Introduction
In the previous article we discussed the importance of building an online data management as the core decision-making platform, and gives a general framework for data generation company management and decision-making platform construction approaches. Every business wants to improve competitiveness, it must start from the subtle management, and research data value, which is a very important point.
To study the value of data, you must first do a good job of governance, management of data, because data mining, OLAP and other data analysis techniques are based on the data warehouse. In this paper, soft sail in the power industry for many years experience in data applications in the field of cooperation, it is from the perspective of the plant, to discuss the value of the number of positions at the power companies to build their ideas for reference.
Keywords: data warehousing, big data, visualization, power generation companies;
The value of a power generation company warehouse number
1, a look at the scene
the director of a Group A power plant in mid suddenly want to understand the situation of energy-saving and environmental protection in recent months, but was told the monthly environmental protection and energy required to make out more than three days. Operations to reflect that the report need both can get some data from the Provincial Environmental Protection Bureau Web site, but also the value of each of the other department heads, such as Taiwan reported up the excel table to get data from an environmental power plant word report integration, and each department consolidation report data is relatively slow, it can not be completed quickly. A power plant for this purpose headache director, Operations, Information Department is also very helpless.
2, Scene resolve the problem
root cause of the above problems is the lack of mechanisms for data management and protection. Resulting in uneven quality of data, basic data scattered, not unified, inconsistent data.
To present information system is currently the company's five power generation groups, for example, most existing information systems company owned power plants vary widely, the power plant inside the existing financial information systems, production management systems, ERP systems are relatively independent, internal information systems between the lack of a unified platform for data correlation, and integration of China Unicom, unable to measure enterprise-wide business profile from a unified perspective, it is difficult to completely release the true value of the data, showing "a lot of data, using now turn to" situation.
And a large number of positions is the role of integration and data management, and therefore the number of positions to build can be a good solution to this problem.
3, the value of the data warehouse
greatest value is the number of positions available to decision-makers a new way to observe the years of accumulated data from the macro or micro perspective, so that decision makers can quickly master the operation status of running his own business, operating costs, development trends of business development and decision-making information is important to facilitate make more timely, accurate and scientific decision-making.
Second, the concept of power generation companies and the number of warehouse reference architecture
1, the number of positions of the basic concepts of
data warehouse is a subject-oriented, integrated, relatively stable, reflecting the historical changes in data collection. It is a new data processing architecture of the various departments within the enterprise business data unified and integrated central data warehouse, to provide the necessary information for enterprise decision support systems and executive information systems.
2, the number of power generation company warehouse reference architecture
3, ODS and DDS DESCRIPTION
1) ODS: Operational Data Store (Operational Data Store) module
It is one of the primary data store to the entire number of bins. All business data include enterprises, the construction can be distributed. It is an integrated and centralized data storage, enterprise-level data from the variety of topics, including the bottom, fine-grained, need long-term preservation of data, but low query efficiency. Words, "Central ODS is shared with enterprise-class entity-relationship (ER) data model to store business data pool."
2) DDS: storing multidimensional data (Dimension DataStore) module
Is designed for fast query design is derived from the ODS in the raw data, the data is often query. DDS uses a multi-dimensional relational model sub-themes (star, snowflake, models, etc.) for storage of data to support in-depth analysis of data, provide data for OLAP.
Third, thinking of building decisions generation company several positions in
the implementation of several warehouse is a large and long-term process, to successfully implement a data warehouse project, which requires investment in human and financial resources, but also need to implement staff with experience using standardized implementation methodology .
For large-scale power plants, power generation group is indeed the need to build a number of positions. But from the perspective of power plants, consider the many reasons of cost, time and so on, small and medium sized power plant is not necessary to build a complete one-time number of positions, the best strategy is to build the corresponding data mart to meet current demand based on business topics, at the same time step by step planning and construction of complete number of positions.
Three basic principles of corporate development of the number of positions of power:
1, the development of the periodic
system development cycle data warehouse is a dynamic feedback heuristic. The number of warehouse development and application cycle can be divided into three phases: planning and analysis data warehousing, data warehouse design and implementation, using a data warehouse. Because, in general, the number of possible positions in a complete cycle. Therefore, these three stages continuous cycle, improve, enhance, periodic spiral development, forming a loop, is constantly reciprocating.
2、数据驱动
数据仓库的开发是从数据出发的，从存在于业务处理系统环境中的数据出发进行构建，要尽可能利用已有数据、代码等，而不是全部从头开始做。这就要在进行数仓设计前，首先识别原有的数据库系统中已经有什么数据。
3、联合使用自顶向下和自底向上的策略
首先自下而上从多个业务角度构建相应的数据集市，将各个业务模块内部的数据整合，初步梳理出相应的问题，在解决部分问题时要兼顾考虑全局，跨业务模块的数据整合。
数据仓库的开发策略有三种：自顶向下、自底向上、俩种联合使用。
1、 自定向下策略：
在实际应用中较为困难。因为需要在一开始企业决策层和管理人员完全知道数据仓库使用的预定目标，并明确数据仓库要在哪些决策中发挥作用。而在数据仓库的开发初期往往不能明确了解数据仓库用户的使用需求，容易导致数仓失去其应有的价值。
2、 自下而上策略：
针对特定的管理决策问题进行开发，适合在数据仓库的应用目标并不是很明确以及数据仓库对决策过程影响不是很明确时使用。其能以较小的投入获得较高的数据仓库应用效益，容易取得成效。
3、 俩种策略的联合使用
则能满足既能够快速的完成数仓的开发应用，同时仍可以建立具有长远价值的数据仓库方案。
遵从上述的原则，再结合给出的参考架构，依据维度建模的理念进行具体的开发设计，相信电力企业推进数仓的开发工作将会更加的顺利。
结语
作为电力的基础组成单位，发电厂是整个发电集团基础数据库最大的数据源，其数据量大、数据种类多、分布专业范围广、对其规范化分类没有现行的行业标准，因此凌乱无序的数据会形成数据垃圾。只有利用大数据工具对数据进行有效的采集、衔接、规范、分类和分析，才能为电厂和集团公司管理者提供有效数据，为电厂从“信息化电厂”到“数字化电厂”最终向“智能化电厂”转变提供有效的基础。
构建面向电厂的数据仓库能有效的协助电厂管理者们从庞大的数据环境中解脱出来，并提供给管理者们有关运行、生产、管理的智能化辅助决策服务，更高效的发挥数据的价值。关于发电企业数据治理的工作，只是电力企业在数据化过程中面临的问题之一。帆软，在电力行业深耕多年，已经积累发电、输配电领域合作客户80多家，基于丰富的合作经验以及客户基础，我们将于2018年11月29日至12月1日举行电力行业数据化研讨峰会，届时将有发电电厂、光伏企业等为大家分享电力企业数据化建设经验，并就智能电网、数字电厂等行业关注的热门话题进行深入研讨，对大会议题感兴趣的电力行业同仁们，欢迎扫描下方二维码，了解会议详情。
