Skip to content
Quantum.world

The Signal

Kaseb tests DelftBlue-to-Tuna-17 workflow through Quantum Inspire

TU Delft researcher Zeynab Kaseb tested an integrated workflow that sends quantum circuits prepared with DelftBlue through Quantum Inspire to the Tuna-17 processor. QuTech reports that a six-qubit encoding produced a lower best residual than a four-qubit run, but its best candidate appeared less often in the measurements.

Kaseb tests DelftBlue-to-Tuna-17 workflow through Quantum Inspire

A route from classical preparation to quantum hardware

TU Delft researcher Zeynab Kaseb became the first user to test an integrated workflow linking DelftBlue, the university’s high-performance computer, with Quantum Inspire and the Tuna-17 quantum processor. QuTech’s account, published on 13 November 2026, describes a small power-flow calculation prepared on DelftBlue, sent to the processor through Quantum Inspire and then analysed using its measurements.

The workflow pairs DelftBlue’s classical calculations with Quantum Inspire, QuTech’s open cloud quantum computer. DelftBlue prepares and optimises the calculation; Quantum Inspire sends circuits to Tuna-17 and returns the processor’s measurements for analysis. The test used this route to examine power flow in an electrical network, a calculation used to determine grid voltages.

Encoding a two-point grid problem

For the demonstration, Kaseb studied a small network with two connected points, known as buses. The method she explored was the Quantum Approximate Optimization Algorithm, or QAOA. It encodes possible answers as bitstrings and uses a quantum circuit to sample candidate solutions.

DelftBlue constructed the mathematical objective and optimised the circuit parameters using a simulator. Kaseb then submitted the circuits to Tuna-17 through Quantum Inspire. Measurements from the processor were decoded into candidate voltages and checked against the original power-flow equations.

Two circuit sizes, different results

Two versions of the calculation ran on Tuna-17, a 17-qubit quantum processor. Their circuits used four and six qubits, with 1,024 measurements in each run. An exhaustive classical calculation established the best candidate available for each encoding, providing a reference for assessing the hardware results.

QuTech reports that the best candidate in the four-qubit run appeared in 215 measurements, or about 21 per cent. The six-qubit encoding represented voltage more finely and yielded a lower best residual: 0.003262, compared with 0.01053 for the four-qubit version. A residual measures how closely a candidate satisfies the power-flow equations; zero is the optimum.

That encoding also required a deeper circuit: 165 layers, compared with 30 for the four-qubit circuit. QuTech says the six-qubit run’s best candidate appeared 16 times, or 1.56 per cent of its measurements. The account describes that frequency as close to what uniform sampling would produce across the 64 possible bitstrings.

Measurements returned for analysis

The test linked classical preparation and optimisation on DelftBlue with circuit execution on Tuna-17 through Quantum Inspire, followed by analysis of the processor’s measurements. QuTech presents the results as showing how candidate quality and the frequency with which a candidate appears can both be examined, alongside the effect of circuit complexity.

The test focused on a small, two-bus power system and compared two encodings on the quantum processor. Its measurements provided candidate voltages for comparison with the power-flow equations, while the exhaustive classical calculation supplied a reference for each encoding.

Source

  1. QuTech

Related stories

Explore

More articles