Feng Kang

Feng Kang: The Man Who Invented the Finite Element Method Twice

Somewhere in the 1950s, isolated from Western engineering journals by Cold War politics and recovering from a bout of spinal tuberculosis that had confined him to self-directed study, a Chinese mathematician worked out, independently and from first principles, one of the most consequential numerical techniques of the twentieth century — the same method that American and European engineers were discovering, under a different name, at almost exactly the same time. Feng Kang never got the Western credit that came with a familiar name in a familiar journal. He got, instead, the quieter distinction of having built an entire national discipline around a discovery nobody told him someone else had also made.

From Electrical Engineering to a Sickbed Education in Mathematics

Feng was born September 9, 1920, in Nanjing, and raised in Suzhou, Jiangsu province. He enrolled at National Central University in 1939 to study electrical engineering, but after two years switched to physics, graduating in 1944 with a growing appetite for the mathematics underlying physical theory. That appetite deepened by accident: shortly after graduation he contracted spinal tuberculosis, a serious and immobilizing illness, and turned the recovery period into an intensive, self-directed course in advanced mathematics — training himself in areas well beyond what his formal physics degree had covered, largely through sustained independent reading rather than classroom instruction.

Building Chinese Mathematics from the Ground Up

He began teaching mathematics at Tsinghua University in 1946, then in 1951 was appointed assistant professor at the newly formed Institute of Mathematics of the Chinese Academy of Sciences. That same year he traveled to Moscow's Steklov Mathematical Institute, studying until 1953 under Lev Pontryagin, one of the twentieth century's most important figures in topology and control theory. Feng's early published work, through the 1950s, stayed within pure mathematics, concentrating on topological groups and Lie groups — abstract territory with no obvious connection to the applied, computational work that would define the rest of his career.

That changed in 1957, when he became associate professor at the Institute of Computer Technology and turned toward computational mathematics at a moment when China had almost no computing infrastructure and no established discipline in numerical methods to draw on. He became, in the words of colleagues who later established a prize in his name, "the founder and pioneer of Chinese computational mathematics," effectively creating the field domestically from a standing start. He went on to found and direct the Computing Center of the Chinese Academy of Sciences, serving as its first director from 1978 to 1987 and its honorary director afterward, and was elected an academician of the Chinese Academy of Sciences in 1980.

An Independent Discovery, Made Twice

Feng's most significant achievement came out of that turn toward computation: in the late 1950s and early 1960s he developed what he called the "finite difference method based on variation principles" — a technique for turning difficult partial differential equations into large systems of simpler algebraic ones by breaking a physical domain into small connected pieces. Engineers in the United States and Europe were developing essentially the same idea in the same period, publishing it under the name that eventually stuck internationally: the finite element method. Feng's version was arrived at independently, without access to the parallel Western literature, behind the practical isolation that separated Chinese and Western scientific communities during the period. The technique is now foundational to structural engineering, aerospace design, and physical simulation of every kind, and the Feng Kang Prize established in his memory explicitly credits his version of the discovery as "a milestone of computational mathematics" in its own right, not merely a footnote to the Western literature.

He did not stop there. In the 1970s he developed embedding theories for discontinuous finite element spaces and extended classical elliptic equation theory into higher dimensions, work that gave rigorous mathematical footing to methods used for analyzing composite elastic structures, and he originated the natural boundary element method, another numerical technique still used in engineering computation. After 1984, in the final decade of his career, he turned to an entirely different problem: constructing symplectic algorithms for simulating Hamiltonian systems, the equations that govern classical mechanics — work recognized with China's First Prize of the National Natural Science Award, one of the country's highest scientific honors, for placing long-term numerical simulation of physical systems on a sound mathematical footing that preserved their essential structure over time rather than accumulating drift.

Why Feng Is Called a Genius

The specific cognitive achievement at the center of Feng's reputation is independent discovery under conditions of near-total isolation: arriving, largely through self-teaching during an illness and subsequent work cut off from the Western engineering literature, at a numerical method that other researchers elsewhere needed an entire discipline's collaborative momentum to develop. Colleagues who created the prize bearing his name describe him plainly as the founder of an entire national field, and characterize his finite element work as a milestone in computational mathematics on its own terms — language that credits foundational, structure-building intellect rather than mere technical competence. The honest complication is that Feng's discovery, however independent, was not unique in world terms: the finite element method as commonly credited emerged from several converging lines of engineering and mathematical work in the 1950s, meaning his genius lay less in being first globally than in being first and alone, deriving a major mathematical tool from variational first principles without the benefit of the international conversation that shaped its Western development, and then building an entire research institution and generation of successors around it inside a country that had essentially no prior computational mathematics tradition to build from.

Legacy

Feng died on August 17, 1993. The following year, the Chinese Academy of Sciences established the Feng Kang Prize, awarded biennially to young Chinese researchers in computational mathematics and scientific computing — a deliberate act of institutional memory that keeps his name attached to the field he built rather than letting it recede behind the internationally standardized name of the method he co-invented. Every subsequent Chinese computational mathematician trained in a tradition that begins, more or less directly, with the discipline he assembled out of isolation, illness, and independent derivation.

Achievements

• Held posts at Tsinghua University