Skills Census (solar fleet inventory; power electronics skills identification)
• Doc #35 — Battery Management and Lead-Acid Production (battery charging from solar DC; lead-acid as lithium-ion replacement)
• Doc #41 — UV Protection (ozone depletion effects on panel degradation)
• Doc #65 — Hydroelectric Station Maintenance (complementary generation; grid context)
• Doc #67 — Transpower Grid Operations (grid connection requirements; distributed generation value)
• Doc #68 — Rural Distribution and SWER (distribution network context for rooftop solar)
• Doc #69 — Transformer Rewinding and Fabrication (distribution transformer failure affects grid-tied solar; distributed generation reduces transformer loading)
• Doc #72 — Micro-Hydro Design and Construction (complementary distributed generation)
• Doc #89 — NZ Steel Glenbrook (potential source of steel and aluminium for mounting structures)
• Doc #91 — Machine Shop Operations (fabrication support for inverter enclosures and mounting hardware)
• Doc #128 — HF Radio Network (DC solar power for radio equipment)
• Doc #130 — Device Life Extension (electronic component salvage and repair skills)
• Doc #134 — Computing Self-Sufficiency Roadmap (long-term semiconductor context)
• Doc #138 — Sailing Vessel Design (trade route for electronic components)
• Doc #138 — NZ–Australia Relations (trade source for electronic components and solar spares)
• Doc #157 — Trade Training Priorities (electrician and technician training pipeline)
• Doc #162 — University Reorientation (power electronics research and education)

---

FOOTNOTES

---

1. NZ installed solar PV capacity: The Electricity Authority’s Electricity Market Information (EMI) database tracks distributed generation connections. https://www.emi.ea.govt.nz/ As of late 2024, NZ had surpassed approximately 50,000 grid-connected solar installations. Total capacity figures are approximate because the EMI database includes connection data but capacity figures for all installations are not always complete. The Sustainable Electricity Association of New Zealand (SEANZ) publishes regular market reports with installation statistics. The 400–600 MW estimate should be verified against the most recent EMI and SEANZ data. NZ solar generation typically represents 2–4% of total electricity generation but is growing rapidly.↩︎

2. Solar panel degradation rates: Jordan, D.C. and Kurtz, S.R., “Photovoltaic Degradation Rates — An Analytical Review,” NREL, 2012. https://doi.org/10.1002/pip.1182 — This study of nearly 2,000 degradation measurements found a median degradation rate of approximately 0.5% per year for crystalline silicon modules, with a range of 0.3–0.7% for most installations. More recent panels (manufactured after ~2015) may degrade more slowly due to improved cell interconnection and encapsulant chemistry, though long-term field data is still accumulating.↩︎

3. Inverter lifespan and warranty: Industry data from major inverter manufacturers. Fronius, SMA, and Enphase offer standard warranties of 5–10 years with extensions to 15–20 years. The 10–15 year field life estimate is based on industry experience and failure analysis literature. See: Golnas, A., “PV System Reliability: An Operator’s Perspective,” IEEE Journal of Photovoltaics, 2013. https://doi.org/10.1109/JPHOTOV.2012.2215015↩︎

4. NZ solar industry workforce: Estimated from industry size. The Sustainable Electricity Association of New Zealand (SEANZ) and the Electrical Workers Registration Board may have more precise data. The 2,000–5,000 range includes installers, designers, sales staff, and support workers. The subset with inverter-level electronic repair skills is considerably smaller — probably in the low hundreds.↩︎

5. Installed cost estimate: Based on approximate installed costs of NZ$5,000–15,000 per residential system (3–10 kW) and proportionally for commercial systems, across 50,000–70,000 installations. This is a rough estimate of total historical investment. The value to the recovery is better measured in ongoing electricity production than in pre-event dollar cost.↩︎

6. NZ installed solar PV capacity: The Electricity Authority’s Electricity Market Information (EMI) database tracks distributed generation connections. https://www.emi.ea.govt.nz/ As of late 2024, NZ had surpassed approximately 50,000 grid-connected solar installations. Total capacity figures are approximate because the EMI database includes connection data but capacity figures for all installations are not always complete. The Sustainable Electricity Association of New Zealand (SEANZ) publishes regular market reports with installation statistics. The 400–600 MW estimate should be verified against the most recent EMI and SEANZ data. NZ solar generation typically represents 2–4% of total electricity generation but is growing rapidly.↩︎

