Energy industry entering into large data age

With the continuous growth and rapid expansion of data, large data era has come, oil, electricity and other energy sectors have opened up large data development and application of the prologue. How to obtain the information efficiently from the http://www.aliyun.com/zixun/aggregation/13584.html "> Mass data, to deeply process and finally obtain the useful data is the aim of the energy enterprise to get involved in large data."

For the oil industry, many companies are applying more new technologies to strategic decision making, science and technology research and development, production and operation and safety and environmental protection and other fields in order to excavate more wealth and value from large data resources. Large data application is an inevitable trend of deep integration of information technology and it and business in petroleum industry, and the prospect of application in petroleum and petrochemical industry will be more and more broad in China. With the gradual decrease of petroleum reserves, the exploration and Development of petroleum and petrochemical industry is becoming more and more difficult, and the maturity of informatization has become the primary factor affecting the growth range of the industry. Accurate and rapid Geological Survey has become one of the core competencies that the world's energy giants rely on, and the application of High-performance computing and large data technology is a key factor.

High-performance computing for massive data processing in oil and gas exploration

Geophysical methods are the most commonly used in petroleum exploration at present. Geophysical method is a method of geologic exploration using modern physical methods, including electric method, magnetic method, gravity method, radioactivity method, seismic wave method, etc., in which seismic wave method is the most important. In order to understand and simulate the geological structure of thousands of meters underground, seismic wave reflection to collect massive data, the general two-dimensional data can reach 1~2TB, three-dimensional data can be as high as hundreds of TB or even PB class, and then a large number of intensive calculation and simulation, the results will be converted into visual visual images, It is convenient for experts to explain the data and provide reference for oil and gas drilling positioning. Therefore, the processing of these massive data can only achieve the best exploration benefit with the help of high performance calculation, which is also the main reason of the high performance computing demand in the field of petroleum exploration.

Because of the particularity and complexity of petroleum exploration industry, petroleum exploration has put forward a very demanding requirement for high performance computing. In the past ten years, the calculation and processing of petroleum exploration have used mainframe or high-performance computer, but the High-performance computer system has already faced many problems in computing performance, system construction and operating cost. The problem that makes oil exploration enterprises have a headache is mainly concentrated in three difficulties: first, the computing power demand and CPU processor performance drop more and more, at present through increasing CPU processor's working frequency to improve the computing performance of the technical route has been gradually towards its limit; Second, the rapid growth of oil exploration data and storage capacity is increasingly mismatched; thirdly, the energy consumption is more and more serious, the large size of high performance computer, the power consumption and so on, have become a great challenge to the data processing of oil exploration.

Large data characteristics of seismic data

BGP (China Petroleum Group Oriental Geophysical Corporation) is a wholly-owned geophysical and technical Service Company of China Petroleum and Natural Gas Corporation, mainly engaged in land, shallow water seismic exploration acquisition, processing, interpretation and geophysical equipment and software research and development, business distribution in the world 34 countries, land seismic exploration market share ranked first in the world. Now has 26,000 employees, more than 3,000 IT personnel and more than 300 software developers, in the Global 23 processing centers, about 90,000 CPU core and 800,000 GPU core, computing capacity of about 2PFlops, storage capacity of more than 25PB.

Niland, chief engineer of the Institute of Oriental Geophysics, China Petroleum Group, has made a keynote speech on the application of large data technology in the field of petroleum exploration in the 2013 of Huawei's cloud computing conference, and elaborated the process of collecting and processing the massive data of the earthquake.

The large data produced by the oil exploration process has its own unique "4V" feature:

1, data mass: to BGP for example, every day produces more than 7TB production data, in the seismic data processing will also produce a large number of intermediate process data.

2, the data source is single: the seismic data is stimulated by the artificial seismic wave, receives and collects by the Fixed-point collection instrument, the data source and the data format are comparatively unitary.

3, the calculation of a large amount: to BGP, for example, 54TB of raw data through the calculation of 4,000 CPU cluster processing, it takes more than 50 days.

4, processing process complex: Seismic data processing involves frequent IO and database, Operation complex.

The rapid growth of seismic data has put forward a great demand for storage and a new challenge to the traditional HPC software and hardware architecture.

Huawei Petroleum Exploration HPC solution achieves high computing and large storage capability

In response to the characteristics and needs of the oil industry, Huawei has proposed a HPC solution for petroleum exploration, which includes the following parts: