Development of a Verification Process for Truck Pneumatic Systems

Sammanfattning

This thesis investigates the development of a verification process for pneumatic systemsin heavy-duty trucks, with a focus on air-leakage detection at the End-of-Linestage. The study was conducted at Volvo Group Trucks Technology & IndustrialDivision in Tuve, Sweden, and evaluates the feasibility of integrating digital pressuredecay measurement equipment to improve quality assurance, reduce operatordependency, and enable data-driven decision-making. A combined methodology wasapplied, including a literature review, stakeholder interviews, benchmarking acrossproduction sites, and experimental testing utilizing PDCA methodology. Key processparameters were assessed through statistical hypothesis testing, including equivalencetesting and paired one- and two-tailed t-tests. These were utilized to evaluatethe effects of test duration, stabilization time, filling method, and subsystem activationon leakage detection accuracy. The results demonstrate that pressure decaytesting can be significantly optimized. Equivalent detection accuracy was achievedwith reduced test durations compared to the current 4-minute reference, enablingcycle time reductions of up to 2–3 minutes under controlled conditions. However,the inclusion of a stabilization phase was shown to be critical for reliable measurementsdue to transient pressure variations caused by thermodynamic effects. Therequired stabilization time was also found to vary across trucks, supporting theimplementation of a dynamic threshold-based approach to balance measurement reliabilityand efficiency. Furthermore, subsystem activation was shown to influencemeasured pressure decay, indicating the need to adapt leakage thresholds dependingon test conditions. Based on these findings, a standardized verification process withautomated OK/NOK decision-making is proposed, integrating digital measurementequipment into the existing production environment. The process improves consistency,enables data collection, and reduces operator dependency. While the conceptis technically feasible, further validation is required before full-scale implementation.