7. NZ installed solar PV capacity: The Electricity Authority’s Electricity Market Information (EMI) database tracks distributed generation connections. https://www.emi.ea.govt.nz/ As of late 2024, NZ had surpassed approximately 50,000 grid-connected solar installations. Total capacity figures are approximate because the EMI database includes connection data but capacity figures for all installations are not always complete. The Sustainable Electricity Association of New Zealand (SEANZ) publishes regular market reports with installation statistics. The 400–600 MW estimate should be verified against the most recent EMI and SEANZ data. NZ solar generation typically represents 2–4% of total electricity generation but is growing rapidly.↩︎

8. NZ solar irradiance: NIWA (National Institute of Water and Atmospheric Research) maintains NZ solar radiation data. https://niwa.co.nz/our-science/climate — NZ’s solar resource varies from approximately 1,300 kWh/m²/year in Southland to approximately 1,700 kWh/m²/year in Marlborough, Nelson, and Bay of Plenty. See also: Anderson, T. et al., “New Zealand Solar Energy Resource Assessment,” NIWA, for detailed maps and regional data.↩︎

9. Nuclear winter solar irradiance effects: Robock, A. et al., “Nuclear winter revisited with a modern climate model and current nuclear arsenals: Still catastrophic consequences,” Journal of Geophysical Research, 2007. https://doi.org/10.1029/2006JD008235 — Southern Hemisphere irradiance reduction depends on stratospheric aerosol transport from Northern Hemisphere detonation sites. For a major NATO-Russia exchange, models suggest Southern Hemisphere surface irradiance reduction of 10–30%, peaking 1–2 years after the event and persisting for 5–10 years as aerosols gradually settle. NZ’s position at 35–47°S provides partial shielding from the worst effects.↩︎

10. Crystalline silicon market share: International Technology Roadmap for Photovoltaic (ITRPV), published annually by VDMA. As of 2024, crystalline silicon (mono and poly) accounts for approximately 95%+ of global PV production. Thin-film technologies (CdTe by First Solar, CIGS) represent the remainder.↩︎

11. Light-induced degradation: Caused by boron-oxygen complex formation in boron-doped p-type silicon under illumination. First reported by Fischer and Pschunder (1973). Modern panels with gallium-doped silicon or n-type cells show reduced LID. Most NZ-installed panels are p-type boron-doped and will have already experienced LID in their first year of operation.↩︎

12. Solar panel degradation rates: Jordan, D.C. and Kurtz, S.R., “Photovoltaic Degradation Rates — An Analytical Review,” NREL, 2012. https://doi.org/10.1002/pip.1182 — This study of nearly 2,000 degradation measurements found a median degradation rate of approximately 0.5% per year for crystalline silicon modules, with a range of 0.3–0.7% for most installations. More recent panels (manufactured after ~2015) may degrade more slowly due to improved cell interconnection and encapsulant chemistry, though long-term field data is still accumulating.↩︎

13. Potential-induced degradation: Pingel, S. et al., “Potential Induced Degradation of Solar Cells and Panels,” 35th IEEE Photovoltaic Specialists Conference, 2010. PID is influenced by system voltage, humidity, temperature, and panel construction. It is more prevalent in high-voltage string configurations and humid environments. NZ’s moderate humidity and relatively short string lengths (due to smaller system sizes) reduce but do not eliminate PID risk.↩︎

14. Nuclear winter solar irradiance effects: Robock, A. et al., “Nuclear winter revisited with a modern climate model and current nuclear arsenals: Still catastrophic consequences,” Journal of Geophysical Research, 2007. https://doi.org/10.1029/2006JD008235 — Southern Hemisphere irradiance reduction depends on stratospheric aerosol transport from Northern Hemisphere detonation sites. For a major NATO-Russia exchange, models suggest Southern Hemisphere surface irradiance reduction of 10–30%, peaking 1–2 years after the event and persisting for 5–10 years as aerosols gradually settle. NZ’s position at 35–47°S provides partial shielding from the worst effects.↩︎

15. Temperature coefficient of crystalline silicon: Standard crystalline silicon PV cells have a power temperature coefficient of approximately -0.3 to -0.5%/°C, meaning output increases as temperature decreases below the standard test condition temperature of 25°C. This is a well-established characteristic documented in panel datasheets and PV engineering references.↩︎

16. Inverter operating principles: Standard power electronics textbook material. See: Mohan, N. et al., “Power Electronics: Converters, Applications, and Design,” Wiley; or Erickson, R.W. and Maksimovic, D., “Fundamentals of Power Electronics,” Springer.↩︎

