RELIABILITY ASSESSMENT FEATURES Management systems

In any automated control system, one can distinguish a set of technical means (technical systems, objects - the meaning of the terms is the same), a team of people (you can distinguish between operators, maintenance specialists, work managers) and programs that implement control algorithms.

The complex of technical means (KTS) ACS has a number of features. Such complexes are subject to high demands regarding the reliability and timeliness of processing large amounts of information. Often, individual CCCs are geographically fragmented. The structure of individual complexes is often unique, although standard elements of a few types are usually used. Hierarchical structures are widely used.

Until recently, the theory of reliability dealt only with technical systems. The problem of reliability of technical systems is still the most important and attracts the most attention. Thanks to the development of automated control systems, interest in works on the reliability of human and technical systems has increased. Such studies are especially important to ensure the safe functioning of transport and industrial systems.

By the reliability of the “man and technology” system, we understand its ability to perform specified functions for the required period of time or the required operating time, while maintaining its operational performance within specified limits.

The study of the reliability of the "man and technology" system is reduced to considering the reliability of the technical system taking into account the activities of operators or to considering the timeliness (reliability) of people performing a system of work to achieve a given goal. For CCS ACS, the first statement of the problem is of greater importance.

The experience in the development and application of ACS also indicates that the most important problem is the reliability of complex control programs operating in real time.

We can talk about the reliability of programs as their property to fulfill the requirements for the program for a certain period of time in real operating conditions. Due to the presence of hidden errors in the programs, emergency situations can occur and the efficiency of the automated control system is significantly reduced.

Thus, when considering the reliability of the designed ACS, it is advisable to separately evaluate:

reliability of the technical system;

reliability of the “man and technology” system as the reliability of a technical system taking into account the activities of operators;

reliability of algorithms (work system plans);

reliability of control programs.

The resulting vector of values ​​of the reliability indicators can be used to directly judge the reliability of the ACS or used as input to estimate the average losses due to unreliability, which indirectly characterize the reliability of the ACS.

When developing a method for studying the reliability of technical means, it is necessary to take into account that modern automated control systems are usually developed for a specific organization in a single copy, that is, systems are unique. At the same time, elements of technical equipment in most cases are serial.

The reliability of the calculations and other methods for determining the reliability of the system is of great importance. It is advisable to use reliability assessment methods that do not require the introduction of doubtful assumptions.

When analyzing the reliability of technical means, it is also desirable to take into account the multifunctionality of modern ACS, which are usually designed

Fig. 3.1. Failure tree example. A1, A2, A3 - subsystems; B1, B2 - blocks; C1, C2, C3 - nodes; D1, D2, D3, 41-elements.

to solve several sets of problems. This can be done using the technique described below, conventionally called the deductive method of researching reliability.

There are two ways to formulate the concept of system failure, which can be called inductive and deductive methods of reliability analysis. With the inductive method, the types of failures of the elements are established and the influence of the failure of each element on the performance of the system is determined. In this case, it is usually possible to reduce all the failures of the elements to a small number of species. For example, for electronic components, breaks, short circuits, and drift are usually considered. Considering a combination of possible conditions; elements, you can find faulty system conditions. When sequentially considering the failures of all elements, it is unlikely that an accidental skipping of possible system malfunctions is unlikely. However, the method is very time-consuming, it is necessary to consider all the failures of the elements, since the criticality of the element becomes known only after analysis.

With the deductive method of analyzing the reliability of the system, all possible system failures are listed and which blocks, elements, etc., can lead to a failure of the type in question. During the analysis, a fault tree is constructed (Fig. 31). Therefore, the method is sometimes called the fault tree method.

The deductive method is advisable to apply at the early stages of design to identify weak links in the system before performing reliability calculations. At the same time, the developer’s attention is focused on dangerous situations that are not hidden behind assumptions and simplification

3.1. Asu industrial type. The problem of reliability.
We introduce the following concepts:

Technological process (TP) - we will consider that TP includes both the simplest technological operations and such complex systems as the development of modern machines or computers, that is, all modern production is represented by a combination of different TP.

Technological object (TO) - those technical (non-technical) means by which TP is performed. Technological objects include both the machine and the plant.

All maintenance forms a control system (SU), which are divided into manual (without the use of technical means), automated (with the participation of a person) and automatic (without the participation of a person).

For ACS we will distinguish automation levels:

With a low saturation of technical means.

Systems in which a person is assigned secondary operations.

ACS are divided according to complexity, and a clear line is drawn between the ACS and the maintenance that it manages.

The combination of maintenance and automated control systems will be called an automated technological complex.

ACS used in industry are divided into 2 groups:

Automated process control systems (APCS).

Automated enterprise management systems (ACS).

Integrated automated control systems are also being created, which include direct process control.

The relationship between efficiency and reliability of the ACU.
The rank of the used characteristics of the product is established, and for the product they establish a single indicator of the quality of functioning, to which all other technical characteristics are reduced - this is an indicator of efficiency.

