RU2364926C2 - Way of control of multiparameter object state - Google Patents

Abstract

FIELD: physics; measurement.

SUBSTANCE: invention concerns area of measuring equipment and can be used for the control and the analysis of a state of the difficult multiparameter objects which are units of communication systems and automation. It is reached by the job of not correlated parametres for the control, grouping of structural units of the multiparameter object on a degree of closeness of values of controllable parametres, revealing of statistical dependence between separate controllable parametres and predictions of the tendency of their change.

EFFECT: increase of reliability, accuracy and efficiency of estimation of a state of the multiparameter object in the course of its functioning.

2 cl, 3 dwg

Description

The invention relates to measuring technique and can be used to monitor and analyze the state of complex multi-parameter objects that are elements of communication and automation systems.

The known method of continuous passive control of telephone line parameters, implemented in a device of the type LST-1007 [Lysov A.V. Phone and security. Problems of information security in telephone networks / A.V. Lysov, A.N. Ostapenko. - St. Petersburg: Polytechnic, 1995. - 109 p. (PPSh Laboratory)], in which the telephone line is preliminarily checked for damage and unauthorized connections, the controlled parameters of the telephone line are set, the controlled parameters are measured, stored as standards, the current values of the controlled parameters are continuously measured and compared with the reference ones, and an alarm is generated when mismatch of the measured values of the parameters with the reference, continue to measure when they coincide with the reference values of the monitored parameters.

The disadvantages of this method are the relatively low accuracy of the control results due to external factors (climatic, temporary, etc.), which accounts for the low probability of making an objective decision on the state of the controlled object and the low economic efficiency of the control system, which limits the scope of the method.

Also known is the "Method of measuring the parameters of RLC circuits", according to the patent of Russian Federation No. 2100813, class G01R 27/26, 1997.12.27, based on measuring the duration of transients in a resistive-capacitive or resistive-inductive measuring circuit, to the input of which a disturbing voltage is applied , varying as a function of time, and its parameters are measured at the output of the RLC circuit under study.

This method in comparison with the previous analogue allows you to get more complete information about the characteristics of the controlled object. The disadvantages of this method are the relative complexity of the implementation, due to the need to generate special sounding signals, and a narrow scope, due to the incompatibility of the measurement method with the normal operation of telephone lines and the inability to continuously measure their parameters.

Closest in technical essence to the claimed one is a method for evaluating the effectiveness of large systems, including a large number of monitored parameters, according to RF patent No. 22010112 "Unified Chernyakov / Petrushin method for evaluating the effectiveness of large systems", class G06F 17/00, declared. June 7, 2001.

The prototype method consists in the fact that a pre-multiparameter object (MPO) is set in the form of a hierarchy of its structural elements of the same type (OSE); set the particular characteristics of the state, put in correspondence with each OSE, the normative values corresponding to each particular indicator of the state, the weight coefficients of importance, corresponding to each particular indicator of the state of the OSE; and in advance, a program for calculating private parameters is recorded in the memory of the terminal server and, finally, the information obtained during the survey of experts in this field of knowledge is preliminarily loaded into the memory of the workstation, the calculation method is selected and this procedure is started, and the measured parameters, automatically read information from sensors through converters and write it to the memory of read information in a terminal server, convert the parameter values into corresponding digital data using various special converters, store digital data in a storage device, calculate the partial and generalized characteristics of the state of the MPO according to the program for calculating the state characteristics using a terminal server, compare them with preset values, display and document the results of calculations and comparisons on a video monitor and printer.

Compared with the analogues discussed above, the prototype method has a wider scope for both simple and multiparameter objects.

The disadvantage of the prototype is the relatively low accuracy and efficiency of assessing the state of MPO. This is due to the fact that the state of the MPO is determined by the values of the state indicators, the calculation of which uses predefined constant weighting factors. In addition, the control results indicate only critical values of the parameters and do not indicate their relationship with other parameters.

The aim of the proposed technical solution is to develop a method for monitoring the state of MPO, which provides higher accuracy, reliability and efficiency of the current assessment of the state of MPO due to preliminary grouping of OSE by the degree of proximity of the values of the controlled parameters, taking into account the statistical dependence between the individual controlled parameters and predicting the trend of their change.

The claimed method extends the arsenal of funds for this purpose.

