Fig 10 - uploaded by Leonardo Di Stasio
Content may be subject to copyright.
Cumulative distribution function of time to next event registered in 2015 for the substation at busbar #26
Source publication
This paper refers to the voltage sags, which are among the most critical Power Quality disturbances. Thanks to the availability of measured voltage sags for four years in the same regional electric system, the work is our first attempt of using the registered voltage sags in three years for forecasting the expected voltage sags in the fourth year....
Citations
... This feature was first evidenced in [16]. This constitutes a crucial aspect that influences the statistical modeling of measured voltage sags and their practical application in forecasting sags at every site within an entire electrical system, as demonstrated in [17][18][19][20][21]. Rare voltage sags can be modeled effectively as a Poisson process with a constant and positive rate of occurrence owing to their features of timeinvariance, memory-independence, and independence from each other. ...
... They typically are dependent events because they are linked to common phenomena that are external to electric systems, such as bad weather conditions, storms, and fires [16,19]. In [17][18][19], the forecast of rare sags was proven to be possible only if the clusters were ascertained and removed. In [19] new intermittence indices were proposed to improve the removal of the clusters and the successive forecast of rare sags. ...
... For example, a strong correlation of sag clusters with significant lightning activity in some areas of the system may suggest an improvement in the performance of the lines related to lightning in those areas. In all the papers [16][17][18][19][20][21], the statistics of the measured sag were modeled by the random variable tim to th n xt v nt of th sit , ttn , rather than the numb r of sags, N. The random variable ttn , which measures the time between successive voltage sags at each site denoted by , provided a significantly larger database compared to N. This allowed forecasting rare sags with errors lower than 10% using the sags measured in only 3 years with any frequency. Given that clusters continue to play a crucial role in representing the severity of voltage sags at various nodes within the power system, in successive papers [20,21], the forecast of all the sags (named total sags), that is rare sags and clusters of sags, was first approached. ...
The field measurement campaigns have revealed that voltage sags also occur as clusters and not only as rare phenomena. The clusters of sags represent a stochastic process due to their time dependence; the rare satisfy the requirements for a Poisson distribution process. To forecast both kinds of sags using the statistics of the measurements, different approaches are required. In this study, a general method for predicting both types of sags is proposed with a procedure that can be implemented automatically. The method uses intermittent indices to distinguish between the sites that have a prevalent number of rare sags and the sites where rare sags and clusters occurred. Based on this means of identification, the technique offers two distinct models for predicting each kind of sag. The final goal is to implement the procedure in a measurement system that can automatically pre-analyze the recorded sags and choose the best technique for prediction depending on the type of sag. The first results were satisfying with forecast errors reduced in comparison with those obtained without the proposed procedure.
... rare sags without clusters, i.e. caused by short circuits [12][13][14][15]; ...
... It is based on the plotting of the quantiles of an empirical distribution derived from a dataset against the corresponding quantiles of the theoretical distribution. The two main analyses allowed by the Q-Q Plots are the distributional comparisons between the two distributions [12], and the estimation of one distribution parameter of the theoretical distribution [13][14][15]. This second target was used in this study. ...
... 13 Low voltage ride through and Over voltage ride through characteristics for static generators. ...
The availability of data measured in the field is one of the greatest advances in the last decades for electric power systems planning, management and control. The technological innovations of systems and devices for the measurement of a wide range of quantities allowed their installation at almost all levels of voltages, from distribution to transmission. Increasing number of proposals of new techniques are coming from researchers and system operators for facing the various problems of system operation relying on the availability of measured data in real systems. The data from the field enable data-driven approaches, which can be integrated into traditional model-based methods constructing the so-called digital twin of a system. It is a system digital model whose parameters and linkages are continuously changed and tuned in accordance with the measurements from the actual system in service. New and attractive possibilities for addressing the problems of planning, managing, and controlling the power systems are opening for researchers and system operators.
The studies presented in this thesis used the voltage sags measured in the medium voltage (MV) regional systems of E-Distribuzione for four years. Two main problems were faced by data-driven approaches: the ascertaining the origin of the measured voltage sags, and the forecasting the voltage sags which will occur in the site of the system. The correct ascertaining of the origin of voltage sags is crucial in view of future economic regulation by the national energy Authorities. It gives to the system operator the correct indication if the measured sags were or not due to faults in his own network interconnected with other systems. The possibility of forecasting the future performance in terms of voltage sags per year was a challenge never dealt with in the literature for the absence for field data. The literature, till now, proposed methods, models, and tools to estimate the average performance of a system, derived from model-based approaches.
