The present invention relates to the field of automatic control and can be used in systems with excessive number of sensors, for example accelerometers, the failure of which should result in a control system failure.

There is a method of determining a failed sensor, implemented in [1], including the measurement of each parameter probe movement and verification of the connection is measured by the sensors of the motion parameters characterizing the proper operation of the sensors.

The disadvantage of this method is that it assumes the presence of a large number of redundant sensors (beyond three, not located in the same plane, and completely define the parameter vector movement) and does not identify the failed sensor in any number of redundant sensors, for example with one.

The closest technical solution to the proposed method is the method implemented in [2], including the measurement of each parameter probe motion and periodic with period T₀check connection measured by each sensor motion parameters characterizing the proper operation of the sensors.

The disadvantage of this method is that it assumes the existence of two identical redundant single redundant sensor that allows you to define failed on tcic. However, this method does not allow to identify the failed sensor in a single redundant single redundant sensor.

The objective of the invention is the extension of functionality through the ability to identify the failed sensor in any redundant system.

This task is achieved by a method for determining a failed sensor in a redundant system, including the measurement of each parameter probe motion, periodic with period T₀check the connection of the measured motion parameters characterizing the proper operation of the sensors, assumes at time t_i=i·T₀where i=1, 2, 3,... violations of the ratio of communication of the measured motion parameters, comparisons of readings and_ieach sensor at time t_itestimony and_(i-1)the same sensor at time T_(i-1)=(i-1)·T₀by finding the difference between these two readings Δa_i=(a_i-a_(i-1)) and identification of the failed sensor, for which the difference Δa_iexceeds the specified value.

The drawing shows a coordinate diagram illustrating the location of the measurement axes of the sensors and the direction of the measured parameter. In this diagram: x, y, z, v - direction of the measurement axes of the sensors and the direction of the measured parameter, α, β, γ, ν - angles is between direction and respectively measuring axes x, y, z, v, ϕ - the angle between the direction of the measuring axis v and measuring axes x, y, z.

Implementation of the proposed method of determination of the failed sensor consider the example of a system with a single redundant sensor - accelerometer when controlling motion of a rigid body (spacecraft). In the General case, the system of three accelerometers, measuring axis which do not lie in one plane, completely determines the vector of the linear acceleration of the spacecraft. A redundant system contains one sensor is more than four accelerometer, while any three of the four sensor does not lie in the same plane. Failure of one of the four sensors in such a system there is the possibility of replacing this sensor, using the testimony of other functioning of the three sensors.

Let a redundant system contains four sensors (drawing), measuring axis which is directed along the coordinate axes x, y, z, v, x, y, z are mutually perpendicular, and the v axis is directed along the diagonal of the cube, the edges of which coincide with the axes x, y, z. Let us also suppose that the vector of the measured parameter (acceleration) and is with the axes x, y, z, v, respectively, angles α, β, γ, νand the angle between the axis v and x, y, z is equal to ϕ. The sensors each measuring axis are respectively equal to

In this case, the vector a can be represented in the form

Design vector and the v axis. The result will be

Now we introduce the ratio of communication of motion parameters in the form (4)

As follows from (1) and (3), the value of R is equal to zero (close to zero), if all four sensors intact. In case of sensor failure, the value of R is not equal to zero (greater than a certain value), indicating a malfunction in the system of the four sensors.

Carry out periodic with period T₀check the connection parameters of motion (4) equals zero (belonging to the given interval) values of R. let us Choose T₀therefore, for the change of the measured parameter (acceleration) was insignificant and did not exceed a certain value Δ. Suppose at some time T_i=i·T₀(i=1, 2,3...) violation occurred ratio relation (4), wherein the value of R is outside of the permitted values. Fix the sensor readings and_iat time t_iand the sensor readings and_(i-1)at time t_(i-1)=(i-1)·T₀. Find the difference Δa_ias

Before time T_ivalue Δasub> ifor each of the sensors was less than the specified value Δ. At time t_ifor the functioning of the sensors, this value will be less than the specified value Δ. For the failed sensor is Δa_iwill go beyond the permitted values Δthat is a criterion for determining the failed sensor. After identifying the failed sensor replacement can be performed by determining the ratio (4) his testimony, using the testimony of other three functioning sensors.

Thus, the way to identify the failed sensor in a redundant system allows you to identify the failed sensor in any number of redundant sensors that extends the functionality of the control system.

The proposed set of features considered by the authors in decisions not met for solving the task and not obvious from the prior art, which allows to conclude that the technical solutions according to the criteria of “novelty” and “inventive step”. The implementation of the method of determining the failed sensor means conducting standard transactions commit parameters, compared with the specified value and by comparing the results of the determination of the failed sensor.

Literature.

1. Authorship SS is R N 613291 7. 03. 78, class G 05 B 15/02, G 05 D 1/02.

2. USSR author's certificate N 489078 30. 06. 75, CL G 05 B 15/02, G 05 D 1/02.

The method of determining the failed sensor in a redundant system, including the measurement of each parameter probe motion, periodic with period T₀selected so that the change of the measured parameter does not exceed the specified value Δ, check the connection of the measured motion parameters describing proper operation of sensors, characterized in that at time T_i=i·T₀where i = 1, 2, 3, ....., violations of the ratio of communication of the measured motion parameters to produce a comparison of the readings and_ieach sensor at time T_iwith the testimony of a_(i-1)the same sensor at time t_(i-1)=(i-1)·T₀by finding the difference between these two readings Δa_i=(a_i-a_(i-1)) and identification of the failed sensor, for which the difference Δa_iexceeds the specified value.