измеряемых параметров, допустимые отклонения

значений параметров от стандартных значений, допустимые разности

между максимальным

и минимальным

уровнями измеренного m-го параметра, где m=1,2,…,М, i-го и j-го однотипных структурных элементов, допустимое значение показателя корреляции Р^доп между измеряемыми параметрами, принадлежащими i-му и j-му однотипным структурным элементам. Для группирования ОСЭ измеряют значения П_mi измеряемых параметров, вычисляют коэффициенты сходства λ_ij однотипных структурных элементов по всем m параметрам, где i=1,2,…N и j=1,2,…N, и i≠j, сравнивают вычисленные значения λ_ij с предварительно заданным значением λ^пор и, по результатам сравнения, запоминают в соответствующих областях памяти запоминающего устройства идентификаторы группируемых однотипных структурных элементов, принадлежащих g-той группе, разность величин коэффициентов сходства которых удовлетворяет заданную величину λ^пор. После чего, в процессе функционирования объекта измеряют с периодом Т_u значения П_mi измеряемых параметров. Значения вычисляемых параметров определяют для каждой g-той группы однотипных структурных элементов, где g=1,2,…,G. Для этого вычисляют средние по g-той группе значения измеряемых параметров

и их отклонения

от стандартных значений

, разности

, промежутки времени τ_mg, по истечении которых значения

могут превысить допустимые значения

. Затем, для структурных элементов, имеющих максимальное

и минимальное

значения m-го параметра в каждой g-той группе, вычисляют значения показателей корреляции P_i и P_j между некоррелированными параметрами i-го и j-го однотипных структурных элементов по формуле:This goal is achieved by the fact that in the known method for monitoring the state of a multi-parameter object, which consists in pre-setting the set N≥2 of its structural elements of the same type and K≥N measured parameters characterizing the state of the structural elements of the same type, in the process of functioning of the multi-parameter object, they measure and store the values of the measured parameters of the same type of structural elements, determine the values of the calculated parameters of the same type of structural elements, document dissolved them and take them to a decision state multiparameter object further pre-set G memory areas in the memory, wherein 1≤G≤N, for storing identifiers I _{g g-one} group of the same structural elements, where g = 1,2, ... , G. The threshold value of the similarity coefficient λ of the ^{pores of} grouped similar structural elements, the set M∈K of uncorrelated measured parameters characterizing the state of each of the same type of structural elements, the measurement period T _u and standard values are also set

measured parameters, tolerances

parameter values from standard values, allowable differences

between the maximum

and minimal

levels of the measured m-th parameter, where m = 1,2, ..., M, of the i-th and j-th structural elements of the same type, the allowable value of the correlation index P ^add between the measured parameters belonging to the i-th and j-th structural elements of the same . To group the SSE, the values of P _{mi of} measured parameters are measured, the similarity coefficients λ _{ij of the} same structural elements are calculated for all m parameters, where i = 1,2, ... N and j = 1,2, ... N, and i ≠ j, the calculated values are compared λ _ij with a predetermined value of λ ^pores and, according to the results of comparison, remember the identifiers of grouped structural elements of the same type belonging to the gth group in the corresponding memory areas of the memory device, the difference in the values of the similarity coefficients of which satisfies the given value of λ ^pores . Then, in the process of functioning of the object is measured with a period T _u values P _mi measured parameters. The values of the calculated parameters are determined for each gth group of the same structural elements, where g = 1,2, ..., G. For this, the values of the measured parameters averaged over the gth group are calculated

and their deviations

from standard values

differences

, time intervals τ _mg , after which the values

may exceed acceptable values

. Then, for structural elements having maximum

and minimum

the values of the m-th parameter in each g-th group, the values of the correlation indices P _i and P _j between the uncorrelated parameters of the i-th and j-th structural elements of the same type are calculated by the formula:

- множественное корреляционное отношение k-го параметра со всеми остальными параметрами i-го и j-го ОСЭ. Корреляционное отношение η_к1 и множественное корреляционное отношение η^* _k вычисляют известными методами, описанными, например, в [Митропольский А.К. Техника статистических вычислений / А.К.Митропольский. - М.: Наука, 1971 г. - 576 с.]. Вычисленные значения

