One may argue that, provided that certain conditions are met, leak detection systems on pipelines aren’t required. Leak detection systems on pipelines have become a common element in pipeline management systems used by pipeline operators though. Different reasons may have lead to the decision of a pipeline operator to implement or modernize a leak detection system. Based on initial results from a survey by CONCAWE it seems that the number of 3rd party’s illegal interventions is growing, which is one possible but essential reason for an operator to implement a leak detection system. Once the hurdle of making the decision to consider implementing or modernizing a leak detection system is taken, the challenge of selecting the best and feasible method, making the final investment decision and implementing the system starts. Presentation describes a real-life case of the whole journey the pipeline operator MOL Logistics had to make for the implementation of a proper functioning pipeline leak detection system in a brownfield environment. Topics included are:

• The process and reasons to decide on implementing a leak detection system.
• Environmental regulations
• License to operate
• General safety and operating requirements
• How to determine the leak detection requirements.
• The process of selecting leak detection method(s) and supplier(s).
• How to embed the leak detection system in the overall pipeline management system, both functionally and technically.
• How to proof that the leak detection system actually functions according to the specifications.
• Brief overview about the technical possibilities of leak detection systems
• What are important elements to take into account for the proper functioning of a leak detection system.
• How to know whether critical elements perform as declared by suppliers, also in cooperation with other elements in order to meet the overall leak detection requirements. An example includes the requirement for accurate measurement data by leak detection algorithms and the consequences if the measurement data are of a lower quality.