P/00/01 1 28/5/91 Regulation 3.2 AUSTRALIA Patents Act 1990 ORIGINAL COMPLETE SPECIFICATION STANDARD PATENT Name of Applicant: Psymetrix Limited Actual Inventor Alexander Golder Address for service is: WRAY & ASSOCIATES Level 4, The Quadrant 1 William Street Perth, WA 6000 Attorney code: WR Invention Title: Method of Monitoring a High Voltage Grid Power System The following statement is a full description of this invention, including the best method of performing it known to me: 1 2 1 "Method of Monitoring a High Voltage Grid Power 2 System" 3 4 Field of the Invention 5 6 This invention relates to electrical power 7 transmission systems and particularly to high 8 voltage, high power grid systems. 9 10 Throughout the specification, unless the context 11 requires otherwise, the word "comprise" or 12 variations such as "comprises" or "comprising", will 13 be understood to imply the inclusion of a stated 14 integer or group of integers but not the exclusion 15 of any other integer or group of integers. 16 17 Background Art 18 19 The following discussion of the background art is 20 intended to facilitate an understanding of the 21 present invention only. The discussion is not an 22 acknowledgement or admission that any of the 3 1 material referred to was part of the common general 2 knowledge in Australia at the priority date of the 3 application. 4 5 Assuming no thermal limitations, power transfer 6 limits for a power system frequently arise from 7 concerns about transient instability or voltage 8 instability in the event of a contingency. There 9 are also concerns regarding steady-state 10 instability. In order to quantify these potential 11 instabilities a knowledge of the power system's 12 dynamic characteristics is necessary. Existing 13 techniques used to provide estimates of a power 14 system's dynamic characteristics, and hence power 15 transfer limits, are based on mathematical dynamic 16 modelling studies which are subject to significant 17 uncertainties. Hitherto system engineers have had to 18 build in large factors of safety, effectively 19 discounting the safe power transfer capacity by a 20 considerable margin, and thus unduly limiting the 21 power which can be transferred or requiring excess 22 investment in capacity. 23 24 It has previously been proposed to use a "signal 25 energy" approach to the setting of power transfer 26 limits, based on the observation that "signal 27 energy" increases (and damping deteriorates) 28 asymptotically as the power flow increases. 29 30 However, these proposals suffer from the facts that: 31 (a) they rely solely on mathematical dynamic 4 1 modelling, with the attendant problems discussed 2 above; 3 (b) the use of "signal energy" without splitting 4 this quantity into frequency components obscures the 5 nature of the problem; and 6 (c) the relationship between signal energy and/or 7 damping and MW power flow is not, in practice, at 8 all uniform. 9 10 Disclosure of the Invention 11 12 In accordance with a first aspect of the present 13 invention, there is provided a method of identifying 14 the source of an oscillation in an electrical power 15 transmission system, comprising the steps of: 16 (a) deriving dynamic characteristics from 17 measurements of small perturbations in 18 operational parameters from a plurality of 19 measurement points in the system; 20 (b) determining which of the plurality of 21 measurement points that the source of the 22 oscillation is closest to by comparing the 23 derived dynamic characteristics; and 24 (c) identifying one or more possible locations 25 of the source of the oscillation based on 26 the determination of which of the plurality 27 of measurement points that the source of the 28 oscillation is closest to. 29 30 Preferably, the method comprises measuring further 31 operational parameters at a plurality of measurement 32 points in the system and wherein step (b) further 5 1 comprises including the measured operational 2 parameters in the determination of which of the 3 plurality of measurement points that the source of 4 the oscillation is closest to. 5 6 Preferably, step (b) comprises establishing fixed 7 multivariate relationships between the dynamic 8 characteristics and the operational parameters when 9 determining which of the plurality of measurement 10 points that the source of the oscillation is closest 11 to. 12 13 Preferably, establishing fixed multivariate 14 relationships comprises using multivariate 15 statistical analysis by forming a matrix of 16 observations from which can be derived the 17 parameters for the multivariate relationships. 18 19 Preferably, the dynamic characteristics comprise the 20 mode damping, mode frequency, mode amplitude and/or 21 the mode phase of the system. 22 23 Preferably, the method comprises measuring the 24 operational parameters comprising measuring one or 25 more of the following: voltage; current; real power; 26 reactive power; and system frequency. 