Protecting Infrastructure – Cyber, Physical, and EMP Attacks

News feeds warn of cyberattacks on the infrastructure that underpins the U.S. economy, including electric, water, and gas utilities, data centers, cellular networks, and major food suppliers. Terrorists using exploding drones to destroy substations serving this infrastructure is a new tactic that could happen close to home. When the lights, Wi-Fi, and cellular services go out, it may not be as simple as waiting for them to come back on. When all the city noises and cars stop, that silence may indicate the occurrence of an electromagnetic pulse (EMP).

The Grid

Imagine 3,500 spiders, each with their own style, getting together to create a giant web. Now imagine that the web is the power grid, comprising thousands of power companies with thousands of generators, sensors, control and communication systems, trading and billing platforms, transmission and distribution systems, and 140 million customers. With this complexity comes vulnerability to combinations of cyber, physical, and electromagnetic attacks on the power grid.

• Cyberattacks can cause regional or nationwide blackouts, as in Ukraine and India. When hackers weaponize substations to create large power spikes, they can induce physical damage to equipment such as motors and computers.
• Physical attacks like shooting holes in transformer oil tanks have become frequent enough that utilities are erecting ballistic protection around substations. While this protection helps prevent rifle attacks, it does not contemplate the use of bomb-dropping drones attacking from above. These drones are being used extensively in conflicts globally and by criminal gangs in Mexico. Bomb-dropping drones used against substations or grid control centers are possible.
• Electromagnetic attacks on the grid can come from solar storms, radio frequency weapons, and high-altitude EMPs from nuclear detonations in the upper atmosphere.

Impact and Outage Duration

Adversaries may attack when they have a high likelihood of success and a significant impact. Each type of attack has a different level of impact.

Cyberattacks can create temporary outages for hours or days and induce physical damage to equipment. The damaged equipment could prolong the outage for weeks or months, depending on the equipment repair or replacement time. While cyberattacks are crafted for specific types of equipment, cyber weapons can have considerable collateral damage beyond the intended target. For example, Russian cyberattacks against Ukraine and Chinese attacks against India caused power outages from hours to days.

Physical damage to substations, such as the Metcalf substation attack, which caused $15 million in damage and shut down 17 transformers, can affect specific targets such as Silicon Valley. When a single site is attacked, power can often be rerouted from other substations to restore service. The attacked site may be down for months since the equipment replacement time for transformers is 18 to 24 months. Coordinated attacks on multiple substations could cause cascading grid failures.

Electromagnetic damage to equipment can be specific or nationwide. There are two types of EMP attacks. One type of attack uses a fast radiated pulse (called E1) from a detonation at an altitude of 46 miles, affecting an area 1,000 miles across. With this type of attack, at least two detonations would be needed to affect the entire U.S. The other uses a single detonation at an altitude of 75 miles to maximize the induced ground current (called E3) and can affect a 2,000-mile diameter.

Unintentional damage can occur from spikes on the grid caused by sudden changes in electrical demand. For example, the rapid charging of millions of vehicles at random times and locations could cause spikes in the power grid as the U.S. moves to electric vehicles. This concern will increase as 13 million commercial trucks, which require as much as one megawatt of power each, are connected to the grid. If not properly designed, spikes from large charging systems can damage nearby residential and business equipment.

Radio frequency (RF) weapons generate pulses through antennas targeted at specific sites. RF weapons can be hand-held or vehicle-mounted – on cruise missiles or drones. The EMPs these weapons generate transfer energy into data cables, power cables, or metal pathways on printed circuit boards. The intent of RF weapons is to upset or damage computer chips used in control systems or data centers, causing the equipment to malfunction or shut down, possibly for weeks.

Solar storms can damage equipment by inducing ground currents, which couple into equipment power lines and coaxial cables. According to NASA:

In 2013, Lloyds of London predicted that the most extreme space weather storms could affect 20-40 million people in the U.S. and cause up to $2.6 trillion in damages, with recovery taking up to two years.

Fortunately, disconnecting from long power lines and cable networks can minimize the effect of the solar storm on electrical equipment. Powering systems from microgrids provides effective mitigation for solar storms.

High-altitude EMPs from nuclear detonations in the upper atmosphere can affect entire continents. This means that EMPs can damage generators, transformers, control systems, and communication systems. They can also cause a collapse in demand by damaging equipment that uses electricity. It could take years to replace the heating, air conditioning, appliances, lighting, and computers across 140 million homes and businesses. Compared to the Lloyds of London estimate for the most extreme solar storm, the cost of an EMP event would be at least three times greater (140 million homes and businesses vs. 40 million people affected), or about $7.8 trillion in damages.

