RU2515996C2 - Method for dynamic detection of accidental electric discharge and device for its implementation - Google Patents

Abstract

FIELD: electricity.

SUBSTANCE: application: in the field of electric engineering. According to the method a recorded signal i(t) of load current is supplied into two channels, each channel has three processing steps, at that at the first step of the first channel deviation of current signal from the periodic function is assessed, a regular component being the determinate part of the source signal is then defined by means of the source signal averaging in a short period of time - more than ten times less than the reporting period T and its mean-square value is calculated; at the second step high frequency components are defined excluding signal components slowly varying during several reporting periods T; at the third step the signal is filtered in order to exclude influence of certain pulses with large amplitude and history of its previous values is considered for the long-term period of time with reduction of each value contribution into the result depending on its time limitation; at that at the first step of the second channel an irregular component is defined by means of a regular component deduction from the signal i(t) and its mean-square value is calculated; at the second step high frequency components are defined and at the third step the signal is filtered, signal levels obtained at the outputs of the first and second channels are compared with the preset threshold values and when they exceed these values an alarm command is generated in order to cut off the controlled mains.

EFFECT: improving reliability.

8 cl, 4 dwg

Description

The invention relates to protection schemes in the field of electric power and fire safety, performing automatic shutdown and directly responding to an unacceptable deviation from normal electrical operating parameters with subsequent restoration of the connection or without it, namely, to methods for the dynamic detection of emergency electrical discharge.

In electrical installations and cable lines, due to mechanical, chemical and temperature influences, insulation damage is formed, which leads to a violation of the functioning of the electrical circuit, a change in the set of characteristics corresponding to it and the emergence of emergency operating modes. The greatest difficulty in detection is emergency conditions, in which physical processes with electrical discharges occur.

A feature of modern electrical installations is the presence in them of devices with single or periodically repeating electrical discharges or devices that, in normal operation, generate a current shape similar to the discharge current. In this regard, when developing a method for the dynamic detection of an emergency electrical discharge, it became necessary to distinguish it into two classes - a regular electrical discharge and an emergency electrical discharge.

By standard electric discharge (SER) we mean a discharge of any type - spark, arc, etc., inherent in a number of electrical machines and devices (collector motors, switches, contactors, relays, welding machines, discharge lamps, etc.). The regular electrical discharge can also include electrical processes occurring in a controlled network, which have a number of characteristic features - transient and switching processes that accompany the operation of switching power supplies and voltage converters, frequency drive devices, etc.

An emergency electric discharge (AER) will be considered any type of discharge - spark, arc, etc., which is not typical for electrical installations, machines and devices in good condition, which can lead to ignition or ignition of insulating materials and surrounding objects and substances.

Emergency electrical discharge, in turn, it is advisable to divide according to a number of signs and features of detection into three types: serial, parallel and ground discharge.

Serial emergency electrical discharge occurs when a reliable electrical contact is broken in the circuit with the connected load. The series emergency electrical discharge current is limited by the rated load current.

A parallel emergency electrical discharge occurs during arc or spark discharges between phase and (or) neutral conductors.

Emergency electrical discharge to earth is a special case of parallel emergency electrical discharge and occurs during arc or spark discharges between phase conductors and grounded parts of the electrical installation (protective earth conductor).

In an emergency, single spark discharges initially appear, characterized by an extremely short discharge current. The released heat at a sufficient frequency of occurrence of spark discharges contributes to carbonization of the insulating material and the flow of the emergency electric discharge into the next, more stable phase - the arc discharge. Similarly, the process develops when the electrical contact in the terminal blocks and other switching products is broken.

The share of fires resulting from unreliable contact and damage to electrical components that do not lead to short circuits and overloads reaches 80% [D. Pavlov Research and development of an intelligent intrinsic safety device for automation systems: dis. Cand. tech. Sciences: 05.13.05 Vladimir, 2006 RSL OD, 61: 06-5 / 3843, p. 12], therefore, the method of protection against emergency electrical discharge is relevant and should be used on electrical installations of all categories, both industrial and domestic.

