1. BACKGROUND AND PROGRESS REQUIRED BEYOND THE STATE OF THE ART
European power networks have been developed in the 1960s and 1970s and some components are now approaching the end of their expected lifetime. Utilities thus have to anticipate investment waves to come. At the same time, it is expected that the networks will undergo changes to evolve towards the “Smart Grids” placing networks under greater stress, resulting for example from the increased variability of energy flow due to renewable energy integration or new loads such as electric vehicles that may increase peak demand. The management of existing aging networks together with the new developments in progress clearly raise key issues in asset management for utilities.
In particular, transmission grid owners have to assure an appropriate payback to the capital invested in the grid asset by their shareholders. In order to achieve this goal, TSOs are trying to reduce their costs without compromising the grid availability level, as they are liable for grid outages. To this respect, asset management, in addition to becoming a key issue, is regarded as a core-competence of the asset owner.
Asset Management is based on understanding and managing the lifecycle of each of the transmission network assets in order to maintain acceptable levels of risk and performance at the lowest possible long run cost. One way of reducing costs and improving reliability is to review and optimize the current maintenance strategy programs within the transmission company. Another issue consists of investment savings through maximizing lifetime of assets and critical power components. The lifecycle of each asset is simply illustrated in the figure below.
It highlights that as assets move through their lifecycle the relative effectiveness of maintenance on the asset changes. While at the start of an asset useful life, asset management focuses on routine maintenance, this focus transitions to refurbishment and replacement as the asset condition deteriorates. Notably, the condition of asset changes over time depending on a wide range of factors, including electrical stress under normal operation and under transients or faults, environment, maintenance history, design and the skills available to maintain those assets (not simply age).
For all these reasons optimization of maintenance management is a topic that transmission companies have to urgently address in order to improve their competitiveness.
In this view the utilities try to evolve their asset management policies from the classical time-based maintenance to a condition-based maintenance (CBM) in order to reduce maintenance costs and increase the reliability of the network. Correspondingly, corrective maintenance (detection of component failure followed by component replacement) moves towards predictive maintenance programs.
According to condition-based maintenance, a maintenance action (e.g. an overhaul) depends on the component status, i.e. the maintenance activity is accomplished only when the component shows (through the monitoring of indicators of degradation), a decay in its status or performance that in a short time would compromise its functionality.
CBM can be further refined to RCM, Reliability centred maintenance or Standardization of component health monitoring , which is using risk based analysis (RBA) to give the priority to maintenance and replacement activities. In fact, even though a wide implementation of Condition Based Maintenance by European transmission asset owners, combined with centralized decision support systems, is the realistic target of the next few years, the real long-term challenge would be to develop and deploy a unified Risk-Based Maintenance of transmission assets. In this case, Risk-Based Analysis (RBA) must be combined with condition-based maintenance to give the priority to maintenance and replacement activities. Today, maintenance and renewal policies are mostly component oriented, often time- or condition-based, rather than risk-based.
Dynamic management of outage planning and maintenance schedules should also be further developed.
B. Tools supporting condition-based maintenance
While the CBM methodology is well consolidated, the tools supporting its practical applications in HV transmission grid field are not as common.
A HV grid poses a number of demanding challenges for the adoption of CBM systems, such as:
- deployment of on-line condition monitoring systems to keep track of the component status;
- coping with an asset that is geographically widespread and consists of hundreds of thousands differentiated components/devices;
- tight integration with the company legacy systems containing a detailed description of the grid asset (stations, bays, breakers, transformers, lines, etc.);
- implementation of decision support tools suggesting the maintenance actions when they are really needed. To this aim the tools have to take into account all the relevant present and past data of each component.
The biggest hurdle we face today regarding substations is that utilities cannot schedule outages because of the load and demand on their substations. In the past, systems were built to take on extra load and utilities planned on having sufficient extra margins to allow other parts of the system to take over. Now, in order to decrease costs, systems are managed with lower margins, so that this method cannot be applied any more. On-line asset monitoring can be used to point out abnormal stresses and ageing of a component, before a failure occurs, and to obtain data essential for the planning of suitable maintenance (or replacement) measures. This operation can be carried out in a period when the unavailability of that particular part of the system does not impair the overall system operation, i.e. when other redundant part can guarantee the operation. Under these conditions the system achieves an interruption-free diagnostic.
