Orifice plates are often used in internal pipe networks to reduce pressure and flowrate, but their presence introduces flow recirculation and turbulence leading to potential vibration and noise generation. Turbulence generation is particularly sensitive to the geometric detail at the inside edge, or lip, of the orifice plate, and sharp corners can lead to large scale vortex shedding, leading to a self-sustained whistling instability. Recent studies have demonstrated that the acoustic response of such components is highly sensitive to minor changes in lip geometry. Introducing an optimal chamfer at the orifice lip has been shown to completely suppress the dominant instability, which in turn leads to strongly reduced acoustic energy. However, prior work has mainly focused on simple lip modifications, such as chamfers or fillets [1]. The influence of fully three-dimensional irregular lip shapes on acoustic absorption remains largely unexplored. To achieve this a mean turbulent flow through the orifice is computed via incompressible RANS and hybrid RANS-LES, with reference to prior numerical work by Benhamadouche et al. [2]. Small-amplitude acoustic perturbations are then superimposed and the linearised Navier–Stokes equations are solved [3]. This allows the prediction of the acoustic energy absorption or production in the frequency domain for a given configuration. Finally, the simulation workflow is set within a multi-objective Bayesian optimisation framework with the aim of maximising acoustic damping while minimising pressure drop. We will undertake a detailed evaluation of the flow physics in order to explore the potential of this approach in identifying novel solutions, with the aim to demonstrate potential of this framework in practical scenarios. These findings can provide the basis for an advance performance of pipe flow acoustics by providing new insight into sensitivity of orifice geometry to control flow-induced noise and instabilities.