Electromagnetic relay (EMR) is widely applied in industrial control system and communication as a kind of basic electric component, also playing a role in switching circuits, transferring signals, and eliminating interference. However, when a short-circuit fault occurs in the power system, it causes power quality issues, such as voltage sag and short interruption (Nagata et al., 2018; Wang et al., 2019; Ye et al., 2019), which affect the normal operation of EMR. Voltage sag even causes the malfunction and damage of EMR in severe cases (Wu and Fan, 2015; Jianbo and Qi, 2018), then leading to the failure of the whole industrial process and resulting in the huge economic losses (Mohammadi et al., 2017; De Santis et al., 2018; Gambôa et al., 2019). For example, (Bollen, 2000) records that the tripping of EMR under one voltage sag event “cause the shutdown of a large chemical plant, leading to perhaps $100,000 in lost production.” However, the tolerance of EMR under voltage sag presented in Bollen (2000) is from IEEE Std. 1346 (IEEE, 1998) which published the sensitivity of EMR 20 years ago. The data in IEEE Std. 1346 may be not suited for the industry today, because of the improvement of the technique of EMR and the updating industry processes. It is important to understand the tolerance of EMR under voltage sag nowadays to provide the voltage sag mitigation scheme and the EMR purchase plan for the industry users.

EMR is known as the sensitive equipment under voltage sag because of industrial users’ complain (IEEE, 1998; Bollen, 2000; Zhai and Yang, 2008). IEEE Std. 1346 (IEEE, 1998) pays close attention to it and present that tolerance of EMR is 60–75% of rated voltage of magnitude, and is 10–30 ms of duration; however, this standard states that the presented tolerance “should not be considered typical for these types of devices but only a samples of what is available.” It is difficult to provide the detailed guide for the users. The simulation was done to investigate the dynamic characteristics of EMR interfered by voltage sags and short interruptions to offer the qualitative understanding of the sensitivity of EMR (Zhai and Yang, 2008). It is a necessary way to get the tolerance information of sensitive equipment than conducting the extensive tests on equipment. The tolerance of other sensitive equipment is also obtained through test, for example, the tests on adjustable speed drive (ASD) (Xu et al., 2019a), programmable logic controller (PLC) (Xu et al., 2019b), AC coil contactor (ACC) (Djokic et al., 2004; Hardi et al., 2010; Weldemariam et al., 2016), and so on (Ouyang et al., 2015; Ouyang and Liu, 2017).

The different single-event characteristics of voltage sag is considered in the test. Magnitude and duration are the essential characteristics which should be included in the test. Furthermore, other characteristics should be included, such as point-on-wave (Alvaro et al., 2019; Wang et al., 2020) and phase angle jump (PAJ) (Wang et al., 2015). The former is also called phase angle, and the latter is called phase shift in IEC 61000-4-30 (IEC, 2021). POW is the phase angle of the fundamental voltage waveform at which the voltage sag occurs; PAJ is the change of phase angle during the voltage sag (Djoki et al., 2007; Wang et al., 2015; IEEE, 2017; Alvaro et al., 2019; Wang et al., 2020; Ren et al., 2021). There are two views about considering the other characteristics. For one thing, IEEE Std. 1346 recommends that “phase shift and point of initiation not be considered,” since these “characteristics are not typically available in the sag environment data.” However, for the improvement of the voltage sag generator and the monitor technical, the problem of “sag environment data” has not been the barrier of the testing. Another opinion is that POW and PAJ have significant impacts on electromagnetic equipment (Djoki et al., 2007; Wang et al., 2015; Alvaro et al., 2019; Wang et al., 2020; Ren et al., 2021); similar thinking is also recommended in the standards. IEC 61000-4-30 (IEC, 2021) states that for some equipment “drop-outs, the phase angle at which a voltage dip begins is an important characteristic”; IEEE Std. 1668-2017 (IEEE, 2017), which is the newest standard related to testing on the sensitive equipment under voltage sag, suggests that “…characteristics such as phase-shift, point-on-wave of initiation, and recovery of the instantaneous voltage waveform during the sag may have an effect on equipment performance as well.” Thus, the various single-event characteristics of voltage sag are included in this paper to test the sensitivity of EMR.