Existing protection methods using circuit breakers (AB) and residual current circuit breakers (RCDs) do not allow detecting an emergency electrical discharge and timely shutting down the damaged part of the electrical installation, therefore, to significantly increase the safety level of electrical installations, it was necessary to develop new protection methods.

A technical solution is known [Russian patent 2254615 G08B 17/06, G08B 25/10, publ. 06/20/2005], according to which the total electric current is measured at the input of the electric network. From the measured total current by filtering, a signal is isolated that contains higher harmonics that appear during sparking and burning of an electric arc. The generated signal is amplified, rectified and accumulated over a set time. During the accumulation process, the value of the total signal is compared with a given level, which is selected for the corresponding degree of fire hazard of the network. Depending on the value of the accumulated signal corresponding to the achieved level, a warning signal and (or) a command to turn off the electric network are generated.

The disadvantages of this technical solution is the low reliability of the detection of the electric arc when used in electric networks with a changing load and a high switching frequency, since the load current that changes when it is turned on and off, as well as the mismatch of the specified limits of the degrees of fire hazard to real conditions, are used as an informative indicator functioning of electrical installations.

The closest in its technical essence and the achieved result to the claimed invention is a method [application for US patent 2009/0059449 A1, publ. 03/05/2009], in which the registered current signal of the controlled network is supplied to two independent channels; the first channel contains a high-pass filter (HPF), designed to remove from the signal of the main harmonic of the AC network, the second - a low-pass filter (LPF), which serves to suppress the high-frequency components of the current arising, in particular, during sparking. The current signals of each channel after filtering are digitized and processed using a microprocessor as follows. The signal of the first channel is analyzed to identify one or more statistical characteristics. The signal of the second channel is used to calculate the rms value of the load current. The statistical characteristics of the signal of the first channel are compared with certain ranges of values depending on the rms value of the load current. If the values of the statistical characteristics fall within the limits set for a certain value of the load current, the values of the logical counter increase. When the limit value is reached, a signal is generated leading to a blackout of the controlled network.

Significant disadvantages of this technical solution are: the possibility of false operation of the device during normal operation of loads with high inrush currents and a high level of regular electric discharges - collector motors and switching power supplies, failure to ensure fire safety in the presence of an emergency electric, including arc, discharge, which is one of types of emergency electrical discharge of low power, the dependence of the limit level of emergency electrical discharge Yes, from the load current of the controlled network, leading to an unacceptable increase in its threshold value, the impossibility of stable registration of short-term spark discharges.

The objective of the proposed method for the dynamic detection of emergency electrical discharge is to ensure stable operation and the absence of false alarms for all types of loads, including in the presence of regular electrical discharges and inrush currents of any level, and to ensure a stable and constant threshold level of emergency electrical discharge based on experimental data on the minimum discharge energy necessary for ignition, registration of both arc and spark electric x bits.

The problem is solved in such a way that in the method for the dynamic detection of an emergency electric discharge, characterized in that the registered signal i (t) of the load current of the monitored network is supplied to two channels, each of which undergoes three processing stages, while in the first channel at the first stage the deviations of the recorded current signal from the periodic function are estimated, the regular component, which is the deterministic part of the initial signal, is isolated by averaging the initial signal over a small interval at time (more than ten times smaller than the analysis period T, and take its rms value, in the second stage, to increase the reliability of detection of an emergency electric discharge, high-frequency components are excluded, excluding slowly changing signal components over several periods of analysis of T, the signal is filtered at the third stage to exclude the influence of individual pulses of large amplitude and take into account the history of its previous values over a long time period with a decrease in the contribution of each value to the result m depending on the prescription; in the second channel, at the first stage, an irregular component is selected, which is the stochastic part of the original signal, by subtracting the regular component from the registered signal i (t) and taking its rms value; at the second stage, high-frequency components are isolated, excluding slowly changing signal deviations for several periods of analysis, and at the third stage, the signal is filtered to exclude the influence of individual pulses of large amplitude and take into account the history of its previous values for a long time period equal to three or more analysis periods, with a decrease the contribution of each value to the result, depending on its prescription, then the signal levels obtained at the outputs of the first and second channels are compared with predetermined threshold values and Achen, if they are exceeded, an alarm command is formed, leading to a controllable network blackout.

