Consortium on Insulation Life in Emerging Environments (COIL)
Designing and operating electrical insulation systems for the next 50 years
Purpose
Decades of experience have led to robust methods of predicting insulation life under 50/60 Hz ac and the transients that occur in those systems. However, dc voltages are making significant inroads in both the utility grid and in electrified transportation, which is introducing different failure mechanisms, making it challenging to design and specify new dc systems.
The goal of this research consortium is to permit industry and government agencies to jointly fund research that provides critical information needed to predict insulation life in dc systems. Using the knowledge developed, participants can develop their proprietary materials, insulation systems, and components based on these materials. The proprietary development will be outside of the consortium.
Technical
Failure in electrical systems is typically caused by a combination of thermal, electrical, mechanical, and
chemical factors and often presents as electrical insulation failure. Under ac, the incipient failure is frequently indicated by the partial discharges and the characteristics of these discharges provides life estimates that have been vetted by decades of experience. This situation is shown graphically in this figure. The life (L) is the time (t) that the insulation can be expected to withstand the design electric field (E). For organic insulation materials, if partial discharges (PD) are not eliminated by design, the system life is very short. If a combination of, for example, thermal and electrical forces, stimulate the initiation of partial discharges during operation, again life is shortened.
This is insight gained from ac experience that is expected to be valid under dc, but the lines are likely to have different slopes under dc than ac. There are two important factors shaping the need for better data under dc. One factor is that, globally, the voltage levels at which designers would like to operate dc systems are increasing well beyond the limited performance data that exists. Increasing voltage while maintaining reliability requires more engineering data. The second factor is the role of power electronics. Power electronics are offering higher switching speed, higher voltage operation, and higher temperature operation than were previously possible. This can induce accelerated ageing across the entire voltage range, likely going down to a few hundred volts, thus becoming a concern in electric cars. This effect can reduce life in cables, motors, generators, transformers and even the power electronics packages and circuits themselves. The pictures show a new transformer and a transformer after being in service in a solar facility for five months, which is one of several transformers that failed in service.
The issue is the pulsed waveforms that can be produced with very rapid rise and fall times. The difference between the partial discharge activity with inverter waveforms and a 60 Hz sinusoidal system are shown in this figure. The figure shows the partial discharge patterns when the voltage is produced by a two-level inverter, a five-level inverter and a sine wave. It is clear that the inverter waveform is producing quite different patterns. This research will focus on the linkage of the various behaviors to insulation life.
Organization
The University of Texas at Austin will manage the consortium. The University researchers leading the activity will be Dr. Gian Carlo Montanari, who will lead the technical activities, and Dr. Robert Hebner, who will provide programmatic oversight. Both Dr. Montanari and Dr. Hebner have decades of experience in insulation research. Dr. Montanari has focused on partial discharge behavior while Dr. Hebner has contributed to the understanding of discharge dynamics.
The research will be funded by membership fees. The annual fee is $50,000 per participant. Companies that want to provide greater support and exert greater influence on the research direction can purchase two memberships. Companies that join after the first year of operation may also be required to pay a one-time entrance fee set by the participants to compensate for start-up costs.
All of the research will be shared by the participants and is expected to eventually be published. All participants will receive a non-exclusive royalty free worldwide license for all intellectual property developed. The University maintains the right to continue practicing and expanding upon any consortium intellectual property.
The research priorities are proposed by the researchers and selected by a vote of the participants. For voting, the University has two votes. It is anticipated the prioritization will be a result consensus and compromise.
The participants will have, via semiannual meetings and intermediate reports, first access to the information produced. Moreover, they will guide the research direction. In that respect, the semiannual meetings are expected to provide continuous insight into the state of the art in this field. As students will be involved in the research, the participants will have an opportunity to meet students who have important skills for their companies. So, the consortium may also serve as a useful recruiting tool.
Although the research results are shared among the participants in the consortium, the University has found in other such groups that there is occasionally a benefit for a participant to negotiate an additional separate project with the university for which the results would not be shared with the consortium. This is acceptable.
Motivation:
The COIL technical program was developed to help smooth the transition to the emerging electrical environment with less cost and disruption that other changes have produced. This opportunity arises from the growth of dc and hybrid ac and dc systems with the attendant power electronics to control these systems.