17. MPPT efficiency vs. direct connection: Without MPPT, a load forces the panel to operate at a voltage determined by the load’s impedance rather than at the panel’s optimal power point. The power loss depends on how well the load impedance matches the panel’s maximum power point impedance. For poorly matched loads, extraction may drop to 50–70% of available power. See: Duffie, J.A. and Beckman, W.A., “Solar Engineering of Thermal Processes,” Wiley, Chapter 23 on photovoltaic-load coupling.↩︎

18. Electrolytic capacitor aging: Capacitor life models are published by major manufacturers (Nichicon, Nippon Chemi-Con, TDK). The halving of life per 10°C above rated temperature is a well-established approximation based on Arrhenius-type electrolyte evaporation kinetics. See manufacturer application notes on capacitor lifetime prediction.↩︎

19. Power semiconductor failure mechanisms: Ciappa, M., “Selected failure mechanisms of modern power modules,” Microelectronics Reliability, 2002. https://doi.org/10.1016/S0026-2714(02)00042-2 — Discusses bond wire fatigue, solder fatigue, and die attach degradation as primary failure mechanisms in power semiconductor modules.↩︎

20. NZ lightning activity: NIWA maintains a lightning detection network. NZ averages approximately 40,000–100,000 cloud-to-ground lightning strikes per year, concentrated in the upper North Island during summer. This is moderate by global standards — far less than tropical regions but sufficient to cause regular equipment damage. Surge protection devices (SPDs) in inverters have finite life and degrade with each surge event.↩︎

21. Inverter lifespan and warranty: Industry data from major inverter manufacturers. Fronius, SMA, and Enphase offer standard warranties of 5–10 years with extensions to 15–20 years. The 10–15 year field life estimate is based on industry experience and failure analysis literature. See: Golnas, A., “PV System Reliability: An Operator’s Perspective,” IEEE Journal of Photovoltaics, 2013. https://doi.org/10.1109/JPHOTOV.2012.2215015↩︎

22. Inverter repair success rates: These estimates are based on field experience reported by power electronics repair professionals and solar industry service departments. Published data on inverter component-level repair success rates under field conditions is limited. The rates cited assume availability of correctly rated replacement components and a technician with board-level soldering skills and access to basic test equipment (multimeter, oscilloscope, ESR meter). Success rates decline significantly when substitute components of different specifications must be used.↩︎

23. Inverter repair success rates: These estimates are based on field experience reported by power electronics repair professionals and solar industry service departments. Published data on inverter component-level repair success rates under field conditions is limited. The rates cited assume availability of correctly rated replacement components and a technician with board-level soldering skills and access to basic test equipment (multimeter, oscilloscope, ESR meter). Success rates decline significantly when substitute components of different specifications must be used.↩︎

24. Charge controller types and efficiency: PWM (pulse-width modulation) charge controllers are simpler, cheaper, and more widely stocked, but waste energy when the panel’s maximum power point voltage is significantly above the battery voltage — typical efficiency 70–80% of available panel power. MPPT charge controllers use a DC-DC converter to match the panel’s optimal voltage to the battery, achieving 90–95% efficiency, but require more sophisticated electronics including an inductor, power MOSFET, and control IC. See: Messenger, R.A. and Abtahi, A., “Photovoltaic Systems Engineering,” CRC Press, for comparative analysis.↩︎

25. MPPT losses with direct-connected resistive loads: A resistive load (such as a water heating element) presents a fixed resistance to the solar panel, forcing it to operate at a voltage determined by the load resistance and the panel’s current output. This operating point is generally not at the panel’s maximum power point, resulting in a power loss typically estimated at 10–25% depending on the match between load resistance and panel characteristics. A well-matched element (resistance chosen to approximate the panel’s maximum power point at typical operating conditions) can reduce losses to 10–15%; a poorly matched element may lose 20–25%. See: Duffie, J.A. and Beckman, W.A., “Solar Engineering of Thermal Processes,” Wiley, for analysis of directly coupled PV-load systems.↩︎

26. Square-wave inverter construction: An H-bridge inverter using four power transistors (or even electromechanical relays for very low power) switched at 50 Hz is among the simplest power conversion circuits. It was the standard inverter topology before pulse-width modulation became practical with fast-switching semiconductors. The output is a square wave that delivers the correct RMS voltage and frequency for resistive loads but produces significant harmonic distortion unsuitable for many AC devices. Design details are available in any introductory power electronics text.↩︎

27. Lithium-ion battery warranties: Manufacturer warranties for residential energy storage systems (Tesla Powerwall, BYD, LG Chem/LG Energy Solution, Enphase) typically guarantee 60–80% retained capacity after 10 years or a specified number of cycles (typically 4,000–6,000 cycles). Actual field performance data for the current generation of residential batteries is still limited, as the market is relatively young.↩︎