According to the inventive method in the first stage to determine the deviation of the registered current signal from the periodic function from the signal values for the current analysis period, the values for the previous analysis period are subtracted and the rms value x ₁ (t) is calculated, which is an estimate of the aperiodicity of the current signal by the formula:

, $$x_{\mathrm{one}}\left(t\right)=\frac{\mathrm{one}}{2\tau \sqrt{T}}\sqrt{{\displaystyle \underset{t-T}{\overset{t}{\int }}{\left({\displaystyle \underset{\rho -\tau }{\overset{\rho +\tau }{\int }}\left(i\left(\theta \right)-i\left(\theta -T\right)d\theta \right)}\right)}^{2}d\rho }}$$

,

where i (t) is the registered current signal;

T - analysis period - a fixed time period for alternating current and direct current, determined by the response time of the protective devices - circuit breaker or residual current circuit breaker, installed together and providing additional protection for the electrical installation;

t is the moment of arrival of the registered signal in the channels for processing;

τ is the radius of a small neighborhood of t;

ρ - time moments lying on the interval [t-T; t] over which averaging occurs;

θ is the moment of signal conversion.

According to the claimed method for isolating the irregular component in the first stage, the values of the regular component are subtracted from the values of the registered current signal, then the root mean square value x ₂ (t) is calculated by the formula:

. $$x_2\left(t\right)=\frac{\mathrm{one}}{\sqrt{T}}\sqrt{{\displaystyle \underset{t-T}{\overset{t}{\int }}{\left(i\left(\rho \right)-\frac{\mathrm{one}}{2\tau }{\displaystyle \underset{\rho -\tau }{\overset{\rho +\tau }{\int }}i\left(\theta \right)d\theta }\right)}^{2}d\rho }}$$

.

According to the claimed method, at the second stage in each channel, to increase the reliability of determining the emergency electric discharge, high-frequency components are isolated, excluding slowly changing signal components for several periods of analysis, and their absolute values are taken, calculated by the formula:

; k∈{1,2} $$y_k\left(t\right)=|\; |\; |x_k\left(t\right)-\frac{\mathrm{one}}{rT}{\displaystyle \underset{t-T}{\overset{t}{\int }}{x}_k\left(\rho \right)d\rho }|\; |\; |$$

; k∈ {1,2}

where r is the number of analysis periods over which averaging occurs.

k = 1 - for the first channel, k = 2 - for the second channel.

According to the method in the third stage, the signals received at the outputs of the second stage in each channel are filtered using amplitude filters and summed with the signals filtered for the previous periods of analysis, filtered using low-frequency filters, according to the formula:

$$\{\begin{array}{c}{z}_k\left(t\right)={\displaystyle \underset{0}{\overset{t}{\int }}{z}_k\left(\theta \right)\exp \left(a_k\left(\theta -t\right)\right)d\theta +\sqrt{2e}\sqrt{b_k}\frac{y_k^2\left(t\right)}{A}\exp \left[-b_k{\left(\frac{y_k\left(t\right)}{A}\right)}^2\right],t\ge 0}\\ z\left(t\right)=0t<0\end{array}$$

k∈ {1,2}

Where $$A=\frac{\mathrm{one}}{\tau }{\displaystyle \underset{t-\tau }{\overset{t}{\int }}{y}_k\left(\theta \right)d\theta }$$

a _k - parameter of calm, characterizes the degree of influence of previous values on the current;

b _k - suppression parameter, characterizes the degree of influence of single pulses of large amplitude;

k = 1 - for the first channel, k = 2 - for the second channel.

The essence of the method is to extract from the registered signal a current proportional to the load current of the monitored network, two components — components that have different properties and characteristics, and an analysis of the dynamics of the characteristics of these components over time. One of the components that we defined as regular is the deterministic component of the recorded signal. The second component, defined as irregular, is the stochastic component of the registered signal, which is a random or, in the case of devices with single or periodically repeating electrical discharges, a pseudorandom value.

The proposed method is illustrated in figure 1-2.