The vision of interruption-free diagnostics of a complete network with substations, lines and cables requires an holistic approach identifying the critical components and possible supervision methods including both existing and new ways of monitoring at system level . Today, data about the condition of different equipment are spread among different persons, organizations and measurement equipment. This includes everything from paper reports of manual inspection to digitally stored disturbance recordings. The main challenge is to convert raw data islands to useful information for Asset Management and Operation. This includes an efficient communication infrastructure and data management, where for example high speed LAN/WAN according to IEC 61850 can be utilized. The major task is however still to analyze data and present useful information for different users.
The overall availability of the transmission system does not rely uniformly on each individual link: some of them play a more important role and deserve a higher attention. Monitoring at component level is therefore important to identify the critical links and assure that their conditions are monitored either with continuous monitoring or manual inspection. Phase-shifting Transformers (PST) and other equipment are monitored for temperature, HVDC Circuit Breaker (DCCB) ) and gas-insulated substation (GIS) for SF6 density, etc. But there are still no overall diagnostics, which can accurately predict if and when a more detailed test and maintenance should be done. In fact, each component has its typical failure modes and related diagnostics parameters and it is difficult to correlate diagnostics data of different components. In addition, although a lot of data exist today, there are still major parts of a substation with connected lines and cables, which are not being monitored or diagnosed. This includes insulation, operating mechanism, conductor joints, disconnectors as well as influence of pollution and influence of weather. However, some possible methods do exist which can be used for on-line monitoring and on-line diagnostic system for a part of this equipment. The leading parameters to be monitored and/or tested as well different methods for monitoring assets conditions are summarized in the figure below.
Partial Discharge (PD) monitoring is one of the monitoring actions, which is now available to monitor the insulation of generators, transformers, cables and GIS without taking the equipment out of service. The advantage of this type of indicator is that a degeneration of the insulation can be detected well in advance and save the equipment from a complete and costly failure. The main challenge is to have efficient on-line monitoring and diagnostics, which can establish when the measured PD is dangerous and requires an action.
Circuit Breakers are essential for the system availability. The reliability of circuit breakers has improved during the years. International statistics show that modern SF6 breakers with spring operating mechanisms have a low failure rate and a low maintenance requirement. It is also possible to monitor the breaker contacts by monitoring the number of operations and calculating I2t as an estimation of contact wear. Besides, for reactor breakers, wear is normally not a problem. According to statistics, the operating mechanism is the major source of a failure. Traditionally, contact travel time or dynamic resistance have been used as diagnostic indicators during breaker testing. It is however difficult to use these methods for on-line monitoring. Vibration monitoring is a new method, which can be used for monitoring (continuous or intervals) during normal open-close operation.
Both PD monitoring and vibration monitoring give high frequency data, which have to be compared with an original signature to indicate any trend. This requires a database with old data as well as a user-friendly tool to analyse the result.
Modern protection and control systems are to great extent self-monitored. This includes electronics and communication, as well as connection to current and voltage transformers and trip circuits. These so called IEDs (intelligent electronic devices) also include disturbance recorder, event recorder and fault recorders, which can be used to analyze faults but also detect incorrect functions before they result in a fault. Modern IEDs designed for IEC 61850 protocol can provide new flexibility, redundancy, functionality and communication. The standard separates hardware from functions and allows new architecture with duplication of all functions, interchange of data for more intelligent automation, etc. More and more utilities will also install high speed Internet or WAN in the station which allow rapid access of all data .
In order to achieve a real condition-based maintenance, there is also the need for the implementation of asset management tools using cost-benefit analysis in the decision making process. The OPEX savings coming from the optimization of the maintenance process as well as the capitalization of the distinctive company knowledge about asset management and maintenance are the major expected benefits.
Therefore, for the critical components or aggregates of components selected to be monitored, also the set of tasks implementing the maintenance process must be carefully modelled. Then a set of technical/economic criteria must be developed and implemented, in order to support a common practice for each task. These decision logics need quite a number of degradation and failure data, which usually rely on the long experience achieved by the TSO.
Therefore, the few existing decision making tools are at the moment peculiar for each single TSO, for different reasons:
there is no common condition approach;
modelling often depends on private data;
decision support systems are integrated with the ERP (Enterprise Resources Planning) and IS (Information System) already available (which are peculiar for each organization); and
standards for communication and data management are not unified.
As for a common condition approach, the health index methodology is a powerful tool to represent the overall health of equipment and provide asset managers with a technical point of view and rank the health condition of links from ‘very good’ to ‘very bad’. This methodology can only be used if data are available.
Failure statistics should be addressed in such a way that they can be optimized for health index considerations. At the same time, the health index methodology must be developed to take into account data quality: basic data must be collected to help asset managers define asset management policies at national and/or regional area, while more detailed data are necessary to support decisions at a local area (to guide equipment renewal).