Prevention Beats Cure

Although it is impossible to prevent all hostile actions, reducing the impact of attacks is possible. Protecting water, power, communications, computing, healthcare, first responders, and transportation systems will be important for the economy to continue functioning after a catastrophic grid outage. By protecting 20% of these infrastructures each year, starting with the most critical, the impact of an attack would continually lessen. Whether the critical infrastructure is owned by private companies, local municipalities, or the federal government, approval by public utility commissions to recover the costs from ratepayers will be critical for the implementation of these protections. The U.S. could be resilient in five years with the protection of the following infrastructure totals:

• 153,000 public water systems (1,560 provide over 10 million gallons/day)
• 25,378 electric generators power the grid
• 55,000 substations
• More than 1,200 natural gas main compressor stations
• 4,238 banks with 70,644 branches
• 4 credit card processing companies with redundant data centers
• 2,653 data centers
• 417,215 cell sites
• 45 undersea cable landing sites in the U.S.
• 10 satellite teleports
• 6,120 hospitals
• 52,291 fire stations
• 17,985 police departments
• 34,661 food and beverage plants
• 40,000 grocery stores
• 147,900 waste collection vehicles
• 13 million commercial trucks
• 38,453 locomotives
• 360 ports
• Department of Defense, Department of Homeland Security, and Federal Bureau of Investigation biometric repositories
• 5,193 public airports, 14,776 private airports

Protection Steps – The Convergence of Physical and Digital (Cyber) Protection

Cyber protection of critical systems starts with rethinking the design of computers and networks, which are built in layers, like a stack of pancakes. The lowest layer is the physical equipment, and the top layer is the application consumers use. By making the lower layers secure, it is possible to better protect the upper layers of the stack. One of the lower layers is the binary layer, which runs processes (subroutines) in memory. If hackers can determine the memory locations of the processes, they can bypass the security controls and make computers do things they were not intended to do.

Equipment manufacturers can randomize the locations of these processes in memory so they are not in the same place on every device. This device uniqueness helps prevent the spread of malware. Uniqueness can also be the basis of device identity for zero-trust implementations. As manufacturers use computer chips to build control systems, they can include additional protections. Just as there are physical stops on a dial, control system manufacturers can add hardware limitations to prevent systems from being reprogrammed to run outside of safe operation.

Physical protection of critical systems can be enhanced by realizing that the homeland is now a contested space where saboteurs can destroy equipment. A change in thinking to address these new threats can help envision ways to prevent saboteurs from disrupting systems. Instead of building substations that are defenseless to bomb-dropping drones, utilities could consider placing critical substations underground and disguising the surface to make it difficult for drone operators to identify their targets. Adding non-conductive netting (e.g., to keep quadcopters from landing on transformers with explosives or to prevent dropped munitions from reaching the transformers) and RF counter-drone systems to existing substations might be possible. Non-conductive netting is intended to keep quadcopters from landing on transformers with explosives or to prevent dropped munitions from reaching the transformers. Counter-drone systems prevent drones from reaching their target using one of several approaches. They can take over the drone’s control channel, disable it, or use a defensive drone to physically capture it.

Protecting against high-altitude EMPs typically involves placing electronic equipment inside 1/4” thick, seam-welded plate steel enclosures with filter banks for the power connections. Filter banks use inductive and capacitive elements in a circuit to eliminate spikes by regulating both voltage and current (referred to as LC filters). As technology has advanced, there are additional options. Conductive materials made from nickel-coated carbon fiber can provide lightweight passive shielding. Advances have also been made in ultra-fast switching that can redirect EMPs on incoming power lines to ground before they can damage the equipment. EMP-shielded cabinets and surge suppression can protect control electronics, sensors for synchronizing the grid, operation centers, and communication systems. By installing low-voltage EMP surge suppression at business and residential meters, utilities can protect the end user equipment and preserve the demand for electricity so the grid can continue to function.

The Investment

Utilities and co-ops want to serve their customers and provide reliable service. Many are willing to install new technology, especially if it helps provide higher reliability and lower operating costs. Since utilities are rewarded with a return on assets, additional investment in infrastructure increases the utilities’ overall returns. Public utility commissioners act as regulators of the utilities on behalf of ratepayers. They want reliable service but keep a close watch on spending to maintain rates as low as possible. Increasing resilience and lowering the impact of cyber, physical, and electromagnetic attacks on infrastructure requires educating the utilities, the public utility commissioners, and the ratepayers. They need to know how the threats have evolved. Educating policymakers is also important. New legislation could encourage and empower public utility commissioners to approve resiliency investments that minimize the consequences of adversarial-caused disasters.