30-Second Overview: On August 15, 2017, a natural gas supply disruption caused rotating blackouts across 17 counties and cities, affecting 6.68 million users; on March 3, 2022, an operational error at the Xingda Power Plant affected over 5.49 million households. Neither incident was a simple matter of "insufficient generation capacity"; rather, they demonstrated how a single equipment failure, a single procedure error, or a single hub malfunction can escalate into a nationwide event. Consequently, Taipower (Taiwan Power Company) has proposed a ten-year grid resilience plan involving NT$564.5 billion, shifting from a centralized grid toward a decentralized, robust, and defensive model. But the real question remains: when a blackout occurs, who—hospitals, communities, remote areas, or cities—can restore life first?

Taiwan's Power Problem is More Than Just Generation Capacity

Blackouts are most easily understood as a supply-and-demand issue: Is there enough electricity? Is the spinning reserve capacity high enough? Which energy source should be increased? But when a light goes out, it is not just illumination that vanishes. Elevators stop, trapping residents in high-rise buildings; base stations and network equipment require backup power, or families cannot communicate; hospitals must rely on generators to maintain surgeries, ventilators, and refrigerated medicines; cold chains, water pumping, traffic signals, and continuous industrial processes must all decide how to endure within a very short window.

Therefore, grid resilience is not about "never having an accident," but rather whether the system can limit the scope of a failure, maintain critical services, quickly identify the problem, and restore power after an incident occurs. Taipower's definition of resilience also includes the ability to respond to accidents and restore stable operation within a short period. ¹

This definition shifts the problem from "is there enough electricity" to four much harder questions: Is electricity concentrated in a few large nodes? Can a failure be isolated? Do local areas have their own backups? Who decides the order of repairs? If one only increases generation capacity without addressing transmission lines, substations, protective relays, operating procedures, and local distribution, a city may still lose widespread power due to a seemingly localized accident.

815: How One Valve Blacked Out Half of Taiwan

The "815" blackout on August 15, 2017, was the turning point when Taiwan's grid resilience was witnessed by society on a massive scale. An Executive Yuan investigation pointed out that the natural gas supply to the Datann Power Plant was suddenly interrupted because a contractor for CPC Corporation (Taiwan) used temporary jumper wires while replacing a power supply at a natural gas metering station; after removal, the controller's power and communication were interrupted, causing the controller to restart and automatically close the electric valve, ultimately leading to a complete gas supply interruption. ²

The core of the accident cannot be explained away by "one person pressing the wrong button." The investigation also noted that important electric valves shared controllers, the Datann Power Plant's gas supply was not sufficiently decentralized, and there was insufficient change management and risk assessment, allowing a local maintenance action to have consequences far exceeding its immediate scale. Following the natural gas interruption, units shut down, system power capacity plummeted, and finally, rotating blackouts affected 6.68 million users across 17 counties and cities. ³

The Environmental Information Center, in reviewing 815, noted that the impact of the accident was not just the blackout itself, but also reignited public debate regarding energy transition, reliance on natural gas, grid stability, and backup capabilities. ³ This debate is often simplified into supporting or opposing a specific type of energy, but another lesson provided by 815 is: regardless of whether electricity comes from gas, coal, nuclear, or renewable energy, as long as critical supply, control, and transmission/distribution systems lack multi-layered protection, a single error can still be amplified.

The Executive Yuan investigation therefore proposed directions such as implementing multi-layered protection, increasing fuel supply margins, strengthening the disaster resilience of gas and power supply systems, and improving independent power supplies for public transport and life-support systems. ² These recommendations bring "energy security" from inside the power plant walls out to the entire city: critical services cannot simply wait for the large grid to recover; they must have their own safety boundaries.

303: More Than Just an Operational Error

The "303" blackout on March 3, 2022, placed "procedures" and "organizational culture" at the center of grid discussions. An investigation by the Control Yuan noted that during the overhaul of Unit 2 at the Xingda Power Plant, procedures for isolators and auxiliary cards were not followed; testing was conducted before the circuit breaker's insulating gas had fully recharged, causing a 345kV busbar flashover to ground fault, which subsequently triggered system splitting and widespread blackouts. The accident occurred on the morning of the day in question, with full power restored by 9:31 PM, affecting over 5.49 million households during that period. ⁴

A review report released by the Ministry of Economic Affairs also pointed out that the reason a single operational error could cause a major blackout was related to insufficient multi-layered protection and the risk of the southern grid being concentrated at the Longqi extra-high voltage substation. Improvement directions include establishing dedicated risk management units, promoting decentralized grids, reviewing protective relays and switchyard systems, and strengthening professional training and external expert diagnosis. ⁵

There is a systemic issue here that is easily overlooked: if every unit is only responsible for its own local task and no one can see "how one operational error crosses multiple lines of defense," the system builds safety on the assumption that no one will ever make a mistake. True resilience is the opposite; it assumes people will err, equipment will fail, and weather will deteriorate; therefore, it must allow errors to be discovered and isolated rather than letting them penetrate the entire system.