Decision-making tools need failure prediction models (including the modelling of the influence of pollution and of the influence of weather) which can suggest the maintenance actions.
So far, asset management relies on an average life-span of equipment as a function of a few critical working parameters (e.g., working temperature, number of switching, etc.). A challenge is to develop lifetime prediction models based on extended parameters that can be easily monitored (based on a trade-off between the extra cost for monitoring and the expected lifetime expansion). The other challenge is to account for the reliability of the monitoring system itself.
Weather condition forecasting and observation are also key to plan maintenance actions in an effective manner, in particular where environmental conditions (e.g. salt from the sea or snow and ice) have a great influence on grid component status.
Only a strong collaboration between TSOs would allow overcoming the mentioned difficulties and challenges, which are mainly related to the lack of common practices for condition-based and risk-centred maintenance. Agreement should be achieved on common tests to be carried out (for example on cables) under defined conditions, on common condition approach (e.g. health indexing), on information exchange of degradation/failure data and models, on use of common standards for communication and data management (for example high speed LAN/WAN according to IEC 61850).
TSOs have started to adopt an asset management process including business values, key performance indicators and risk management. In spite of great similarity on a high level there is diversity in the details (e.g. KPI used). Risk management is strongly developing but there is no common practice yet for risk assessment, prioritization and reduction. Studies on networks asset management are carried out separately by each utility/TSO today. There is no common health index / condition approach. Standardization of component health monitoring and interoperability of health monitoring systems are expected to be developed in the near future.
There is a lack of up to date and accessible asset data and condition and health information. Databases are being developed, but the type and number of indicators are diverse.
Therefore it would be useful to gather knowledge and skills of TSO experts to benefit from coordinated training, sharing of resources and efforts, data, tools and methods, so as to enable findings of relevant ideas and best practices.
This goal has been preliminarily shared by a European coordination project related to the asset management of distribution and transmission networks, called SmartLife, where 26 European partners representing TSOs, DSOs, R&D institutes and universities of nine countries were brought together . The priority objective of SmartLife was to optimize the management of both current aging and future assets by considering the ratio of network performance to renewal cost. For this, maintenance, renewal and refurbishment of components must be optimized, and general methods, policies and strategies need to be examined. The SmartLife initiative was investigating both these tasks.
After a thorough analysis of failure and aging mechanisms, maintenance and diagnostic techniques and current practices on asset management, the project identified the needs in order to develop asset management approaches that can maximize the lifetime of critical power components of existing and future networks. They concluded that, in the field of transmission, coping strategies must be set up to face realistic reinvestment waves of equipment on a global scale. The development of enabling methodologies was proposed to meet future challenges: health and risk indexing techniques, dynamic loading schemes and methods and total cost of ownership.
They concluded that management of current aging networks and the evolution towards future “smart grids” call for the optimization of network asset management and rely on a solid knowledge of aging mechanisms, a useful characterization of equipment performance and relevant asset management processes. This can only be done via widespread collaborative exchanges among utilities and R&D entities.
2. OUTCOMES PROVIDED BY THE PROJECTS THAT ADDRESS THE CHALLENGES OF THE CLUSTER
A. The MBI project
A comprehensive maintenance management tool supporting the whole HV transmission grid was developed in an Italian project, called MBI (Monitoring and Business Intelligence), implemented by the national TSO (Terna SpA). By using such Asset management Decision Support System (DSS), a new concept of on-line diagnostic system, called EDS (Expert Diagnostic System) was developed, with the goal of better handling maintenance intervention.
The system is based on the utilization of simple sensors with the aim of acquiring the most significant parameters from various substation equipment. In order to automatically schedule the maintenance activities, the system transmits signals trends and logic signals to a central computer. In addition, any further control with dedicated instruments could be programmed depending on the information received. The on-line diagnostic system has successfully been installed and tested since 2007 and is now in service on 450 Air Insulated and Gas Insulated Substations as well as on 60.000 km of HV OverHead Lines. Therefore, it could be demonstrated that the system allows the monitoring of a variety of substation equipment (with reduced impact on the existing circuits) in order to have a condition-based maintenance, hence reducing outage time and related cost as well as increasing the apparatus lifespan.
B. The GARPUR project
The aim of the GARPUR project is to design and evaluate new power system reliability criteria to be used within the key activities of TSOs at different time scales: system development, asset management and power system operation. If successful, these criteria could be progressively implemented at the pan-European level, optimally balancing reliability and costs. Indeed, the increasing uncertainty caused by (among others) the massive renewable energy integration calls for the use of probabilistic reliability criteria to supplement and enhance the pure preventive N-1 criterion.