Regarding the origin of the voltage sags measured at the MV busbars of the High Voltage/Medium Voltage (HV/MV) stations, this thesis analyses and compares two methods which use only the residual voltage, the time the voltage sags occurred, and their duration. The analysis was conducted also studying the effect of the presence of DG (Distributed Generation). The data from field, moreover, revealed that in some specific cases the sags caused by faults in the MV systems propagated to HV networks. This problem was also studied simulating a portion of a real system, which presents the interconnections between HV and MV network.
Regarding the forecast of voltage sags, the thesis proposed two main methods which use at least three years of measurement. The common choice of these two methods is the selection of the random variable, different from the statistical variable used by all the methods in literature. The random variable used in this study is the time to next event, that is the time intercurred between each couple of sags, instead of the variable number of voltage sags. This choice allows a huge increase of the data sets with the positive consequence of reasonable measurements time to obtain a forecast with acceptable accuracy. The first method, based on Poisson model, is suitable for rare sags, that are the sags occurred with a time to next event of the dimension of hours. The second method, based on Gamma model, is suitable for all the voltage sags, comprehensive of sags occurred close each other as groups. The latter, named clusters, are typically due to exogenous causes from the power systems, like adverse climatic conditions or fires.
Intermittent indices are also proposed introduced for an initial screening of the measured sags to focus if, and how many, clusters are present in the data base of the measured sags. Such analysis drives the successive steps of the statistical analyses for discriminating the adequacy of Poisson and Gamma models.
... Model-based simulations, both numerical [14], [15], and analytical methods [16], [17] give the expected performance at the nodes of the systems in terms of the average annual number of voltage sags, which is not a forecast in strict sense. Some recent papers [18]- [21] addressed the forecast of the expected number of voltage sags. The key factor of such approaches was not only the availability of large amounts of measured sags, but also a different choice of the random variable to describe the statistics of the sags. ...
... This value is computed estimating the average sag annual rate at that site, λk, from the measured sags in a precedent period. In [18]- [21] the time needed to obtain acceptable accuracy was proven to be at least three years. ...
... Once known λk, the wanted value Nf,k is given by the relation [18]- [22]: ...
... In the opposite case, a gamma distribution fitting was confirmed. In [59], some ideas for voltage sags forecast using the parameter time to the next event as the probabilistic variable are given. Further elaboration is provided in [60]. ...
In this paper, a novel method for real-time prediction of voltage sag duration is proposed. It is based on the recently introduced new characteristic of voltage sag, named harmonic footprint, and is formulated using a logistic regression model. The concept is mathematically formulated and statistically analyzed using an extensive set of real grid measurement data which are recorded in distribution grids. Furthermore, the proposed method is applied as a part of an advanced grid-tie converter control. It is included in previously developed methods for fast sag detection and magnitude of voltage sag prediction. The algorithm is applied to the control of grid-tie converters used in distributed generators and tested with real/grid measurement data in the IEEE 13-bus test grid by simulations and in the IEEE 33-bus test grid using a hardware-in-the-loop (HIL) microgrid laboratory testbed. It is shown that this method can prevent unnecessary tripping of distributed generators (DG) and improve low-voltage ride-through (LVRT) support. In addition, the model has the potential to be applied to a wide range of devices or algorithms for the protection, monitoring, and control systems of distribution grids.
... Recently, some papers have approached the problem of assessing the future occurrence of the sags at network sites from a completely different point of view [13][14][15]. The main novelty was to pass from the estimation of the average performance of a system obtained from simulations to the forecast of the number of sags at the sites of a system obtained from the statistics of recorded sags. ...
... Concerning the rare events, they are sags with a frequency in each site in the range from some sags at year to some sags at month. All the relevant literature was aimed to estimate the average performance of a systems, focusing the attention on rare sags [7][8][9][10][11][12][13][14][15]. ...
... The presence of the sag clusters [13][14][15][16] makes the sag frequency dependent on the time. The sags, which contain both rare sags and sag clusters, present the features of a stochastic process. ...
The forecast of the occurrence of voltage sags at
the sites of a system is nowadays feasible thanks to the
availability of huge quantity of recorded data. To forecast future
performance from the statistical analysis of recorded sags, the
stochastic modelling of the voltage sags is required since the
events are not statistically time independent. The presence of
groups of sags, named clusters, brings the phenomenon far from
the conditions of Poisson model. This paper proposes the
Gamma distribution to model the sags, which also include the
clusters. Different techniques for assessing the parameters of the
Gamma distribution are presented and applied to forecast the
number of sags expected at selected sites in the year 2018, i.e.,
the year successive to those when the sags were measured. The
outcomes of the forecast are compared with the sags effectively
occurred in those sites in the year 2018, using different criteria
for evaluating the forecast error. The results showed the
viability of the approach and encourage further studies to
improve the accuracy and extend the forecast to entire systems.