Figure 1 shows a structural diagram illustrating the method.

Figure 2 shows the area of regular and emergency electrical discharges in accordance with the change in the values of z _k .

The load current signal of the monitored network, registered by the current sensor 1, connected to one of its conductors, is supplied to two parallel channels of the processing unit, in each of which the registered signal passes through three processing stages (X, Y and Z).

In the first channel in the first stage (stage X in figure 1), the regular component is extracted from the registered signal as follows: the periodic component of the signal is eliminated by summing the input signal with the inverted signal of the previous analysis periods (delay line 2, inverter 3) and averaging over a small period of time preceding the moment the registered signal arrives in the channel for processing using the averaging element (delay line 4 and adder divider 5), and then determine yayut rms regular component for the time interval T using the RMS detector 6.

At the second stage (stage Y in FIG. 1), the obtained rms value of the converted signal is summed with the arithmetic mean of the converted signals inverted by the inverter 15 for the r previous periods of analysis T obtained using the delay line 11 and the adder-divider 12.

Thus, in fact, the repeated fluctuations of the signal value are eliminated, in the AC network corresponding to the frequency of the main harmonic of the network, or in the DC network - the switching frequency of switching power supplies or frequency converters, and the allocation of high-frequency components from the signal for further processing. At the final stage of the second stage, the absolute value of the signal is taken using a bridge rectifier 7.

Next, the signal enters the third stage (Z), where it is filtered using an amplitude filter 8, which reduces the sensitivity to single pulses of large amplitude, significantly exceeding (10 or more times) the average amplitude of the converted signal. After that, the signal is summed with the levels of the output signal accumulated over the previous time and obtained at the output of the low-pass filter included in the positive feedback circuit and formed by the delay line 20 and the integrating link 19, which ensures the influence of the history of previous signal values on the formation of the current output signal level .

In the second channel, at the first stage (stage X in Fig. 1), the irregular component is extracted from the registered signal: the signal is summed with the signal of the first channel inverted with the inverter 9, so that the extracted regular component is excluded from the original signal, after which the rms value of the irregular components over period T using a rms detector 10.

At the second stage (Y), the processing of the converted signal received at the previous stage (stage X) occurs similarly to the processing at the second stage for the first channel (elements are used: delay line 13, adder-divider 14, inverter 16, bridge rectifier 17.

The transformations in the third stage (Z) of the second channel are carried out similarly to the transformations of the first channel using an amplitude filter 18 and a low-pass filter in the feedback circuit formed by the delay line 22 and the integrating link 21.

After passing through the third stage of processing, the output signals of the first and second channels are fed to comparators 9 and 23, respectively, by which signal levels are compared with predetermined threshold signals α and β, which are supplied from reference voltage sources 25 and 26. Logical signals from comparators are fed to the device logical multiplication And, from the output of the latter, a discrete signal is supplied to the control device 24. In case of simultaneous exceeding of both threshold values, the control device based on the received of it a logical TRUE signal generates a command to disable the controlled network sends an alarm to the remote controller, etc.

Figure 2 shows the relationship of the threshold values α and β with the modes of operation of the monitored network, z ₁ and z ₂ are the output signals of the first and second channels received at the input of the comparators 9 and 23. An increase in the values of z ₁ and z ₂ signals the transition of the operating mode networks from the non-discharge mode or with the presence of a regular electric discharge to the emergency electric discharge mode. There is a transitional region - the zone of uncertainty, in which the device can operate in the absence of a pronounced emergency discharge. However, such trips cannot be considered false, as they are caused by modes close to emergency (excessive wear of the brushes of the collector motors, burnout of the contact pads of switching devices, etc.)

This method can be implemented using a device, the prototype of which is described in patent application US 2009/0059449 A1. The prototype, for the reasons stated earlier in the description of the known method, does not provide the necessary reliability of the detection of an emergency electric discharge, and the device has false positives.