After 303, Taiwanese society's expectations regarding blackouts became more precise. Citizens no longer just ask "why did it stop," but also "why does an accident at one power plant affect so many people," "why wasn't the same error stopped at the previous layer," and "why are the restoration order and information release not clear enough." These questions are essentially demanding that the grid become a public system that is understandable and subject to oversight.

From Centralized Efficiency to Decentralized Resilience

Taiwan's past grid architecture had its historical reasons. Taiwan is densely populated, land for power plants is hard to find, and electricity demand is concentrated in urban and industrial zones; thus, the grid gradually formed a centralized structure where large power plants are interconnected via north-south transmission lines and a few hub substations. Taipower explained that the existing system consists of three 345kV main trunk lines, with the Longtan, Zhongliao, and Longqi hubs connecting the northern, central, and southern regions. ¹

Centralization can bring efficiency and allow different regions to support each other; the problem is that when a hub becomes a bottleneck, efficiency can conversely become a common vulnerability. The plans proposed by the Executive Yuan and Taipower after the 303 accident both listed "reducing centralized risk" as a core objective. The Grid Resilience Construction Plan announced in 2022 plans a ten-year investment of NT$564.5 billion, with three main pillars: striving for decentralization, continuous reinforcement, and enhanced defense. ¹

Decentralization is not just about building more power plants. The plan includes direct supply from plants to science parks and industrial parks, decentralized green energy supply, increasing distribution nodes, grouping hub substations, expanding energy storage, moving substations indoors (undergrounding/enclosure), and real-time dynamic defense. ¹ The logic is to bring electricity closer to consumption centers, make fault points easier to isolate, and ensure that when an accident occurs on a main trunk line or at a substation, the entire region does not have to wait together.

Taipower stated that the grid resilience plan will involve building 28 additional substations and promoting the enclosure of some substations; meanwhile, by combining green energy with energy storage, decentralized energy will become not just a source of generation, but also a tool for regional power stability. ⁶ This direction is intertwined with the energy transition: renewable energy has decentralized potential, but wind and sunlight are volatile; without storage, dispatch, and distribution capabilities, decentralized sources do not automatically equal decentralized resilience.

Figure 1: Distribution resilience occurs not only at large power plants and extra-high voltage substations, but also in the transformers and feeders within neighborhoods. Photo by Yoxem, from Wikimedia Commons file page, licensed under Creative Commons Attribution-ShareAlike.

The Distribution Network: Where Blackouts Most Frequently Occur

While nationwide blackouts are most conspicuous, they are not the only form of blackout citizens encounter daily. Data compiled by the Office of Economic Advisors of the Executive Yuan from Taipower records shows that the number of distribution accidents causing blackouts dropped from 21,019 in 2012 to 6,115 in 2023—a 70% decrease over a decade. ⁷

These numbers demonstrate that grid resilience also happens on streets, utility poles, transformers, and feeders. There are over ten thousand feeders, approximately 410,000 kilometers of lines, 1.55 million switches, 1.48 million transformers, and 3.2 million utility poles across Taiwan; more than half of accident-related blackouts are related to external and environmental factors such as animals, trees, lightning, salt damage, construction, vehicles, or user equipment. ⁷

Therefore, resilience is not just engineering for extra-high voltage substations. Feeder automation can detect and isolate fault intervals when an accident occurs, restoring normal lines first; disaster-preventative undergrounding of lines can reduce blackouts caused by trees, strong winds, or foreign objects. ⁷ However, there are also choices in engineering: to what extent should densely populated areas or critical medical and transport zones be prioritized? For remote lines with long distances and far repair reaches, is a different evaluation method needed? Undergrounding all lines may not be the only answer; the key is to place limited resources where they can most effectively reduce public risk.

Microgrids: Bringing "Recovery" Back to Localities

While the large-scale grid remains the primary foundation of life in Taiwan, during disasters, whether a locality can temporarily operate independently from the main grid may determine whether shelters, communications, and basic lighting can be maintained. According to Taipower Monthly, areas such as Fushan in Wulai, Lijia in Alishan (Chiayi), mountainous regions in Pingtung, and outlying islands have established disaster-preventative microgrids that combine solar photovoltaics, energy storage, and energy management systems with "islanding" capabilities. ⁸

The "islanding operation" of a microgrid is not intended to make an entire township permanently self-sufficient, but rather to disconnect the local power system when the external grid is damaged, prioritizing the maintenance of shelters, communications, lighting, and essential equipment. This capability is especially important for remote areas, mountains, and islands, because when main trunk lines or roads are damaged, it takes longer for repair crews and equipment to arrive.

