Indian Space Research Organisation

This paper proposes a spectrally optimized, multichannel filter-bank channelizer with time-multiplexed hardware reuse and block-based [Formula: see text] per-sample FFT complexity based on a unified framework for deep-space and satellite-onboard digital signal processing. This framework achieves fine-grained channelization using a single prototype filter with substantially reduced memory and computational overhead. The proposed approach derives all analysis and synthesis subband filters from a single prototype FIR filter using dyadic dilation and structured cascading. This eliminates the need for independently designed subfilters, achieves uniform channel spacing with consistent group delay, and reduces spectral distortion across all bands. A multi-objective cost-function-based prototype filter was obtained for the multilevel tree-structured uniform filter bank by minimizing transition-band energy, stop-band energy, and hardware complexity under QPSK recovery constraints. Strong channel isolation, a sharp roll-off, and reliable reconstruction at every level of the hierarchy are demonstrated by comprehensive simulation-based spectrum analysis. These channelizers are well-suited for onboard digital processors in next-generation satellite and deep-space missions. Simulation results demonstrate an average adjacent-channel isolation exceeding 98dB (with a worst-case value of 79.6dB) for a 64-channel 10kHz frequency grid, while maintaining the QPSK error vector magnitude (EVM) below 10%. Fixed-point simulations and FPGA synthesis results confirm that the architecture maintains communication-relevant isolation (>60 dB worst-case) under practical finite word-length constraints, demonstrating realistic hardware feasibility.

Abstract Accurate prediction of turbulent heat transfer and pressure drop in rectangular microchannels remains limited, as most existing correlations are constrained by narrow geometric ranges, low Reynolds numbers, or a single working fluid. To address this gap, the present study develops new correlations for the local and mean Nusselt number and the Darcy friction factor in rectangular microchannels with three-sided heating under turbulent flow. To establish the new models, a database of 40 conjugate three-dimensional CFD simulations was generated in ANSYS Fluent using the SST k-ω turbulence model. The simulations span hydraulic diameters of 200?1000 µm, aspect ratios of 0.25-2, Reynolds numbers of 2500-12,000, and Prandtl numbers of 4-12.5, covering a substantially broader parameter space than previously available studies. The resulting correlations incorporate the effects of asymmetric heat input and wall conduction, and are validated against in-house experiments using both water and acetone. All predictions remain within the ±15% experimental uncertainty. The proposed closed-form correlations are straightforward to implement and offer reliable predictive capability for the design and optimization of microchannel heat sinks operating under realistic three-wall heating conditions.