The proposed device for dynamic detection of emergency electric discharge contains a load current sensor of a controlled network with an input protection device connected to the output, the signal from which is fed to a broadband amplifier, from the output of which it is fed to an analog-to-digital converter, using which a recorded signal is read out discretely in time i (t) of the load current of the controlled network, the analog-to-digital converter is connected to a functional module - digital logic device which is prepared by using the M samples i _m in the interval of time T _j, set further from the obtained set of subtracted samples elementwise samples obtained during the previous time interval T _j-1:

, $$i_{m\mathrm{,one}}=\{\begin{array}{c}{i}_m-i_{m-M},m\ge M+\mathrm{one}\\ i_m,m\le M\end{array}$$

,

where i _m is the m-th element of the set of reports obtained from the sampled signal of the load current, read over the time interval T _j ;

M is the number of samples obtained over a period of time T _j ;

the obtained sets of samples are averaged using the averaging operator

, $$i_{m\mathrm{,\; 2}}=\frac{\mathrm{one}}{m-\max \left\{m-S\mathrm{,one}\right\}+\mathrm{one}}{\displaystyle \sum _{j=\max \left\{m-S\mathrm{,one}\right\}}^m}{i}_{j\mathrm{,\; 2}}$$

,

where S is a fixed number of samples falling into the radius of a small neighborhood of τ;

and calculate the norm of the result by the formula:

; $$x_{n\mathrm{,one}}=\sqrt{\frac{\mathrm{one}}{M+\mathrm{one}}{\displaystyle \sum _{j=m-M}^m}{|\; |\; |i_{j\mathrm{,\; 2}}|\; |\; |}^2}$$

;

to select high-frequency components, the averaging operator is applied to the initial set of samples i _m according to the formula:

, $$i_{m\mathrm{,\; 3}}=\frac{\mathrm{one}}{m-\max \left\{m-S\mathrm{,one}\right\}+\mathrm{one}}{\displaystyle \sum _{j=\max \left\{m-S\mathrm{,one}\right\}}^m}{i}_j$$

,

To increase the reliability of the estimates, slowly varying components of the sequence of samples of several analysis periods are excluded: from the initial set of samples i _m, i _{m, 3 is} subtracted elementwise, and then the norm of the obtained set of numbers is calculated:

; $$x_{n\mathrm{,\; 2}}=\sqrt{\frac{\mathrm{one}}{M+\mathrm{one}}{\displaystyle \sum _{j=m-M}^m}{|\; |\; |i_j-i_{j\mathrm{,\; 3}}|\; |\; |}^2}$$

;

then from the obtained values x _{n, 1} , x _{n, 2, the} arithmetic mean values of the sets of samples x _{m, k} obtained for the last r time intervals T _j are subtracted and their absolute values are taken:

^, k∈{1,2}; $$y_{n,k}=\{\begin{array}{c}|\; |\; |x_{n,k}-\frac{\mathrm{one}}{r+\mathrm{one}}{\displaystyle \sum _{m=n-r}^n}{x}_{m,k}|\; |\; |,n\ge r\\ |\; |\; |x_{n,k}|\; |\; |,n<r\end{array}$$

^, k∈ {1,2};

then, applying to each of the sequences of values of y _{n, 1} , y _{n, 2 a} transformation equivalent to a two-stage filtering scheme consisting of a nonlinear filter for suppressing isolated pulses and a linear accumulation filter:

$$z_{n,k}=\{\begin{array}{c}{\displaystyle \sum _{m=n-p}^{n-\mathrm{one}}}{z}_{m,k}\exp \left(a_k\left(m-n+\mathrm{one}\right)\right)+\sqrt{2e}\sqrt{b_k}\frac{y_{n,k}^2}{B}\exp \left[-b_k{\left(\frac{y_{n,k}}{B}\right)}^2\right],n\ge \max \left\{p,q\right\}\\ y_{n,k},n<\max \left\{p,q\right\}\end{array}$$

k∈ {1,2}

get the values of z _{n, 1} and z _{n, 2} and compare them with the given threshold values α and β, i.e. check the fulfillment of the condition:

(z _{n, 1} > α) & (z _{n, 2} > β),

if this condition is met, a signal is generated to turn off the electrical installation, which is fed from the output of the control and analysis unit to an indicator, the output of which is connected to a relay, which actuates the actuator acting on the contact group, by means of which the controlled network is de-energized.