This paper proposes a general test procedure and has done extensive tests to investigate the tolerance of EMR to get the quantitative results and support the voltage sag mitigation for EMR. The rest of this paper is organized as follows. Section 2 describes the structure and response mechanism of EMR. Detailed test plan is elaborated in Section 3, including test principle, test procedure, equipment under test (EUT), and voltage sag generator. Test results are presented in the form of voltage tolerance curves in Section 4. Section 5 discusses the potential application, and Section 6 makes a conclusion of the test results.

This study proposes a general test procedure to improve the test effective and the accuracy. The comparisons of the effectiveness and the accuracy between the proposed test procedure and the test method proposed by IEEE standard (IEEE, 2017) are listed in Table 5.

The results in Table 5 are the test results for R2, when 0° POW and 0° PAJ. The “real” critical points of the tolerance under the voltage sags with the different residual voltage are testing results by the traditional method, which is called “step by step” method. The step size is 1 ms; the small step size ensures the testing accuracy but with the heavy testing work. The real tolerances are 19, 19, 19, 6, 5, 6, and 14 ms when magnitudes are 0, 10, 20, 30, 40, 50, and 60%, respectively. However, the traditional method costs the extensive work, for example, it has to do 1,000 tests, when the critical point is 1,000 ms for a certain residual voltage.

IEEE test method (IEEE, 2017) is the fixed step test method. For example, the first step to the seventh step is testing the tolerance of EUT when duration is 0, 10, 20, 50, 100, 200, and 500 ms, respectively. The greater test step sizes result in the greater test errors. The example in Table 5 is not with the great errors, because of the critical points are less than 20 ms, the step size in this range is small by IEEE method. The worse situation is that when the critical point is 300 or 400 ms, the test result should be 500 ms, leading the larger error.

The proposed method can decide the next test point automatically according to the test results of the two previous test points, can quickly approach the critical point. The errors of the testing results for R2 are less than 1 ms compared with the real tolerances. The test work is 8 times for a certain residual voltage, which is quite less than the test work by “step by step” method, whose testing works are 19, 19, 19, 6, 5, 6, and 14 times for the different residual voltage for R2. The proposed method is with the high efficiency and test accuracy.

First, through numerous tests, it finds that POW and PAJ are the major affected factors, as well the two-stage sag. For the reason that, it can carry out a subset tests described in this paper to analyze the voltage tolerance of EMR, only considering magnitude, duration, POW, PAJ, and the two-stage sag.

Secondly, based on the test results got, the general VTC of EMR are obtained as shown in Figure 12. The general VTC is helpful for industry consumer to choose the proper EMR considering their own operating environment. Moreover, it is useful for the manufacturer to design the EMR with higher voltage tolerance for specific consumer. For example, EMR can work steadily in a longer duration environment.

Third, EMR can tolerate a voltage sag with the magnitude more than 74% of U _nom no matter how long the sag lasts from the general VTC in Figure 12. The result is profitable for providing appropriate sag countermeasures. For example, the voltage is only compensated to 74% of U _nom to ensure the EMR operate properly, and the entire industry process would not be interrupted, accordingly. It means that industrial users can choose a voltage-compensating device with a smaller capacity to maintain the whole process and save money, more importantly.

This paper has performed a large number of tolerance tests on eight different EMRs, and the test results are presented graphically as VTCs. The tolerance to rectangular and non-rectangular voltage sag is investigated in the tests; the influence factors considered here include POW, PAJ, frequency variation, harmonic, magnitude variation in pre- and post-sag segments, two-stage sag, and slow recovery sag.

The obtained VTCs clearly show that the response of EMRs to voltage sag or short interruption can be rather complex. The magnitude and duration of sag are not the only parameters which influence EMR’s sensitivity; POW and PAJ also have an important influence on the EMR’s tolerance.

The following specific conclusions can be drawn from the tests performed.

1) From the tolerance test under rectangular voltage sag, the tolerance capability of EMR is obtained. Although the EMR’s VTC is different depending on its type, the shapes of VTCs of all tested EMRs are basically similar. Through more than 10,000 tests on each EMR, EMR can tolerate a voltage sag with the magnitude more than 74% of U _nom . When the sag magnitude is less than 48% of U _nom , EMR will trip when the sag duration exceeds 28 ms.

The test results are useful to choose proper EMR with higher tolerance and to use subset tests to quantitatively analyze sensitivity of a new EMR. It is profitable to apply a voltage compensating device with a smaller capacity to keep the process unaffected using the general VTC.