Efficient Provision of KPIs with High Information Content as a Complement to Conventional Data Management Systems (DCS, PI) Operating Experience with Expert Systems

摘要

Thanks to the developments in the area of 1. distributed control systems (DCS) 2. data historian capabilities as well as 3. application software for real-time data management (e.g. OSIsoft PI System) operators and supporting engineers enjoy a variety of convenient ways to access operational data of their power plant unit, their wind turbine, their bio-plant etc. Many measurements are available at a central data base and archived for later use. However, challenges for operators still remain: the operation of plants in most parts of the world in general has become more flexible whilst the operating hours are often limited. It is difficult to assess the condition or performance of the unit under constantly changing external factors (weather, fuel quality, load, market demands, etc.). At the same time, the demands of official agencies, management and customers increase: additional reports are required and not-spotting of defects is not accepted any longer. Many companies establish internal experts who are meant to support the operator via timely evaluations and practical recommendations. These experts often have to use the tools already available and typical issues they have to deal with are confusing and stored in fragmented data archives. Additional, too strong data compression (losses of important information) and lacking significance of the values and results make the evaluation tricky. Therefore, the conversion of data to valuable information is crucial. With advanced data management as well as handy calculation tools users can rely on state of the art modules such as neural networks, high precision water steam tables, fast fluid property algorithms, early warnings by means of statistical process control, self-organizing maps and fuzzy logic modules. The tools and modules of SES System Technologies cover these tasks: EBSILON~RProfessional is the software to model first principle models and conduct data reconciliation; in connection with the SR::x Interface and Data Management System EBSILON~RProfessional serves as the calculation kernel for the online performance evaluation SR::EPOS. SR::SPC is the package for predictive maintenance and early warnings, and the Finger Print system allows monitoring of large amounts of measurements. However, the underlying approaches are more or less universally applicable, therefore the System Technology tools – also known as the SR::Suite (Figure 15: Examples of software-based expert systems) – mainly serve as an example within this paper. Which approaches and tools support the current situation depends on the current task – but the paper will share experiences from recent projects in terms of benefits to the customer, tricky obstacles on the way and long term usage / effect. The paper will also address two main fields: First, relevant background information will be discussed (chapter 2). Second, the specifics as well as advantages and disadvantages of the different grades of KPI (key performance indicators) will be shown based on project KPIs of SESs recent projects (due to confidentiality requirements the actual project will be masked, chapter 3).