Р_ij сравнивают с предварительно заданными допустимыми значениями

Р^доп. После этого документируют значения параметров состояния структурных элементов. Если выполняется условие

и/или

и/или P_i>Р^доп, и/или P_j>Р^доп, то принимают решение о неисправном состоянии многопараметрического объекта и дополнительно документируют информацию о параметрах структурных элементов, для которых

и/или

и/или Р_i>Р^доп, и/или Р_j>Р^доп. Если же выполняется условие

и

и Р_i≤Р^доп, и Р_j≤Р^доп, то принимают решение об исправном состоянии многопараметрического объекта и продолжают периодические измерения и вычисления значений параметров структурных элементов.where η _kl is the correlation ratio of the kth and lth measured parameters of the i-th and j-th OSE, where k = 1,2, ..., M and l = 1,2, ..., M, k ≠ l;

- the multiple correlation ratio of the k-th parameter with all other parameters of the i-th and j-th OSE. The correlation ratio η to _k1 and the multiple correlation ratio η ^* _k are calculated by known methods described, for example, in [Mitropolsky A.K. The technique of statistical computing / A.K. Mitropolsky. - M .: Nauka, 1971 - 576 p.]. The calculated values

P _ij compared with predefined acceptable values

R ^add . After that, document the values of the state parameters of the structural elements. If the condition is met

and / or

and / or P _i > P ^add , and / or P _j > P ^add , then make a decision about the malfunctioning state of the multi-parameter object and additionally document information about the parameters of structural elements for which

and / or

and / or P _i > P ^add , and / or P _j > P ^add . If the condition is satisfied

and

and P _i ≤P ^add , and P _j ≤P ^add , then decide on the working condition of the multi-parameter object and continue periodic measurements and calculations of the parameters of the structural elements.

Due to the new set of essential features in the method, it is possible to detect changes in the controlled parameters in an arbitrarily small (limited only by the capabilities of the measuring instruments) interval of permissible values, detect an abnormal correlation between predefined non-correlated controlled parameters, minimize the influence of external factors on the control results by calculating the correlation indicators R _ij , as well as forecasting the trend of change in the controlled parameter s, what is achieved in the claimed method to improve the accuracy, efficiency and reliability of the current assessment of the state of MPO.

The used correlation index allows you to take into account the correlation background of the controlled parameters, due to the systemic influence of external factors. Due to this, the actual statistical relationship between the parameters is more accurately reflected. Identification of the parameters that have the highest values of the correlation indicators provides an increase in the efficiency of identifying the nature and cause of the MPO malfunction.

The use of forecasting the trend of changes in the controlled parameters of the OSE state provides an increase in the economic efficiency of the control system.

The claimed method is illustrated by drawings, which show:

Figure 1 is a drawing illustrating a multi-parameter object and its structural elements of the same type, grouped by the degree of proximity of the values of the measured parameters;

Figure 2 - figures explaining the process of grouping the OSE, assessing and predicting the trend of changes in the parameters of the OSE groups;

Figure 3 is a drawing illustrating the achievement of a positive effect in the implementation of the claimed control method.

The efficiency and accuracy of assessing the state of MPO is determined by the size and composition of the set of controlled parameters, methods of their measurement and processing of measurement results.

The values of highly informative controlled parameters can be strongly influenced by the conditions of their measurement. In this regard, taking into account the influence of external factors on the controlled measured parameters significantly increases the accuracy of their measurement and reduces the likelihood of errors in determining the state of MPO.

The implementation of the claimed method is conveniently considered on the example of MPO, which represents the controller of the fire alarm system (K) with a network of loops and sensors (of the same structural elements) (Fig. 1.). For similar structural elements, a set of uncorrelated controlled parameters {P _m } is specified, where m = 1.2, ..., M, which determine the operational characteristics of the SSE (temperature, concentration of smoke in air, etc.).

