Applied Sciences

This study addresses the problem of improving the efficiency of fine grinding of bulk materials in an original-design double spring–rotor grinder equipped with a separating diaphragm with a variable discharge orifice. The purpose of the work is to determine rational operating parameters that ensure a balanced trade-off between grinding quality, throughput, and energy consumption. The methodology is based on a full-factorial experimental design (Hartley plan) with five controllable parameters—rotational speed, material filling ratio, overlap of the working zones, grinding chamber clearance, and grinding duration—followed by response surface modeling and multi-objective optimization. The main responses included grinding fineness, throughput, drive power, specific energy consumption, and specific metal intensity. Adequate second-order regression models were obtained (R² > 0.93), and analysis of variance confirmed the statistical significance of the main effects and interactions. Multi-objective optimization enabled the identification of operating regimes that increase throughput by 15–20% while reducing specific energy consumption by 8–12% compared with empirical settings. The proposed approach provides a quantitative basis for selecting compromise operating conditions and can be applied to the tuning and control of spring–rotor grinding equipment in processing industries.

Fungal spores are the main active ingredients in fungal preparations. In this study, we evaluated vegetative spore (oidia) production of the Latvian isolate of Phlebiopsis gigantea PG 182 using liquid-surface (LSF) and solid-state (SSF) fermentation processes in the 450 mL and 700 mL jars, respectively. The effects of medium depth (0.5 or 0.7 cm), malt extract (ME) syrup concentration (25, 50, and 75 g/L) and duration time of cultivation (7, 14, 21 and 28 days) on oidia production and partly on mycelium biomass yield were evaluated in the LSF experiments. The highest spore yields (0.88 ± 0.22) × 10⁷ and (1.10 ± 0.31) × 10⁷ (95% CI) (oidia/g liquid medium) were achieved on day 28 in the LSF process using a medium depth of 0.5 cm and ME concentrations of 25 and 50 g/L, respectively. While in the SSF process, pine sawdust enrichment with wheat bran (0, 5, 10, 15, and 25%) and cultivation time (14, 21 and 28 days) were evaluated under conditions of 8 cm substrate depth. The most promising result was obtained on day 28 with 10% bran supplementation, reaching (24.73 ± 5.09) × 10⁷ (95% CI) (oidia/g solid medium), which is 1.45 and 3.17 times more than after 21 and 14 days of cultivation, respectively. Our findings indicate that SSF with a 10% wheat bran additive produces superior spore yields for P. gigantea isolate PG 182, exceeding benchmarks set by comparable research. Potential for further improvement remains by optimizing the wheat bran (WB)-to-substrate ratio and refining the thermal processing of the solid substrate.

This study proposes an explainable machine learning framework for estimating the total project cost (TPC) of AI training-data construction, where cost information is difficult to structure due to heterogeneous workflows and quality requirements. Using 386 public AI training-data projects conducted between 2020 and 2022, we derive 24 numerical predictors from standardized final reports and construct three input tracks: a baseline feature set, a principal component analysis (PCA)-enhanced set, and a factor analysis (FA)–enhanced set capturing latent cost structures. Four regression models (Ridge, Random Forest, XGBoost, and LightGBM) are evaluated using nested cross-validation. XGBoost achieves the best overall performance across all three tracks (Baseline, PCA-enhanced, and FA-enhanced). Among them, PCA-enhanced XGBoost attains the highest predictive accuracy (R² = 0.868; RMSE = 1084.9; MAE = 746.9; MAPE = 0.358; pooled out-of-fold), while Baseline XGBoost yields the lowest MAE (731.4; R² = 0.863). To support transparent decision-making, Shapley Additive exPlanations (SHAP)-based attribution and scenario-based sensitivity analyses are conducted. Results show that project scale and process-level unit costs are dominant cost-drivers, while cloud usage, expert participation, and de-identification requirements exhibit secondary effects. The proposed framework provides an interpretable, data-driven approach to cost information management and decision support for data-intensive AI projects.