Beijing is in the midst of a pollution revolution. In order to decrease its record-busting particulate matter concentrations by 25% in the next three years, the city has declared a “war” on smog. Last week, IBM announced that its research center in China would be joining the fight, too.
Cutting smog pollution by a quarter is an ambitious goal, and one that will have to carefully negotiate the balance between Beijing’s booming economy and its rapidly declining air quality. So while the city government begins to control the number of cars on the road and shut down inefficient factories, IBM’s research center will launch a 10-year green computing project that aims to help the city make better decisions.
I spoke to Jin Dong, the lead engineer in charge of the partnership’s “Green Horizon” program about what kind of technology that entails.
IBM’s used satellite data, sensor networks, and drones to predict crop conditions up to 48 hours in advance, but never before has the company attempted to do the same with air pollution. Jin’s current task is to integrate data from Beijing’s 14 air-quality monitoring stations with satellite data in order to build accurate models that can anticipate pollution levels up to three days ahead of time. Another feature of IBM’s air-quality prediction model will be a map that can pinpoint sources of pollution down to street level.
“If we can forecast air pollution relevance with a long lead time, then we can provide a kind of decision support to the government, so that based on the economic and social impact, they can adjust for specific industries,” Jin said.
Still, it’s unclear if information about polluters will be made available to the public, and what, exactly, officials might choose to do with industrial offenders. The Chinese Ministry of Environmental Protection has at least dropped hints that it’s no longer playing Mr. Nice Regulator, though, in launching a drone pilot program to find factories violating emissions rules at night. When I asked IBM if drones might one day become a useful data stream in the program, spokesperson Jonathan Batty didn’t shut down the possibility.
“The more data you’ve seen in our analytic system, the more insight you’re going to get,” he said. “So while our project is focusing on sensors and satellite data, it could easily incorporate data feeds from elsewhere, such as drones or anything else that was seen as of value.”
In addition to predicting how much smog will show up on any given day, IBM’s researchers will be working on a system that anticipates the availability of wind and solar power. The “Hybrid Renewable Energy Forecasting” (HyRef) system uses turbine sensors, weather modeling, and cloud imaging to make predictions up to a month in advance.
It’s unclear how accurate those predictions become over time, but IBM’s already demonstrated a successful pilot with 30 renewable sources in China, increasing the amount of clean energy in a local grid by 10%. By anticipating how much solar and wind can come online, IBM’s system could one day help officials adjust the rest of their energy portfolio–perhaps one day turning down the flow of electricity from coal or natural gas.
Part of IBM’s goal is to leverage a Beijing-based Internet of things, creating a network of energy monitoring devices to inform massive computing engines. Where industry represents 70% of China’s energy consumption, Jin’s research team will be building a giant, Big Data-informed platform to manage industrial energy use. Jin adds that universities, local government, and private partners will all help coordinate the effort.
“I’m pretty excited to work with the whole society on these kinds of very pressing, but very exciting, very challenging problems,” he added.
But however advanced IBM’s technology becomes, the responsibility to hold polluters accountable still lies with the municipal government. Beijing also has control over how much of IBM’s findings are actually communicated to the public. Technology is one part of the solution, but in the end, governance will likely play a larger role.