(определенные в эксплуатационно-технической документации), а также допустимые отклонения значений каждого из параметров

от нормы и допустимый размах значений m-го параметра

Контролируемые параметры ОСЭ могут принимать значения в диапазоне

(фиг.2а). Для функционирования МПО изменение значений параметров в этом диапазоне не критично, но может быть признаком неисправности или ухудшения условий функционирования ОСЭ. В связи с индивидуальными условиями эксплуатации отдельные ОСЭ в неравной степени подвержены влиянию системных и случайных внешних воздействий (температура в различных помещениях может существенно отличаться). Поэтому для одних ОСЭ некоторые значения контролируемых параметров являются нормой, а для других - критичными. Известные способы контроля не учитывают изменений контролируемых параметров в пределах допустимых значений и статистическую связь между отдельными параметрами, чем обусловлена низкая точность оценки состояния МПО. Уменьшение влияния внешних системных факторов и связанных с ними ошибок контроля в предлагаемом способе достигается путем группирования ОСЭ по степени близости значений измеряемых параметров и вычисления показателей корреляции Р_ij.For each m-th parameter from the set of M-parameters of the OSE, depending on the tasks and operating conditions of the MPO, standard (nominal) parameter values are set

(defined in the operational and technical documentation), as well as permissible deviations of the values of each parameter

from the norm and the permissible range of values of the m-th parameter

The controlled parameters of the OSE can take values in the range

(figa). For the functioning of the MPO, a change in the parameter values in this range is not critical, but it can be a sign of malfunction or deterioration of the operating conditions of the OSE. In connection with individual operating conditions, individual OSE are unequally affected by systemic and random external influences (the temperature in different rooms can vary significantly). Therefore, for some OSEs, some values of the controlled parameters are the norm, and for others, they are critical. Known methods of control do not take into account changes in the controlled parameters within acceptable values and the statistical relationship between the individual parameters, which is due to the low accuracy of assessing the state of MPO. The reduction of the influence of external systemic factors and the associated control errors in the proposed method is achieved by grouping the OSE according to the degree of proximity of the measured parameters and calculating the correlation indices P _ij .

Grouping is carried out by one of the known methods of grouping (recognition) before the start of periodic monitoring. In this case, the measured values of the OSE parameters P _m1 ... P _{mN are measured} and for each pair of OSE differences ΔP _mij are calculated for all M-parameters (Fig. 2a). The obtained results are used in calculating the values of the grouping indices λ _ij for each pair of OSE. As an indicator of grouping can be used, for example, a measure of proximity Voronin [Mandel I.D. Cluster analysis. - M .: Finance and statistics, 1988 - 176 p .; ill.], which is calculated by the formula:

where σ _m - pre-set weights of the controlled parameters.

The criterion for deciding on the proximity of the i-th and j-th OSEs is the pre-set value of the group ^pore index λ ^pores . As a result of grouping, identifiers I _g , where g = 1,2, ..., G OSE, for which the difference in the values of their similarity coefficients λ _{ij are} greater than a predetermined threshold value, are recorded in the memory area of the storage device (memory) of the monitoring system

λ ^then .

After the completion of the OSE grouping, the monitoring system periodically monitors the status of the MPO. For this, during the operation of the MPO, the values of the measured parameters P _m are measured with a period T _u and recorded in the memory arrays corresponding to each OSE group. Then calculate the average values of the measured parameters for each group of OSE, calculate the deviations of the average values with predefined standard values of the parameters. After that, the obtained deviations are compared with the permissible deviations and the results of the comparison are recorded in the corresponding arrays of the memory. The statistics accumulated in this way are used to identify, using known methods of forecasting, trends in the values of parameters of OSE groups and to calculate the estimated time interval τ _mg before the critical state of MPO. The value of the time interval τ _mg for each gth group of the same type of structural elements is calculated by known forecasting methods [Kovaleva L.N. Multifactorial forecasting based on the series of dynamics / L.N. Kovaleva. - M .: Statistics, 1980 - 102 p.]. For the simplest linear relationship, the predicted time τ _mg can be calculated, for example, by the formula:

- среднее значение приращений значений

(фиг.2б).Where

- average value of increment values

(figb).

