Sensors are indispensable in industrial applications for monitoring and control purposes. To reduce system costs, some physical sensors may be replaced with virtual ones that replicate the physical sensor by leveraging data from other sources. This work investigates how equation discovery techniques can be employed to derive closed-form expressions for a pressure signal based on actuation signals (of e.g. valves) and other sensor signals within a real-world hydraulic circuit. The discovered equations are utilized by means of: 1) physics-informing, 2) synthetic data generation, and 3) feature engineering. Depending on the nature of the discovered equation, both physics-informing and synthetic data generation proved to allow for increasing the accuracy of the virtual sensor over purely data-driven approaches.