Why did the RBMK accident at Leningrad Unit 1 in 1975 damage fuel, while Chornobyl Unit 4 in 1986 destroyed the reactor and contaminated a continent?
Alexander Rumyantsev’s 2009 memoir offers a disturbing explanation: between those accidents, the control rod design had been changed without maintaining the analytical continuity needed to understand the consequences.
Rumyantsev had worked on the RBMK’s spatial reactor physics before the first unit entered service. His calculations showed that low total reactor power could conceal severe local power concentrations. Neutron flux, boiling and rod movement could interact to overheat a small part of the core while the overall power indication remained relatively unremarkable.
Even the original shutdown system was problematic. Absorbers entering from above could shift the power distribution downward, concentrating power in the lower core before shutdown became effective.
According to Rumyantsev, a subsequent modification made this much worse.
The graphite displacers attached below the absorbers were shortened, leaving water columns beneath them when the rods were fully withdrawn. During initial insertion, graphite replaced that neutron-absorbing water. The shutdown system therefore added positive reactivity in the lower core, precisely where other effects could already be concentrating power.
A component intended to terminate an excursion could intensify it.
Rumyantsev argued that the earlier geometry might still have allowed an accident, but one comparable to Leningrad. The modified geometry, in his reconstruction, turned local damage into reactor destruction.
That counterfactual cannot be treated as proven. But the positive shutdown effect itself is well established. It was experimentally recognized in 1983, and effective corrective action still failed to follow before April 1986.
The organizational lesson is as important as the rod geometry.
Rumyantsev describes analytical work ending, knowledge and computational tools being lost, and attempts to obtain a justification for the modification going unanswered. The people making changes were no longer effectively connected to the analysis that should have constrained those changes.
A mechanical modification had altered the reactor’s transient physics. The safety justification had failed to follow it.
Rumyantsev estimated that the direct and indirect damage from Chornobyl exceeded, many times over, all capital investment in Soviet nuclear power. His memoir does not substantiate that accounting, but the scale of the consequences is beyond dispute.
Design authority must include ownership of the underlying physics. When core analysts fall out of the loop, an apparent improvement to a safety component can undermine the safety of the entire plant.
***
A reactor’s geometry can become fixed before its physics is sufficiently understood.
According to Alexander Rumyanchev’s account of RBMK development, the 25 cm graphite lattice pitch was selected on engineering design grounds before the analyses needed to assess its consequences were complete.
That dimension—the spacing between neighboring channel positions—helped
determine how the reactor would respond when cooling water turned into steam.
Rumyanchev describes calculations summarized in 1968 reports that compared three lattice pitches:
- 20 cm: A negative void reactivity effect. Increased resonance absorption outweighed the reduction in neutron absorption by water.
- 25 cm: A positive void reactivity effect at the expected equilibrium fuel composition, potentially exceeding the effective delayed neutron fraction.
- 30 cm: A negative void reactivity effect again, this time because increased neutron absorption in graphite outweighed the reduction in absorption by water.
These were results for the configurations his team analyzed. Nevertheless, the design implication was substantial: both a smaller and a larger pitch offered negative feedback from coolant boiling, through different physical mechanisms.
The selected 25 cm pitch favored achievable fuel burnup with the proposed 1.8% enrichment. But that favorable fuel economy came with an unfavorable reactivity response.
According to Rumyanchev, by the time this became clear, the principal RBMK design parameters had already been approved. Changing the graphite lattice pitch was considered impossible. Proposals to reduce the effective graphite density were also judged impractical.
The geometry had become a commitment before the safety implications were adequately understood.
This is a recurring problem in engineering. Analysis has its greatest value while its findings can still change the design. Once dimensions, interfaces, and manufacturing commitments become effectively irreversible, a fundamental safety finding can become a problem to manage within the chosen configuration.
The calculations may eventually be completed. The reports may be approved. Yet the opportunity to make the most consequential improvement has already passed.
Safety analysis must be part of choosing the basic design. Its timing matters as much as its technical quality.
A project must leave fundamental design decisions open until the analyses that could overturn them are ready.