China National Petroleum Corporation
Operations Research Models Enable Better Planning of Natural Gas Pipelines Resulting in Reduced Costs, Lower Budgets, and Enhanced Revenues
From 2010 to 2015, China’s annual natural gas consumption increased from 107 billion to 193 billion cubic meters, while the total length of the pipeline network grew from 32,800 kilometers to 45,000 kilometers. With the demand for natural gas expected to continue increasing rapidly, and in view of the large scale, China recognized a need to be efficient in planning, constructing, and operating its pipeline network.
China National Petroleum Corporation (CNPC) is China’s largest oil and natural gas producer and supplier and one of the world’s top five players in the energy industry. CNPC now holds over 72% of China’s natural gas resources plus approximately 70% of national pipelines. Since its establishment in 1988, CNPC has integrated a series of upstream and downstream oil and gas businesses.
Prior to 2015, all of CNPC’s annual planning for gas production, importing, transmission, storage, and sales was manually conducted using spreadsheets. Increasing complexity caused this to became no longer feasible.
To pursue efficiency, and conquer complexity, CNPC turned to operations research. Researchers from CNPC, the University of California, Berkeley and Tsinghua University (Beijing) partnered to develop and implement novel analytical models. CNPC’s approach changed from making simple production-pushing or demand-pulling decisions to profit maximization. Optimization models were developed to deal with the complex physical laws of natural gas pipeline transmission dynamics and to provide a unified way to handle different models. The research team worked to produce models that complete in seconds, generate solutions with satisfactory accuracy, are robust, and are user friendly.
The operations research team modeled the dynamic planning of natural gas transmission and usage as a nonconvex, deterministic optimization problem. They invented an innovative solution method to tackle both finding optimality and being sufficiently accurate. Specifically, they modeled the application as a nonconvex and nonlinear mixed-integer problem, then convexly relaxed and solved. They proposed a three-stage framework to improve solution accuracy. And they developed implementing software.
Benefits of the operations research effort have been substantial. Implementation of the software saved more than $530 million for CNPC’s natural gas transmission in post-implementation years 2015 and 2016. In addition, efficiencies discovered within the existing pipeline network postponed the need for new pipelines, resulting in a budget reduction of over $20 billion. And CNPC realized greater revenue from selling more natural gas than would have been possible without the efficiencies. Furthermore the importance of optimization has been recognized by the entire corporation, while the modeling methods and the optimization software system have become core techniques for reducing operational costs at CNPC.
This project also favorably impacts China’s sustainable development. The new analytical software provides accurate evaluation of CNPC’s current transmission capacity, helps CNPC make well-informed new investment decisions for its transmission system, and thereby facilities more efficient operation of the gas network. That increases the percentage of natural gas in China’s energy consumption, and thus has a positive impact on the sustainable development of China’s economy.