In the current work, the productivity of single and twofold collar in decreasing scour rate and profundity around round and hollow extension wharfs was concentrated tentatively. To diminish the strength of wake vortices, various quantities of bars were likewise introduced at the downstream essence of the wharf in mix with a solitary collar. The best outcome was accomplished for twofold collar when the collars were introduced at the streambed level and one wharf measurement beneath bed level. With this game plan the lower collar was not sabotaged following 80 hours of trial, appearing around 56 % decrease in scouring contrasted and an unprotected dock. It was likewise shown that by utilizing collars, the pace of scouring diminished impressively. However establishment of bars didn't fundamentally diminish the greatest profundity of scour, it delayed the start of scouring at the upstream essence of the wharf.

The paper proposes an integrated methodological approach to modeling and forecasting Oracle DBMS performance degradation under high-load industrial environments. Against the backdrop of digital transformation and the consolidation of the AIOps paradigm, there is a shift from reactive monitoring toward proactive intelligent support mechanisms focused on early detection of adverse indicators and incident prevention. A central role is assigned to the integration of multidimensional telemetry: wait events, performance metrics of the input/output (I/O) subsystem, temporal profiles of memory consumption in the SGA and PGA areas, as well as behavioral characteristics of key background processes, including LGWR, DBWR, and ARCH, are taken into account. The proposed methodology for the first time combines classical Oracle diagnostics (AWR, ASH, V$ views) with machine learning and AIOps approaches to predict future degradations before they manifest at the user level.

The architecture of a predictive solution is described, based on a combination of machine learning methods (Random Forest, XGBoost) and deep learning models (LSTM, BERT). This composition ensures robust extraction of latent patterns in operational data streams and enables detection of anomalous states 7–12 minutes before the formation of a critical failure. Substantial attention is devoted to the reconstruction of causal relationships between infrastructural parameters and application-level indicators, which makes it possible to interpret degradation not as a set of disparate symptoms, but as a result of the interaction of resources, internal DBMS mechanisms, and the load profile. Empirical verification on experimental data demonstrates a significant improvement in operational diagnostics, including a MTTD 63%. The material is intended for practicing database operations specialists, designers of high-load solutions, and researchers developing intelligent methods for IT operations.

Creativity and inspiration, poetry and poetry, interpretation and image, plot and fable all come together at a single point - composition. The poet uses words, repeats, intensifies the image and tone, assigns a meaningful meaning to them, arranges them, organizes them, and all this unites in a single aesthetic center. So, can a composition be a masterpiece? What about modern poetry? No, each work has its own compositional quality and limit of philosophical scope. Just as there is a certain order in every phenomenon in the universe, there is also a certain compositional principle in modern poetry.

The article presents an analysis of the test results obtained by pulling out a reinforcing bar from a concrete sample. In experiments, the reinforcement diameter and concrete strength were taken as variable factors. The results were used to analyze the indicators of coupling of the reinforcement with concrete, the tension-strain state and the nature of fracture of the test specimens to establish the patterns of their changes. The summarized data are presented in tabular form.