The device can be structurally and functionally combined with a residual current device, and / or with a circuit breaker for protection against overloads and overcurrents.

The proposed device is illustrated by the drawings of figures 3-4.

Figure 3 shows a block diagram of a device.

Figure 4 - algorithm of the functional module of the device.

The device (figure 3) contains a contact group 2, a surge protection unit 3, a load current sensor 4, a power supply 7, an input protection device 9, an amplifier 11, an analog-to-digital converter 12, a functional module 14, an indicator 13, a relay 10 , an actuator 8 and, if combined with a protection device, thermal and electromagnetic releases, and a differential current sensor 6.

The device is connected to a controlled network in series with loads and operates as follows.

The mains voltage is supplied to the input terminals 1, which then goes to the contact group 2, used to disconnect the load in case of emergency, the surge protection unit 3, which prevents the passage of short-term surge surges to a controlled electrical installation and to the electronic circuit of the device, to the load current sensor 4 and to the output terminals 5.

The signal from the load current sensor 4 is fed to the input protection device 9, which provides additional protection for the electronic circuitry of the device during prolonged overload (if the device does not include a circuit breaker or if it is damaged). Next, the signal is amplified by an amplifier 11 and fed to an analog-to-digital converter 12 for digitization and further processing using the functional module 14.

Functional module 14 contains a microcontroller or other digital logic device with additional elements that provide the desired mode of operation. With its help, they collect, store and process the digitized data of the current signal received from the sensor, generate a signal to the indicator 13 and commands to the relay 10 to actuate the actuator 8 in order to open the contact group 2 and disconnect the controlled network when an emergency occurs.

Functional module 14 processes the incoming signals and issues information messages or generates a command to shut down the monitored network according to the algorithm shown in Fig. 4.

As a result of the implementation of this invention, the reliability of determining an emergency electric discharge is significantly increased, false alarms are excluded while various types of loads are operating (purely active, inductive, capacitive and mixed). Separately, it should be pointed out that the device is operating at a high speed and that an emergency situation is recorded steadily before igniting a stable electric arc leading to a fire. This invention can be applied in networks of both alternating and direct current.

Thus, the introduction of the proposed method and device for the dynamic detection of emergency electrical discharge provides a fire warning in case of a malfunction in the electric network, increases the degree of protection of people, residential, industrial and other objects from the damaging effects of fires.

Claims (8)

1. A method for dynamically detecting an emergency electric discharge, characterized in that the registered signal i (t) of the load current of the controlled network is supplied to two channels, each of which undergoes three processing steps, while: in the first channel, the deviations of the registered signal are evaluated at the first stage of the current from the periodic function, the regular component, which is the deterministic part of the original signal, is isolated by averaging the initial signal over a small period of time - more than ten times m nshim analysis period T, and take its rms value; at the second stage, to increase the reliability of detection of an emergency electrical discharge, high-frequency components are isolated, excluding slowly changing signal components over several analysis periods T; at the third stage, the signal is filtered to exclude the influence of individual pulses of large amplitude and take into account the history of its previous values over a long time period with a decrease in the contribution of each value to the result depending on its prescription; in the second channel, at the first stage, an irregular component is selected, which is the stochastic part of the original signal, by subtracting the regular component from the registered signal i (t) and taking its rms value; at the second stage, high-frequency components are distinguished, excluding slowly changing signal deviations for several periods of analysis; to the third degree, the signal is filtered to exclude the influence of individual pulses of large amplitude and take into account the history of its previous values over a long time period with a decrease in the contribution of each value to the result depending on its prescription, then the signal levels obtained at the outputs of the first and second channels are compared with predetermined threshold values and if they are exceeded, an alarm command is generated, leading to a blackout of the controlled network. 2. The method for dynamically detecting an emergency electrical discharge according to claim 1, wherein in the first stage, to determine the deviation of the registered current signal from the periodic function, the values for the previous analysis period are subtracted from the signal values for the current analysis period and the rms value x ₁ (t) is calculated, which is an estimate of the aperiodicity of the current signal, according to the formula:
$$x_{\mathrm{one}}\left(t\right)=\frac{\mathrm{one}}{2\tau \sqrt{T}}\sqrt{{\displaystyle \underset{t-T}{\overset{t}{\int }}{\left({\displaystyle \underset{\rho -\tau }{\overset{\rho +\tau }{\int }}\left(i\left(\theta \right)-i\left(\theta -T\right)d\theta \right)}\right)}^{2}d\rho }}$$

