Unplanned downtime is an undesirable reality that costs the customer revenue and performance penalties on its contracts for the transport of oil and gas products.
The customer experiences two forms of unplanned downtime. The first type arises due to physical systems faults, while the other arises from malicious interference from human agents. The customer required a system that could infer what type of 'faults' has occurred in real-time, to recover and sustain availability of transport service within 1 hour.
We developed proprietary technology that characterizes the different types of 'faults', diagnosed their occurrence and provide real-time feedback to operators in the field or control center.
The end solution has been architected to perform large scale data fusion, real time characterization using fuzzy-based artificial intelligence methods and drive a robust user interface for ease of human intervention. Where the basis for characterization of the 'faults' is the features exhibited by the product flowing through the pipelines.
The complexity of this project was the application domain and the peculiar geographic location of the pipeline infrastructure.
Hardware
Communications protocols implemented
Software
The technology developed by Mizoli provides the customer with a unique tool that has no commercially available options at the price point of their budget. We continue to learn from the deployed system and are enabling the customer to unlock the potential of AI in this domain.