Taipower also collaborates with universities to promote microgrids, bringing technology into education, demonstration, and community support fields. ⁸ However, a microgrid is not completed simply by installing solar panels and batteries. It also requires knowing who is responsible for activation, which equipment is prioritized, how long fuel and batteries will last, how to operate during communication outages, and how to safely reconnect with the main grid after a disaster.

These questions make microgrids both an engineering and a governance issue. If there is only equipment without local maintainers, it may fail to activate during an accident; if only a few institutions know how to operate it, the community cannot treat it as shared emergency infrastructure. True local resilience must integrate technology, training, responsibility, and drills together.

Resilience Must Be Measurable, Not Just Claimed

Grid investment is often presented in terms of engineering quantities—such as how many new substations are added, how many lines are undergrounded, or how much energy storage is installed; but engineering quantities do not equal the safety of residents during an accident. A more complete assessment should include "probability of failure," "scope of fault propagation," "duration critical services can be maintained," and "whether information is understandable" within a single set of indicators. ¹⁷

Such indicators also prevent "averages from masking disparities." A decrease in the national average blackout time does not mean every community is equally safe; having generators in large hospitals does not mean surrounding long-term care facilities, high-rise buildings, or small businesses can endure the same length of a blackout. The goal of resilience governance should be to fill the gaps most likely to cause life-threatening situations or public service interruptions first, rather than merely pursuing a beautiful total average.

From Equipment Resilience to Institutional Resilience

Hardware upgrades for the grid must occur alongside the observability of institutions. Strengthening substations, undergrounding lines, and building energy storage can reduce the probability of failure, but they cannot alone answer three governance questions after an accident: Who has the authority to decide the restoration priority? Who must disclose accident information publicly? Who is responsible for turning experience into the next drill and rule?

The common lesson from 815 and 303 is that accidents often cross multiple organizational boundaries. Natural gas suppliers, power plant operators, transmission dispatchers, local governments, and public service agencies may each complete their own local tasks, yet no single unit sees the cross-system risks early on. Resilience, therefore, requires shared scenarios across agencies, clear escalation reporting, traceable change management, and investigation methods that look not just for individual responsibility after an accident, but also at institutional lines of defense.

For the public, transparency does not end with publishing a technical report. During a blackout, public information should at least explain the affected scope, the level of certainty regarding the current cause, the time of the next update, locations where drinking water and charging are available, and the assistance available to the elderly, people with disabilities, those dependent on medical equipment, and remote residents. If information only circulates in technical jargon, even if power is restored quickly, society may still lose trust due to uncertainty.

Grid Resilience Will Also Become a Social Distribution Issue

The impact of blackouts is never equal. People living in elevator buildings versus old apartments without elevators experience blackouts differently; those working from home, those needing refrigerated medicine, those dependent on electric medical equipment, and small shops running cold chains do not have the same backup capabilities. Large enterprises may have generators and Uninterruptible Power Supplies (UPS), while small businesses and low-income families can only wait for restoration.

Therefore, grid resilience plans cannot be measured solely by "average blackout time" or "national restoration percentage." One must also ask: Which critical public services have independent power supplies? Which communities have accessible shelters? Which remote areas and islands have microgrids or backups? Is blackout information released in a way everyone can understand? Is the restoration order transparent?

The Executive Yuan's investigation into the 815 accident has listed the independent power supply for public transport and life-support systems as an area for improvement. ² This indicates that the grid is not just a service quality issue for a single industry, but a common condition for healthcare, water supply, communications, transportation, and public safety. When blackouts occur simultaneously during extreme heat, typhoons, or disasters, power shortages can rapidly transform into health risks and social inequality.

Taiwan's Energy Transition Must Simultaneously Be a Backup Transition

Taiwan's energy transition is often debated between types of generation: the costs, risks, and political controversies of gas, coal, nuclear, and renewables. But grid resilience requires another perspective: energy choices must be planned alongside transmission/distribution architecture, energy storage, demand management, fuel supply, and local backups.

When renewable energy is combined with storage and microgrids, it becomes more than just a source of generation; it can also become a local safety boundary during disasters. This transition requires new maintenance capabilities: localities need to know which equipment remains available during a blackout, managers must be able to initiate islanding operations, power companies must be able to safely reconnect, and the government must incorporate these capabilities into disaster drills rather than waiting for an accident to confirm if equipment actually works.