... This study was conducted without performing any simulations, and only the available data were used to conduct the statistical analyses. This paper is an extended version of our earlier paper [25], in which the results of the preliminary studies were presented. These preliminary studies were focused mainly on the evaluation of the statistical techniques that were used to assess the input data that came from the field. ...
... The overriding goal of this study is to give to the network operators proper algorithms that will allow them to forecast how many voltage sags will occur and the sites at which they are likely to occur. Also, this extended version of the paper [25] allows identifying the details of the main critical aspects that are encountered, and it proposes adequate solutions, which are in progress. ...
... In this paper, which is an extended version of an earlier paper [25], we presented the preliminary results of the study, indicating, with more details than were presented in [25], the main critical aspects encountered and some solutions to overcome them. ...
This paper presents the preliminary results of our research activity aimed at forecasting the number of voltage sags in distribution networks. The final goal of the research is to develop proper algorithms that the network operators could use to forecast how many voltage sags will occur at a given site. The availability of four years of measurements at Italian Medium Voltage (MV) networks allowed the statistical analyses of the sample voltage sags without performing model-based simulations of the electric systems in short-circuit conditions. The challenge we faced was to overcome the barrier of the extremely long measurement times that are considered mandatory to obtain a forecast with adequate confidence. The method we have presented uses the random variable time to next event to characterize the statistics of the voltage sags instead of the variable number of sags, which usually is expressed on an annual basis. The choice of this variable allows the use of a large data set, even if only a few years of measurements are available. The statistical characterization of the measured voltage sags by the variable time to next event requires preliminary data-conditioning steps, since the voltage sags that are measured can be divided in two main categories, i.e., rare voltage sags and clusters of voltage sags. Only the rare voltage sags meet the conditions of a Poisson process, and they can be used to forecast the performance that can be expected in the future. However, the clusters do not have the characteristics of memoryless events because they are sequential, time-dependent phenomena the occurrences of which are due to exogenic factors, such as rain, lightning strikes, wind, and other adverse weather conditions. In this paper, we show that filtering the clusters out from all the measured sags is crucial for making successful forecast. In addition, we show that a filter, equal for all of the nodes of the system, represents the origin of the most important critical aspects in the successive steps of the forecasting method. In the paper, we also provide a means of tracking the main problems that are encountered. The initial results encouraged the future development of new efficient techniques of filtering on a site-by-site basis to eliminate the clusters.
... This paper follows two preceding papers [18,19] in which we reported that the measured voltage sags were grouped in clusters. In [18,19], first, we used the same filter of [17], that is the 1-hour filter, on the voltage sags measured over a threeyear period (2014, 2015, 2016) in a regional MV network. ...
... This paper follows two preceding papers [18,19] in which we reported that the measured voltage sags were grouped in clusters. In [18,19], first, we used the same filter of [17], that is the 1-hour filter, on the voltage sags measured over a threeyear period (2014, 2015, 2016) in a regional MV network. Then, the sags, without the filtered clusters, were used to forecast the performance in the fourth year (2017). ...
... Fig. 1 shows the scheme of the method. With the same approach that was used in [18,19], we used the registered data from the field over a period of N years to forecast the sags in the successive year. There are two separate blocks, which are included in rectangles with the dashed lines. ...
This paper provides a means of forecasting the future average performance a regional electric power system in terms of sags per year based on data accumulated over a four-year period. The paper also presents statistical analyses of the measured data. The sags measured in real systems consist of both rare voltage sags and grouped voltage sags (clusters); clusters are stochastic events that must be detected and removed to allow the estimation of the future performance based on a Poisson process. To detect and remove the clusters, new site indices of intermittence are proposed. These indices were measured in every site at which clusters were present, and the measurements included the frequency, time of grouping, and the large number of sags that constituted them. The performance of the forecast was evaluated by comparing the results obtained using the proposed indices with the results derived using a time filter equal for every site. The results showed that the forecast of the rare voltage using a few years' worth of field data had acceptable errors and was not prohibitive. The main aspect, which must be investigated further, is related to the choice of the best threshold on the intermittence site indices.