и П_mi, вычисляют разность ΔП_mi для каждого i-го ОСЭ по каждому m-му параметру с помощью устройства сравнения (например, компаратора). Аналогично, для вычисления значений ΔП_mij считывают из ЗУ экстремальные значения

и

для каждого m-го параметра, сравнивают их и вычисляют разность ΔП_mij с помощью устройства сравнения. После этого считывают из ЗУ предварительно записанные значения

и вычисленные значения

и сравнивают их также с помощью устройства сравнения. После этого для ОСЭ с экстремальными значениями измеряемых параметров в группе вычисляют значения показателей корреляции P_i(j). Используемый показатель корреляции P_i(j) учитывает нелинейную корреляцию между отдельными контролируемыми параметрами

что повышает точность результатов контроля и позволяет учитывать корреляционный фон

обусловленный влиянием внешних факторов.To calculate the characteristics of the OSE state, pre-recorded values are read from the memory

and P _mi , calculate the difference ΔP _mi for each i-th OSE for each m-th parameter using a comparison device (for example, a comparator). Similarly, to calculate ΔP _mij values, extreme values are read from the memory

and

for each m-th parameter, compare them and calculate the difference ΔP _mij using a comparison device. After that, pre-recorded values are read from the memory.

and calculated values

and compare them also using a comparison device. After that, for OSE with extreme values of the measured parameters in the group calculate the values of the correlation indicators P _{i (j)} . The used correlation index P _{i (j)} takes into account the nonlinear correlation between the individual controlled parameters

which increases the accuracy of the control results and allows you to take into account the correlation background

due to the influence of external factors.

Thus, the calculated correlation indices P _{i (j)} are weight coefficients that take into account the individual operating conditions of the OSE when determining the most correlated pair of monitored parameters, and can improve the accuracy of monitoring the state of MPO.

Identification of the statistical relationship between the individual controlled parameters of the structural element of the object allows you to determine the nature and cause of its malfunction. For example, a statistical dependence of the change in capacitance (or insulation resistance) of a cable communication line and the change in the surrounding magnetic field, which is recorded by magnetometric detection means [Shagabutdinova L. M., can be detected. (CJSC "Vostok - Special Systems") / L.M. Shagabutdinova // OPS, fire safety. - 2005. - No. 6. - p. 116-118.]. Such a dependence may serve as a sign of unauthorized influence on the cable communication line (for example, installation of a device for secret reading of information without galvanic connection).

Р^доп. Результаты сравнения являются исходными данными для принятия решения о состоянии МПО. Если значения вычисляемых параметров не превосходят допустимых значений, то принимается решение об исправности МПО, а система контроля автоматически документирует результаты контроля и продолжает периодические измерения и вычисления значений контролируемых параметров. В случае превышения допустимых значений контролируемых параметров принимается решение о неисправном состоянии МПО, и система контроля дополнительно документирует информацию о состоянии неисправных ОСЭ.The thus obtained values of the calculated parameters: ΔP _mi , ΔP _mij , P _{ij are} recorded in the memory and compared with pre-set valid values

R ^add . The comparison results are the source data for deciding on the status of the MPO. If the values of the calculated parameters do not exceed the permissible values, then a decision is made about the health of the MPO, and the control system automatically documents the results of the control and continues to periodically measure and calculate the values of the controlled parameters. If the permissible values of the monitored parameters are exceeded, a decision is made about the faulty state of the MPO, and the control system additionally documents information about the status of the faulty OSE.

The possibility of obtaining a positive effect when using the proposed method was confirmed by comparative step-by-step modeling and comparison of the prototype method and the claimed method. The comparison results are presented in figure 3.

(см. на фиг.3, момент времени

В то же время, при использовании способа-прототипа до момента

состояние МПО ошибочно принимают исправным, а аномальность может быть обнаружена только на (ν+1)-м цикле контроля в момент времени

(см. фиг.3), т.е. при выполнении условия

В ситуации, представленной на фиг.2б, способ-прототип не позволяет обнаружить неисправность, и состояние МПО ошибочно принимают исправным, а заявленный способ позволяет выявить тенденцию изменения контролируемого параметра и после ν-го цикла контроля выявить неисправное состояние МПО. Время обнаружения в способе-прототипе превышает время, затраченное на обнаружение аналогичных отклонений при использовании заявленного способа. Благодаря упреждающему обнаружению тенденции ухудшения условий функционирования МПО возможно своевременное устранение причины неисправности и повышение оперативности обнаружения и точности оценки состояния МПО. Таким образом, оказывается возможным достижение запланированного технического результата.The step-by-step modeling technique was as follows. The time interval T _{u was set} between the νth and (ν + 1) -th cycles of monitoring the state of MPO. Next, control cycles were performed according to the prototype method and according to the claimed method until an unacceptable deviation of the controlled parameters was detected by the control system. In the situation shown in figa, the claimed method allows to detect abnormal behavior of the parameter of one of the OSE on the ν-th control cycle when the condition