,
where i (t) is the registered current signal;
T - analysis period - a fixed time period for alternating current and direct current, determined by the response time of protective devices: circuit breaker or residual current circuit breaker installed together and providing additional protection for the electrical installation;
t is the moment of arrival of the registered signal in the channels for processing;
τ is the radius of a small neighborhood of t;
ρ - time instants lying on the interval [tT; t] over which averaging occurs;
θ is the moment of signal conversion. 3. The method for dynamically detecting an emergency electrical discharge according to claim 1, in which, to isolate the irregular component in the first stage, the values of the regular component are subtracted from the values of the registered current signal, then the rms value x ₂ (t) is calculated by the formula:
$$x_2\left(t\right)=\frac{\mathrm{one}}{\sqrt{T}}\sqrt{{\displaystyle \underset{t-T}{\overset{t}{\int }}{\left(i\left(\rho \right)-\frac{\mathrm{one}}{2\tau }{\displaystyle \underset{\rho -\tau }{\overset{\rho +\tau }{\int }}i\left(\theta \right)d\theta }\right)}^{2}d\rho }}$$

4. In the method for dynamically detecting an emergency electric discharge according to claim 1, in the second stage in each channel, to increase the reliability of determining the emergency electric discharge, high-frequency components are isolated, excluding slowly changing signal components for several periods of analysis, and their absolute values are taken, calculated by the formula :
$$y_k\left(t\right)=|\; |\; |x_k\left(t\right)-\frac{\mathrm{one}}{rT}{\displaystyle \underset{t-T}{\overset{t}{\int }}{x}_k\left(\rho \right)d\rho }|\; |\; |$$

; k∈ {1,2},
where r is the number of analysis periods over which averaging occurs;
k = 1 - for the first channel, k = 2 - for the second channel. 5. In the method for dynamically detecting an emergency electric discharge according to claim 1 in the third stage, the signals received at the outputs of the second stage in each channel are filtered using amplitude filters and summed with the signals obtained from the previous analysis periods filtered using low-pass filters , according to the formula:
$$\{\begin{array}{c}{z}_k\left(t\right)={\displaystyle \underset{0}{\overset{t}{\int }}{z}_k\left(\theta \right)\exp \left(a_k\left(\theta -t\right)\right)d\theta +\sqrt{2e}\sqrt{b_k}\frac{y_k^2\left(t\right)}{A}\exp \left[-b_k{\left(\frac{y_k\left(t\right)}{A}\right)}^2\right],t\ge 0}\\ z\left(t\right)=0t<0\end{array}$$

, k∈ {1,2},
Where $$A=\frac{\mathrm{one}}{\tau }{\displaystyle \underset{t-\tau }{\overset{t}{\int }}{y}_k\left(\theta \right)d\theta }$$

;
a _k - parameter of calm, characterizes the degree of influence of previous values on the current;
b _k - suppression parameter, characterizes the degree of influence of single pulses of large amplitude;
k = 1 - for the first channel, k = 2 - for the second channel. 6. A device for dynamic detection of emergency electrical discharge, containing a load current sensor of a controlled network with an input protection device connected to the output, the signal from which is fed to a broadband amplifier, from the output of which it is fed to an analog-to-digital converter, using which a time-discrete reading is performed the registered signal i (t) of the load current of the monitored network, the analog-to-digital converter is connected to a functional module - a digital logic device count, by which is obtained i _m M samples in the time interval T _j, set further from the obtained set of subtracted samples elementwise samples obtained during the previous time interval T _j-1:
$$i_{m\mathrm{,one}}=\{\begin{array}{c}{i}_m-i_{m-M},m\ge M+\mathrm{one}\\ i_m,m\le M\end{array}$$

