There are many issues around the move to Process Bus and in particular Sampled Values. These come up in various discussions and so I have collated a series of issues herein afor convenience. Process Bus is not just Sampled Values. Process Bus can carry SV, GOOSE and MMS. Process Bus is that part of the LAN cabling that runs out to the primary plant such as CT/VT, CBs, T/F, fans, pumps, Cap Bank, SVC. This is further discussed here - scroll down to "Spot the Process Bus" on the right hand column. SV are not new - the first MUs producing SVs that I know of were in service way back in 1984 and they were associated with optical CT and optical VT. They produced 3 different outputs normal 1 A and 110V, low level analogue 0-10 mA and also a digital frame which eventually became known as IEC 60044-8 which subsequently morphed into IEC 61850-9-2. When we look at what the MU is, it is “simply” the A/D converter module that used to be housed in the relay box, now housed separately. There is nothing really new in that either. Low Impedance Bus Bar Protection schemes have been using a distributed I/O unit or Peripheral Unit since the late 1980’s - this is a remote A/D converter communicating to a central box to do the actual algorithm but using a totally proprietary comms protocol. In terms of reliability and potential failures we can therefore see that whilst the MU is another box in the system with another power supply, processor and comms board etc, the fact that it is permanently “talking”, regardless of whether there is primary current or volts, we are totally certain of the entire system operation – in fact we don’t even need any relays connected to the LAN to verify that. Reliability and Availability are key improvements as a result of using IEC 61850 as discussed here. The question of redundancy is important but needs an engineering assessment. As I mentioned above the MU is just the relay’s A/D module separated out – we never insisted on redundant back planes inside the relay – we had complete redundant systems and so should be the case with MUs. Aah I hear you say, but we are now relying on a single fibre and possibly lots of switches in between. True, but we also relied on lots of wires and links before and we never really knew if they were intact unless current was actually flowing in the wires – a LAN will keep us informed even before the CB is closed. Never the less we have to consider what happens when messages didn’t get through – well, we (probably) never tested if the relay was stable if the VT module was withdrawn whilst in service, or the module connectors or the back plane failed interrupting the signals being sent to the processor – we blindly trusted the vendors had thought of that in the security of their algorithm design. Now we can actually verify what is happening and make sure the IED has the right “graceful degradation” rules as required by IEC 618061850-3-4.3.3, and even apply our own rules and controls to be certain to block the relay if some SVs suddenly disappeared. I hesitate to talk about this issue of messages getting lost between the MU and the SV – many will say the solution to that is to use PRP or HSR LANs. Sigh! Well, that is true – they provide bumpless operation – but just because you can, doesn’t mean you should. Given we have always designed our entire protection systems as inherently unreliable, we have always provided duplicate systems because the “bump” might last hours, days or even months. We didn’t connect two paralleled wires between each CT terminal to each relay terminal in case one wire was eaten through by a rat or someone left a link open, so why do we suddenly want to have bumpless MU-IED comms? Aah again! But in lower voltage systems we don’t have duplicate systems so we rely on one system to be absolutely reliable – really? But if the relay fails itself you have accepted a single, if not the largest single point of failure but have not worried about “bumpless” – we have relied on back up protection. Refer this architecture discussion. Indeed the MU-IED relationship CAN be vendor dependant – i.e. a matched pair. Indeed this is the case with the GE Hard Fibre “Brick” solution. Whilst it produces a SV frame in IEC 61850-9-2 format, this matched pair aspect causes many an argument that the Brick is not true IEC 61850-9-2 as it is uniquely dependant on the matched relay to control the sampling and it cannot use a switched LAN - this is in principle against the prime objectives stated in IEC 61850-1 as vendor independent interoperability. However most asset owners don't appreciate being locked into a vendor because of some other matched equipment. Whilst the overall principle of IEC 61850 is not interchangeability (discussed here), it does go a long way towards minimising the effort associated with changing for a newer version or a different supplier. Given the operational life of the substation is say 100 years with a couple of primary refurbishments and say five or six secondary system refurbishments, can we be certain that the same supplier will exist continually and support all their products with the same life expectancy. Segregation of the MU and the IED with consistent engineering configuration and interoperability provides the asset owner with the freedom and security to choose the right product for the right reasons at each refurbishment/replacement decision. Now then this brings us back to the justification of “why bother”. Certainly the reduction in wires is an issue in the days of higher copper prices. However there is also the engineering costs. CIGRE 2008 Paper B5-102 indicates a copper based system cost $297K to install, a fibre based system $138K. Of course it could also be argued that we could mount the relays out in the yard to save wiring. Perhaps but firstly most relays are only rated to 55 degrees whereas MUs have special versions generally rated to 85 degrees – OK, the relay hardware could be "beefed" up (perhaps even the MU integrate the protection algorithms since that hardware is already outdoor rated) but we need to deal with practical issues of heat dissipation (probably larger boxes, vents, no fans ...) and none-the-least is that the operator interface displays are very hard to obtain with such extreme temperature capability (they generally turn black in high temperatures - or the operator can't read the screen in open sunlight). However we then look at operational issues – commissioning often involves lengthy periods on site with many test sequences over many days – it is generally more convenient efficient and acceptable to the technicians to be able to do this inside a building. In some countries they have provided multiple small bay-based building out in the yard next to the respective primary plant to minimise wiring but retain the building environment – this can lead to other difficulties when dealing with multiple bays in multiple buildings rather than on adjacent panels. We then need to turn to life time costs. With wire-based systems, if you replaced any equipment you had to redo the whole CAD drawings, review and approve, install, test and commission. Once a digital system is configured to send a message from one device to another device, that engineering never has to be repeated even if there is a different vendor involved. That saving amounts to millions of dollars in each sub over say 3 or 4 refurbishments in the life of the substation. The finally is operational benefit. Wire based systems would take hours perhaps even close to a day by the time staff got to site and did all the replacement and recommissioning. Digital systems allow us to provide “5-minute restoration remotely” when an IED fails. Refer paper and presentation 19D and 19P |