This emerging electrical environment is driven in large part by the integration of renewable energy sources in the legacy power grid, including HVDC transmission links, and by the electrification of transportation on land, sea, and in the air.
Emerging Technology:
The driver for the new electrical environment is the switching pulse train produced by power electronics. The growing adoption of wide band gap semiconductors promotes higher voltages, faster switching speeds, and higher temperatures, all of which may reduce insulation life.
Affected Components:
There is a risk to all power system components including motors, generators, cables, splices, terminations, transformers, power electronics packaging, and inverters. Designs and materials that work well under ac power are exhibiting challenges in the new environment. Moreover, the concern is not limited to a single voltage class – life limiting issues can occur in low, medium, and high voltage systems.
Key Technical Activities:
The technical program is intended to address the following important topics. It should be recognized that these are not independent but are interrelated efforts to achieve the common goal of the prediction and control of dielectric behavior in the emerging electrical environment.
• Insulation characterization and design and life modeling
New environments need techniques for insulation characterization and tools to derive the desig stresses. Life models must be devised which can predict life, and thus insulation systems design parameters, under waveforms different from ac (e.g., dc, very low frequency ac, and repetitive impulses, harmonics). All must be included in a stochastic framework, because insulation design is fundamentally probabilistic.
• Degradation mechanisms and diagnostic property investigation
The basis for life modelling and for the identification of diagnostic properties (which is required for condition monitoring and life prediction approaches) is to investigate the aging mechanisms under various types of voltage waveforms, load cycling, and mechanical stress. Note that dianostic property does not mean only partial discharge, but also space charge, dissipation factor, dissolved gas, and several others.
• Measurement technology
The pulsed nature of the signals makes the separation of partial discharges from the electrical noise in the system significantly more challenging than under conventional ac signals. Analog difficulties, for opposite reasons, hold for dc, where the partial discharge repetition rate can be so low that the discharges are masked by noise. While this research team has made progress in this area, further improvement is needed so that these measurements can become routine. It is antic pated that these improvements will be in both hardware and software. Other properties need ad quate testing tools, such as space charge (which, in some cases, may prove to be the leading property associated to premature failure in a dc system).
• Material screening and selection, and developments of new materials
Different materials will have different performance characteristics in the emerging electric envronment. Test techniques and models will be developed to inform material design and selection. It might well be that existing materials will not be able to satisfy the design needs put in place by the new voltage and current environment. Consequently, properties of new materials will be determined, with a focus on nanostructured materials. This research team has multi-year experience in the development and characterization of nanostructured insulating materials.
• Adopting core measurement technology to specific applications
As examples, motors, cables, and power electronics packages will need different measurement approaches due to their different input/output, material combinations and internal structures. This research will identify challenges and solutions to important classes of measurement applications.
• Health monitoring
Condition monitoring is critical for system owners and operators. As an example, partial discharges provide precursors to failure, but there is insufficient experience and engineering data related to the emerging environment to have confidence as to how much of the previous knowledge is relevant. Also, algorithms encompassing a global set of diagnostic properties for the health condition and residual life evaluation have to be developed. The research team has experience in health monitoring via partial discharge (and other property) measurements that is expected to be important in this investigation.
• Standards and specifications
Some of the information developed will be important for future standards and industrial specifications. The industrial partners will lead in the identification and promulgation of any new standards. The university research team can provide technical support if needed.
University Consortium Role:
Buyers and sellers of materials, test equipment, utilities, energy generation (particularly renewable) companies, components and systems manufacturers must have a common set of engineering data on which to base judgements where improved life and performance can be achieved. Based on that data, manufacturers innovate new products and services needed by the community, and the community has the confidence needed to buy and use the technology. This information is only valuable if trusted and widely available, making it unwise for a single company to bear the cost of developing the information. The consortium makes it possible for each member to exploit the comprehensive information developed, while sharing the development
cost broadly.
Setting Priorities:
The priorities for phasing of research projects will be set by the principal investigators in consultation with the industrial advisors. This process is expected to produce an efficient research program that provides industrially relevant information to the participants as quickly as possible. While it is expected that most or all of the research results will be published, publication delays mean that the participants will likely have a year or two advantage over their competitors who are not participants. Moreover, with the industrial input on priorities, the members will know students who are experienced in topics important to their companies prior to the student’s availability for hire.