27 28 Preferably, measuring voltage comprises measuring 29 voltage phasors, voltage phasor angle difference and 30 voltage magnitude. 31 6 1 Preferably, measuring current comprises measuring 2 current phasors, current phasor angle difference and 3 current magnitude. 4 5 Preferably, the mode amplitude is used to determine 6 which measurement point is closest to the source of 7 the oscillation. 8 9 Preferably, the mode damping is used to determine 10 which measurement point is closest to the source of 11 the oscillation. 12 13 Preferably, the mode phase is used to determine 14 which measurement point is closest to the source of 15 the oscillation. 16 17 Preferably, step (c) further comprises identifying 18 electrical plant in the system closest to the or 19 each of the possible locations. 20 21 Preferably, the method comprises taking corrective 22 action at the identified electrical plant thereby 23 mitigating the source of the oscillation. 24 25 In accordance with a second aspect of the present 26 invention, there is provided a method of monitoring 27 a high voltage electrical power system comprising 28 the steps of: 29 30 deriving measurements of the small signal dynamic 31 characteristics of the power system, wherein the 32 measurements are obtained under the prevailing 7 1 system conditions, said system conditions being 2 defined by the operational parameters of the power 3 system; and 4 5 determining the power transfer limits of the power 6 system utilising said measurement. 7 8 Grid systems are continuously subject to small 9 perturbations. 10 11 In accordance with a third aspect of the present 12 invention, there is provided a method of determining 13 power transfer limits in an electrical power 14 transmission system, comprising: 15 16 (a) measuring the dynamic characteristics 17 (including mode damping) of the power system based 18 on measurement of the small perturbations, on at 19 least some lines of the system over a period of 20 time; 21 (b) measuring the power system operational 22 parameters, including power flows, during said 23 period of time; 24 (c) using the data collected in steps (a) and (b) 25 to establish relationships between mode damping 26 characteristics and power system operational 27 parameters in each of said lines; and 28 (d) calculating from said relationship power flow 29 limits, providing the required level of confidence 30 in the security of supply, for each line. 31 8 1 Step (d) is preferably carried out by multivariate 2 analysis of data from a number of points on the 3 system, to identify fixed power system operational 4 parameters/damping relationships and to deal with 5 otherwise unexplained random variations in the 6 relationships. 7 8 Brief Description of the Drawings 9 10 Embodiments of the invention will now be described, 11 by way of example only, with reference to the 12 accompanying drawings, in which: 13 14 Figs 1, 2 and 3 are graphs illustrating the 15 relationship between power flow and mode 16 damping ratio; 17 Fig. 4 is a graph illustrating the relationship 18 between power flow and mode decay time; and 19 Figs 5 to 7 are graphs relating to a further 20 embodiment. 21 22 Best Modes for Carrying Out the Invention 23 24 An embodiment of the method of the invention 25 involves measuring the small signal or steady state 26 dynamic characteristics of the power system. The 27 measurements are obtained continuously and online 28 and are made and associated with the actual 29 prevailing system conditions. 30 31 Specifically, the dynamic characteristics which are 32 measured are the mode damping, mode frequency and 9 1 mode amplitude of the system based on real or 2 reactive power flow, voltage, system frequency, etc. 3 4 These mode values are in turn associated with the 5 actual prevailing system parameters such as real and 6 reactive power, voltage, system frequency etc. 7 8 In addition, the time of occurrence of the 9 measurements is always taken. These measurements 10 are then used to determine the small signal power 11 transfer limits of the system and thus ensure the 12 most efficient power transfer operational 13 conditions. 14 15 Preferably the measurements are modified by a. factor 16 which is determined as follows: 17 18 As a first stage a series of small signal dynamic 19 characteristic values, preferably the mode damping 20 characteristics, and corresponding operational 21 parameters of the power system are obtained from a 22 mathematical model of the system prior to a limiting 23 transient or voltage condition resulting from an 24 identified contingency event. 25 26 From this series of values there is derived a 27 relationship between the small signal dynamic 28 characteristic and the transient limit under the 29 particular 'system conditions. This relationship is 30 then used together with the actual online small 31 signal measurements to determine the power transfer 32 limits of the system.