Analysis from the Global Taiwan Institute points out that the blackout events of 2017, 2021, and 2022 exposed Taiwan's economic structure's high dependence on electricity, and noted that the core of the NT$564.5 billion plan proposed in 2022 is to reduce vulnerable bottlenecks by using decentralization and microgrids to lower the possibility of a single node causing a widespread blackout. ⁹

The Environmental Information Center, reviewing 815, reminded that blackout events involve gas reliance, grid stability, spinning reserves, and social trust in the energy transition. ³ If the energy transition only adds new generation sources without simultaneously explaining how electricity is transported, how it is stored, and how critical services are maintained during accidents, the public may still feel they have "chosen an energy source but haven't solved blackouts."

More mature energy governance should plan "how much electricity" and "where can be restored first" together. Renewable energy, storage, demand response, direct supply to plants, regional microgrids, and distribution automation are not mutually exclusive single answers, but a toolkit of tools to reduce different risks.

Energy transition therefore also requires new methods of public communication. The government should not only announce installed capacity and construction amounts, but also explain how new equipment operates during cloudy days, typhoons, fuel supply interruptions, or damaged communications; local residents need to know the safety distances of facilities, maintenance responsibilities, land-use impacts, and contact windows during accidents. When technical information can be translated into understandable life scenarios, energy policy has a chance to move from abstract supply-and-demand debates to risk management that can be jointly supervised.

After a Blackout, What Should the City Restore First?

The final question of grid resilience is not technical, but one of priority. After an accident occurs, it is impossible for all places to recover simultaneously. The restoration order for hospitals, water supply, communications, transport, shelters, food cold chains, and residences must be discussed in advance, rather than being decided on the fly in the dark. The 815 investigation also required a review of warning notifications for sectional rotating blackouts, tracking affected feeders and user counts, and gradually establishing independent power circuits for public transport, security, and life-support systems. ² These requirements remind us that restoration is not just a question of "how fast," but also an issue of information justice—whether the government can let people know "who is restored first and who is still waiting."

If the restoration order for critical services is not disclosed in advance, during a blackout, support may easily be prioritized for places with the most resources, the loudest voices, or those closest to decision-making centers. Conversely, if local governments and power companies regularly inventory medical, water, communication, shelter, and elderly care sites, and incorporate backup capabilities and restoration priorities into drills, resilience will not stop at an equipment list but can become a public commitment that can be verified by society.

Figure 2: Grid resilience lies not only in the equipment itself, but also in the organizations responsible for planning, maintenance, dispatch, and accident learning. Photo by Xuan Shisheng, from Wikimedia Commons file page, licensed under Creative Commons CC0 1.0 Public Domain.

This also requires the government, Taipower, local governments, medical institutions, communities, and industries to establish a common recovery plan. During a large-scale blackout, citizens need more than just "under repair" messages; they need to know how long it might last, where they can get water and charging, which roads and transport facilities are still usable, and who can receive priority assistance.

In recent years, Taipower has shifted from a centralized architecture toward decentralization, reinforcement, and defense, and distribution accidents have also shown a downward trend, indicating that long-term engineering can indeed improve power supply quality. ¹⁷ But the lessons left by 815 and 303 remain relevant: any system that lacks alternative paths, error isolation, and public review can turn a local problem into a city-wide crisis.

True grid resilience is not guaranteeing that blackouts will never happen, but ensuring that a blackout does not have to mean a total halt of life. It should ensure hospitals have power, remote areas have their own basic support, communities know how to care for the vulnerable, and the government can clearly explain who is restored first, why they were prioritized, and how to avoid the same mistakes next time.

As Taiwan invests ten years and NT$564.5 billion to redesign its grid, the most important outcome should not just be a new line or a new substation, but that after an accident occurs, more places can retain a bit of light, a bit of communication, and a bit of order until the entire city reconnects to power.

Further Reading:

• Taiwan's Oil Price Mechanism and CPC — Within the same NT$600 billion additional budget, alongside Taipower's NT$71.1 billion, are two entries for CPC: a state-owned company called to stand at the forefront, and who ultimately pays for the money they carry first.
• Discussions on Taiwan and Nuclear Power — The generation structure in the first non-nuclear year, the natural gas reserves of that "Hormoz" spring, and the two paths for whether Nuclear Plant No. 3 can return.

References

Images and Licensing

Both photos used in this article are from Wikimedia Commons, with original file pages, authors, and licensing information preserved. The first photo was taken by Yoxem under CC BY-SA; any cropping, modification, or redistribution must comply with the same attribution and share-alike requirements. The second photo was taken by Xuan Shisheng under CC0 1.0 Public Domain, allowing free copying, modification, and reuse.

🧬 Taiwan.md