(see figure 3, the point in time

At the same time, when using the prototype method until

the state of the MPO is mistakenly taken as serviceable, and the anomaly can be detected only on the (ν + 1) -th control cycle at a time

(see figure 3), i.e. under the condition

In the situation presented in fig.2b, the prototype method does not allow to detect a malfunction, and the state of the MPO is mistakenly taken as serviceable, and the claimed method allows you to identify the trend of the controlled parameter and after the ν-th control cycle to identify the malfunctioning state of the MPO. The detection time in the prototype method exceeds the time taken to detect similar deviations when using the inventive method. Thanks to proactive detection of a trend of deterioration in the operating conditions of the MPO, it is possible to timely eliminate the cause of the malfunction and increase the speed of detection and accuracy of assessing the state of MPO. Thus, it is possible to achieve the planned technical result.

The ability to implement and obtain a positive effect in the implementation of the claimed method was also tested by creating a mock control system and conducting a full-scale experiment. During the experiment, various combinations of controlled parameters were used, and the correlation indices P _{ij were} calculated between them when simulating the effects of various external factors. The experimental results showed a significant increase in the reliability of assessing the state of MPO, which also indicates the possibility of achieving the planned result using the proposed technical solution.

Claims (2)

1. A method for monitoring the state of a multi-parameter object, which consists in pre-setting the set N≥2 of its structural elements of the same type and K> N measured parameters characterizing the state of the same structural elements, in the process of functioning of the multi-parameter object, the measured parameters of the structural elements are measured and stored, determine the values of the calculated parameters of the structural elements, document them and make a decision on them about the state of a multi-parameter object a, characterized in that G areas of memory are pre-set in the storage device, where 1≤G≤N, for storing identifiers I _{g of the} gth group of the same structural elements, where g = 1,2, ..., G and the threshold value of their coefficient the similarities of λ ^pores , the set M∈K of uncorrelated measured parameters characterizing the state of each of the same structural elements, the measurement period T _u and standard values

of these parameters, tolerances

parameter values from standard values, allowable differences

between the maximum

and minimal

levels of the measured m-th parameter, where m = 1,2, ..., M, i-th and j-th structural elements of the same type belonging to the same group, the permissible value of the correlation index P ^add between the measured parameters belonging to the i-th and j- mu of the same structural elements, measure the values of P _mi parameters, calculate the similarity coefficients λ _ij , of the same type of structural elements for all m parameters, where i = 1,2, ... N and j = 1,2, ... N, and i ≠ j, compare the calculated values of λ _ij with a predetermined value of λ ^then , according to the results of comparison, are stored in the corresponding x memory areas of the storage device, identifiers I _{g of} grouped structural elements of the same type belonging to the gth group, the difference in the similarity coefficients of which exceed a predetermined value of λ ^pores , during the operation of the object, values П _{mi of} measured parameters are measured with a period T _u and values of calculated parameters determined for each gth group of structural elements, where g = 1,2, ..., G, by calculating the average values of the measured parameters

and their deviations

from standard values

differences

, time intervals τ _mg , after which the values

may exceed acceptable values

, values of correlation indicators P _i , P _j for structural elements having the maximum

and minimum

values of the m-th parameter, compute the calculated values,

, ΔP _mgij , P _i , P _j with predefined acceptable values

,

, P ^add , and after documenting the values of the state parameters of the structural elements, a decision is made about the malfunctioning state of the multiparameter object, if the condition

and / or

, and / or P _ij > P ^add , and additionally document information about the parameters of structural elements that have invalid values, and when the condition

and

and P _ij ≤ P ^add decide on the working condition of the multi-parameter object and continue to periodically measure and calculate the values of the parameters of the structural elements. 2. The method according to claim 1, characterized in that the correlation index P _ij between the parameters of the i-th and j-th structural elements of the same type is calculated by the formula:

,
where η _kl is the correlation ratio of the k-th and l-th measured parameters of the i-th and j-th structural elements of the same type, where k = 1, 2, ..., M and l = 1,2, ... M, k ≠ 1;

- the multiple correlation ratio of the k-th parameter with all other parameters of the i-th and j-th structural elements of the same type.

Images