10 1 2 The technique can provide on-line information as to 3 how close a power system is to its transfer limit. 4 5 The damping characteristics may be measured in terms 6 of damping ratio, mode decay time, or any other 7 suitable form. 8 9 The embodiment of the invention further includes 10 deriving the safe limits calculated by the foregoing 11 process to one or more contingency conditions. 12 13 The invention is based on making use of actual past 14 data of power system operational parameters and 15 mode damping in the system of interest. Such data 16 may be derived from historic data manually or by 17 other known means. 18 19 It is preferable to utilise a power system dynamic 20 model which has been verified by continuous direct 21 dynamics measurements under observable operational 22 conditions so that the required level of confidence 23 can be placed in the model predictions. 24 25 On-line estimates of power system dynamics are 26 derived in terms of mode frequency, amplitude and 27 damping by acquiring power system operational 28 parameters at one or more points on a network and 29 analysing the small perturbations that are always 30 present. 31 11 1 A number of features can be noted from this data. 2 For example, when the real power flow on a line 3 increases, the mode damping on the line 4 deteriorates. Also, in many instances the mode 5 damping ratio is approximately linearly related to 6 the MW power flow on the line being monitored. 7 For example, in Fig. 3, the relationship between the 8 power flow and the mode damping ratios is 9 10 Damping Ratio = 0.105 - 0.0005 x Power Flow 11 12 The small signal power transfer limit is established 13 as the point where the damping ratio is zero, i.e. 14 the power flow is 0.105/0.0005 = 210 MW. 15 16 However, this situation does not always pertain. 17 Suppose that two lines in a network are being 18 monitored (Line 1 and Line 2) and that there are two 19 modes present on these lines (Mode A and Mode B). 20 On Line 1, the damping for Mode A is related to the 21 real power flow on Line 1 as shown in Fig. 1. 22 However, on Line 2 it may be that the damping ratio 23 for Mode A is totally unrelated to the real power 24 flow in Line 2 as shown in Fig. 2. At the same 25 time, the damping ratio for Mode B may be 26 approximately linearly related to the real power 27 flow in Line 1 and in Line 2. 28 29 At first sight it might appear that there is not a 30 clear relationship between damping and real power 31 flow. 32 12 1 To resolve this problem, it is necessary to 2 recognise that the apparent lack of relationship 3 arises because the situation is being viewed as a 4 univariate problem, whereas it is in fact a 5 multivariate problem. In establishing the 6 relationship between the damping for a mode and the 7 MW power flow on a line, account must be taken of 8 the simultaneous power flow and possibly damping on 9 other associated lines. 10 11 By use of multivariate analysis (or neural network 12 as appropriate), one can establish for each line in 13 the system which modes have a fixed relationship 14 between power system operational parameters and mode 15 damping ratio. This can be done, for example, by 16 forming a matrix of observations, from which can be 17 derived the parameters for the multivariate 18 relationships. The fixed multivariate relationships 19 can then be used as discussed above to calculate 20 power transfer limits for each line. 21 22 Another source of 'unexplained' variation in the 23 relationship between damping and MW power flow on a 24 line was recognised to arise from inefficient or 25 mal-operating control systems associated with 26 individual generators and plant. 27 28 As an additional benefit it is possible to identify 29 which observed generators/plant are contributing to 30 system damping, by including observations on the 31 power system operational parameters associated with 13 1 individual generators and other associated plant, 2 when establishing the multivariate relationship. 3 4 It should be noted that the transfer limits derived 5 in this way apply to both voltage and transient 6 instability; in this way the likelihood of whether 7 the power system can withstand any specific 8 contingency can be estimated. 9 10 It is important to realise that this assessment is 11 based largely on observational data, and the only 12 network modelling involved is required in order to 13 derive relationships between pre-contingency small 14 signal dynamic characteristics and transient limits. 15 16 Another important feature of the technique is that 17 by basing transfer limits on the prevailing power 18 system dynamic conditions, 'conditional' rather 19 than 'marginal' probabilities are being used and 20 this fact leads to additional accuracy and 21 flexibility in the assessment of transfer limits. 