Summary of the Results

The main conclusions of the thesis are summarized below, and are the result of applying the proposed automatic method to the field measurements described in Chapter 4.

Influence of the load characteristic in voltage stability

The effect of the load characteristic in voltage stability studies has been investigated. The use of the traditional static load models is changing more and more in favor of the dynamic ones in order to get a better representation of the load, and therefore to optimize the system operation.

The fact that loads are generally voltage dependent is a critical aspect for the design of security margins. A very pessimistic design may over-weight this security and this may result in a poor utilization of the system. On the other hand, a very optimistic design may lead the system to voltage collapse. It has been also studied that the voltage sensitivity of the load may help the stability of the system by providing some system load relief.

Moreover, some types of load such as electric heating are especially critical for stability because of their thermostatic characteristic. After a disturbance in the system, (voltage and power drop), and due to the effect of the thermostats, the aggregated load tends to increase the nominal load to a level close or equal to the pre-disturbance one, at the low voltage. This situation may result in severe conditions for the system operation. However, during the recovery time it may be possible to take some corrective actions such as local reactive compensation, load shedding, starting small-scale gas turbines, which may lead the system to stable operating points instead.

Window length

The influence of the data sequence length has been studied. A good estimation is achieved by choosing a window length of 2.5-3 times the load time constant, after a voltage change. If the identification is limited to determining the transient characteristic of the load, the measuring time can be considerably reduced to an approximate value of 15% of the time constant. Moreover, the measurements have verified that during the first

15-30% of the recovery the load behaves as a resistance, i.e. transient characteristic in the neighborhood of 2.

Normalization on reactive dynamic load models

The influence of the normalization in dynamic reactive load models has been studied throughout Section 6.4. By using measured data from normal operation it is illustrated that the reactive power level Qo traditionally used to normalize reactive load models is inappropriate, since it may be equal to zero due to the effect of reactive compensation. The identification of parameters for the reactive load model when normalizing by Qo provides values that tend to infinity when Qo goes to zero.

If instead active power level Po, or apparent power level So is used as a base, the variability in the parameters that describe the reactive load response is drastically reduced.

Dynamic load parameters

According to the identified results from the determination of the active and reactive load parameters in Chapter 7 the following has been achieved:

Active and reactive time constants

In general, the values of both time constants move in the same order in a range of about 80 to 200 seconds, and they are correlated to each other with a correlation factor of 0.50. The difference in the order of the time constants is probably related to the effect of spontaneous load demand variations, but the main recovery of the reactive load is produced by the increase of the reactive losses due to the active recovery of the load at low voltage.

The average seasonal and daily time constant values indicate, faster recovery of the load during warmer months. The lowest values correspond to evening hours during summertime when the load demand reaches its minimum due to holidays season and to low consumption of electric heating. However, the monthly and daily standard deviation of the parameters, show a large variability in the distribution of the results due to the high diversity of load processes, even during the day. In order to group

the results more accurately, it would be useful to apply alternative analysis techniques as well as other ways of grouping the data.

Active transient load voltage dependence αt

The obtained correlation factor (-0.82) between this parameter αt, and the outdoor temperature during the measurements shows the strong dependency of the active transient characteristic of the load with the season, time of the day and weather conditions. The larger values of this parameter, close to 2, correspond to the colder months of the year, especially during the night, and they show a pure resistive characteristic of the load, which is related to the increase in using electric heating load. On the other hand, the lower values of the parameter correspond to summertime where the heating load consumption is very low.

Reactive transient load voltage dependence βt

The parameter βt is associated to the load composition and the presence of induction motors in it, but also to the effect of distribution transformers operating in saturation.

A high dependency of this parameter on the temperature has been found, where the higher values correspond to summertime and day hours, and the lower to wintertime and night hours. Since the measurements correspond to the same area, the main varying condition is the temperature, and the low variability in the parameter is probably related to connection/disconnection of air conditioners, heat pumps, and other similar loads during winter and summer.

Active and reactive steady state load voltage dependence αsβs

The active and reactive steady-state characteristics of the load, αs and βs, describe the respective active and reactive time recovery. It has been found that during wintertime the values of αs are closer to 0, and therefore to a full restoration of the load, while larger values are obtained during summertime.

Moreover, some of the results exhibit negative values, corresponding to overshooting in the recovery of the load because of the action of discrete tap changers under low voltage conditions, that may contribute critically to the voltage stability of the system.