In-plane centroids of textile composites are simulated as cross-correlated random fields. Each tow position is defined as an average trend quantified from experimental data, added with zero-mean deviations produced as a stochastic field. Realisations of these fields are generated using a framework based on the Karhunen-Lo`eve series expansion that is calibrated with experimental information from prior work. Positional deviations are obtained that are correlated along the tow and between neighbouring tows. The application is a 2/2 twill woven carbon fibre reinforced epoxy consisting of multiple unit cells. Generated in-plane deviations of the warp and weft tows resem-ble the experimental fluctuations with similar wavelengths. Simulation of thousand specimens demonstrates that the virtual in-plane positions possess the experimental standard deviation and correlation lengths on average.