22 23 In another aspect there are frequent incidents in 24 high voltage transmission systems where a poorly 25 damped oscillation 'spontaneously' appears in the 26 power flow on a transmission line. These 27 oscillations can be sustained for a few minutes or 28 extend to several hours. During the time when these 29 poorly damped oscillation are present, the power 30 system is being exposed to a power system supply 31 security risk, that usually takes the form of a 32 voltage or transient instability risk. The source 14 1 of the oscillations is often related to the 2 maloperation of the control systems associated with 3 electrical generators or other plant connected to 4 the transmission system. 5 6 There are many generators and other plant items 7 connected to a transmission system and it is 8 currently very difficult to identify from among 9 these many items, which individual item of plant is 10 at fault. When the individual plant item is 11 identified, corrective action can be taken and so 12 reduce the risk of loss of supply. 13 14 The present invention also provides a means of 15 identifying the individual plant item based on the 16 modal analysis of measurements of voltage, current, 17 real power flow, reactive power flow and system 18 frequency taken from the power system. 19 20 The dynamic characteristics for the power system are 21 measured, in terms of mode frequency, mode damping 22 and mode amplitude. By comparing some or all of 23 these modal measurements between the various 24 measurement locations on the network, identification 25 of the location of the source of the poor damping is 26 made possible. Hitherto, it has been traditionally 27 assumed that the frequency and damping for a mode is 28 constant across the transmission network, with only 29 mode amplitude varying from one location to another 30 (i.e. that the power system behaves in a linear 31 manner in this respect). 32 15 1 Where modal analysis is done on signals acquired 2 close to individual plant, this non-linear behaviour 3 of the power system can be utilised to identify 4 individual maloperating plant. 5 6 Another aspect whereby detection of the source of an 7 oscillation is made possible is by means of 8 examining the phase relationship for modes 9 associated with power system frequency and real (or 10 reactive) power. Where the source of an oscillation 11 is in close proximity to the point of measurement, a 12 greater difference in phase between modes can be 13 expected than the case where the point of 14 measurement is remote from the oscillation source. 15 16 An example of the use of this aspect of the 17 invention will now be described with reference to 18 Figs 5 to 7. 19 20 Fig. 5 illustrates a 'spontaneous' oscillation 21 typical of what can often be seen on transmission 22 systems. Fig. 6 illustrates the results obtained 23 when the real power, reactive power and voltage 24 signals for this event are analysed at point A on 25 the transmission network that is electrically remote 26 from the source of the oscillation. Fig. 7 27 illustrates the results obtained when the real 28 power, reactive power and voltage signals for this 29 event when analysed at point B in the transmission 30 network that is electrically close to the source of 31 the oscillation. 32 16 1 It can be seen in this illustration that the mode 2 decay time constant for the reactive power signal at 3 point B: 4 5 (a) has a higher level in terms of average and 6 maximum value over the period of the 7 oscillation 8 (b) achieves a high value at a point in time that 9 is in advance of the corresponding signal at 10 point A 11 12 It can also be seen that the average and maximum 13 values achieved for the real power and voltage 14 signals during the period of the oscillation are 15 generally higher at point B than at point A. 16 17 These and other similar indications in the modal 18 analysis results for the various signals, identify 19 that the source of the oscillation lies closer to 20 point B on the network than point A. Taking similar 21 measurements at other points on the network that are 22 more electrically remote from point B, confirms that 23 the source of the oscillation lies close to point B. 24 25 By this means the source of the oscillation has now 26 been localised / identified and corrective action 27 can now be targeted on the individual plant, thus 28 removing this risk of loss of supply. 29 30 The exact form of the manifestation of a fault 31 within the various signals, depends upon the type of 32 plant fault or maloperation. The precise nature of 17 1 the manifestation may well be used to classify the 2 fault and provide guidance for the form of remedial 3 action that should be taken. 4 5 Modifications and improvements may be made within 6 the scope of the present invention.