UNCO

Dependable Instrumentation and Control for Renewable Energy Management Systems using Dedicated and Public Communication Channels

Reference: F1R-ING-PUL-08UNCO

Project Leader: Jean-Régis Hadji-Minaglou

Research team: Ralf Hoben (Doctorant)

Research domain: Engineering

Project period:

Partners: CRP Henri Tudor , Institut für Energiesysteme IZES in Saarbrücken

Abstract:

Distributed Control Systems (DCS) are used to optimise processes monitored and controlled by spatially distant components connected by a communication sub-system.
They usually implement a hierarchical structure with sub-systems identified as component layer controlling the hardware access, an interface layer, a process layer for data aggregation and information systems or gateways to enterprise resource planning, a management level with administrative and application development services.
DCS are used in a variety of technical applications, one example is electrical power generation and transmission (so-called SCADA systems: Supervisory Control and Data Acquisition).
As a technical application the DCS must meet strict quality criteria defining reliability and latency as dynamical properties, i.e. possess a (at least statistically) defined Quality of Service (QoS), at least some components need to be realised as so-called Event-Based Systems, in order to ensure short intervention times during abnormal system states (and avoid cost intensive high polling cycles).
On the other hand it is necessary to open the data gathered and the device interfaces provided to external systems working in a not strictly controlled environment using generic methods for data exchange and service provisioning.
This means it is necessary to distribute the functionality according to their time and safety criticality and map the services provided by the different levels while maintaining QoS requirements where needed.
An interesting example of such a system is distributed energy generation, especially when it consists of units providing renewable energy, as they are usually located in areas not covered by standard infrastructure as it is the case for larger-scale power sources and are currently operated on a best effort base without a controlled strategy to maximise overall dependability.
This work concentrates on the definition of a prototype distributed energy generation architecture and aims at defining the requirements in a quantitative manner so they can be expressed as input to the models determining the dynamic behaviour of the complete system consisting of hardware (i.e. electrical components and communication devices) and software (i.e. data communication and aggregation protocols and process adapted algorithms).
A promising candidate to gather reliable and timely data describing the system state are sensor networks, which received considerable research attention in the previous years, mainly because technology enables the deployment of low-cost, wireless nodes, capable of hosting data-centric routing algorithms ensuring the requested QoS.
To link these communication clusters to mainstream applications needed to provide operational data (as e.g. weather and power demand or pricing data) ad-hoc approaches as bespoke programs are still the standard. There are recent approaches to standardise the interface to sensors and actuator as e.g. IEEE1451
The goal for the overall architecture is to show ways to integrate these clusters into existing Open Source or commercial enterprise frameworks by means of enhancing or developing a generic middleware layer fulfilling requirements of a Service Oriented Architecture (SOA).
The challenge lies in splitting this architecture into centralised and de-centralised components enabling a distributed installation in response to the underlying communication networks with intrinsic properties, which cannot necessarily be hidden by an abstraction layer above, especially when it comes to engineering requirements as determinism, real-time behaviour and energy efficiency.
Thus a prototype based on an instrumented, small-scale power generation unit and identified candidate architectures interfacing a cooperation partner's Energy Management System will be implemented as a proof of concept.