,
where i _m is the m-th element of the set of reports obtained from the sampled signal of the load current, read over the time interval T _j ;
M is the number of samples obtained over a period of time T _j ;
the obtained sets of samples are averaged using the averaging operator
$$i_{m\mathrm{,\; 2}}=\frac{\mathrm{one}}{m-\max \left\{m-S\mathrm{,one}\right\}+\mathrm{one}}{\displaystyle \sum _{j=\max \left\{m-S\mathrm{,one}\right\}}^m}{i}_{j\mathrm{,\; 2}}$$

,
where S is a fixed number of samples falling into the radius of a small neighborhood of τ;
and calculate the norm of the result by the formula:
$$x_{n\mathrm{,one}}=\sqrt{\frac{\mathrm{one}}{M+\mathrm{one}}{\displaystyle \sum _{j=m-M}^m}{|\; |\; |i_{j\mathrm{,\; 2}}|\; |\; |}^2}$$

;
to select high-frequency components, the averaging operator is applied to the initial set of samples i _m according to the formula:
$$i_{m\mathrm{,\; 3}}=\frac{\mathrm{one}}{m-\max \left\{m-S\mathrm{,one}\right\}+\mathrm{one}}{\displaystyle \sum _{j=\max \left\{m-S\mathrm{,one}\right\}}^m}{i}_j$$

,
To increase the reliability of the estimates, slowly varying components of the sequence of samples of several analysis periods are excluded: from the initial set of samples i _m, i _{m, 3 is} subtracted elementwise, and then the norm of the obtained set of numbers is calculated:
$$x_{n\mathrm{,\; 2}}=\sqrt{\frac{\mathrm{one}}{M+\mathrm{one}}{\displaystyle \sum _{j=m-M}^m}{|\; |\; |i_j-i_{j\mathrm{,\; 3}}|\; |\; |}^2}$$

;
then from the obtained values x _{n, 1} , x _{n, 2, the} arithmetic mean values of the sets of samples x _{m, k} obtained for the last r time intervals T _j are subtracted and their absolute values are taken:
$$y_{n,k}=\{\begin{array}{c}|\; |\; |x_{n,k}-\frac{\mathrm{one}}{r+\mathrm{one}}{\displaystyle \sum _{m=n-r}^n}{x}_{m,k}|\; |\; |,n\ge r\\ |\; |\; |x_{n,k}|\; |\; |,n<r\end{array}$$

^, k∈ {1,2};
then, applying to each of the sequences of values of y _{n, 1} , y _{n, 2 a} transformation equivalent to a two-stage filtering scheme consisting of a nonlinear filter for suppressing isolated pulses and a linear accumulation filter:
$$z_{n,k}=\{\begin{array}{c}{\displaystyle \sum _{m=n-p}^{n-\mathrm{one}}}{z}_{m,k}\exp \left(a_k\left(m-n+\mathrm{one}\right)\right)+\sqrt{2e}\sqrt{b_k}\frac{y_{n,k}^2}{B}\exp \left[-b_k{\left(\frac{y_{n,k}}{B}\right)}^2\right],n\ge \max \left\{p,q\right\}\\ y_{n,k},n<\max \left\{p,q\right\}\end{array}$$

k∈ {1,2};
Where $$B=\frac{\mathrm{one}}{q}{\displaystyle \sum _{m=n-q}^{n-\mathrm{one}}}{y}_{m,k}$$

;
get the values of z _{n, 1} and z _{n, 2} and compare them with the given threshold values α and β, i.e. check the fulfillment of the condition:
(z _{n, 1} > α) & (z _{n, 2} > β),
if this condition is met, a signal is generated to turn off the electrical installation, which is fed from the output of the control and analysis unit to an indicator, the output of which is connected to a relay, which actuates the actuator acting on the contact group, by means of which the controlled network is de-energized. 7. The device for dynamic detection of emergency electrical discharge according to claim 6 can be structurally and functionally combined with a residual current device. 8. The device for dynamic detection of emergency electrical discharge according to claim 6 can be structurally and functionally combined with an overload and overcurrent protection circuit breaker.

Images