In the presence of an efficiency indicator, a criterion is formulated for optimizing the solution of various issues in the development and design of products, most often it is an extreme criterion of the form:

W → extr (3.1)

W → max or W → min (3.2)

Example of criteria: the maximum probability of error-free fulfillment of a given function and the minimum cost of manufacturing a unit of product.

Denote some generalized reliability indicator of ACS by “a”, then there is a relationship:

Reliability problems asu.
There are a number of state and departmental documents, methods of research, evaluation and improving the reliability of complex systems have also been developed.

There are three main reasons for the poor study of reliability issues when creating automated systems:

The main task: the creation and deliberate implementation of the maximum number of systems. At the same time, issues of reliability efficiency remain in the background.

Lack of standards for reliability of ACS. GOST 171994 - 71 and GOST 17195 - 71 only to a small extent affect reliability issues.

With the current scope of development of automated control systems, the reliability service and design institutes are not able to carry out the necessary work and their main volume falls on

ACS development engineers who do not have sufficient training and experience and are not qualified to perform calculations and calculations.

The most important task is to bring together all the work on the reliability of ACS, develop a common policy and create a unified coordination work plan and develop a set of regulatory technical and methodological documents on the reliability of ACS.

Reliability of components.
There are three main tasks:

Establishment of nomenclatures, characteristics and reliability indicators of GPS products, providing sufficient completeness of information necessary for calculating reliability.

Development of a set of measures to ensure the reliability of reliability indicators indicated in the technical documentation for GPS products.

Establishing sound standards for the value of reliability indicators for all types of GPS products.

Reliability includes four components:

Reliability.

Durability.

Maintainability.

Persistence.

All of them are associated with certain random variables having the dimension of time.

T is the uptime.

Tv is the recovery time.

TD - time to the onset of the limit state.

Tc - retention time.

In the theory of reliability, the following forms of specifying the probability distribution of random variables are used.

The integral function F (x) is the reliability function (3.4)

Differential f (x) - uptime (3.5)

Inverse integral function G (x) = 1 –F (x) (3.6)

Intensity function H (x) = f (x) / G (x) (3.7)

As numerical indicators:

Tsr - average uptime

Tv.av - average recovery time

Tr.s. - average resource

Ts.sr - average retention time

[P (tfix)]

[Fв (tfix)] - probability of recovery time

[Gr (tfix)] is the probability that the resource of the product sample will exceed a fixed value.

Suppose that GPS products have an excess structure, therefore, the T distributions inherent in them are simple and are determined by the interaction of physicochemical processes that occur in elements and parts.

Then we observe the performance of the system.

Efficiency is the state of the product in which it is able to fulfill the functions assigned to it, and its functional parameters are within specified limits.

Product failure can be defined as going beyond the specified limits of at least one of the functional parameters of the product.

If the functional parameters are Z1, Z2, ..., Zn, then the product working conditions can be considered:

Zin ≤ Zi ≤ Ziв, where i = 1,2, ..., n (3.8)

Zin and Ziv are the lower and upper tolerance limits for the ith functional parameter.

x1, x2, ..., xN; y1, y2, ..., y1K;

Zi = φi (x1, x2, ..., xN; y1, y2, ..., y1K), where i = 1,2, ..., n (3.9)

φi (x1, x2, ..., xN; y1, y2, ..., y1K) = Zin

φi (x1, x2, ..., xN; y1, y2, ..., y1K) = Ziв (3.10)

These equations define 2n surfaces in the product functional space — these surfaces bound a certain region D, all points of which satisfy the set of inequalities (1.8) and which is called the stability region.

PPEs participate in the formation of FPI - they determine the ability of the product to withstand various external loads: g1, g2, ..., gu; h1, h2, ..., hS.

The load does not affect the parameters of the FPI, but is the cause of the failure.

If hi> gi– then a failure occurs.

W = W (a) (3.3)

ACS are classified according to three criteria:

Systems for industrial use.

The main indicators of the quality of functioning are indicators of economic efficiency.

The systems are equipped with products included in the state system of industrial devices and automation.

In accordance with feature 1, such systems will be called industrial. In accordance with the second main criterion, there is an “economic” criterion for evaluating and comparing measures that can affect the reliability level of industrial automated control systems, and the reliability level that ensures maximum economic efficiency is optimal.

The third feature is based on the tendency to complete these systems from the arsenal of GPS tools, with the exception of specially developed systems.

Destabilizing processes and classification of failures.
Destabilizing processes include two groups:

Fluctuation processes of the WWF value (both signals and loads).

The aging and wear processes of the elements and parts that make up the product. These processes lead to failures, which differ both in external manifestations and in measures to combat them.

Under operating conditions, these processes change in time randomly, i.e. possess certain statistical properties.

Long trouble-free operation is ensured by the presence of:

Certain reserves of functional stability.

Accuracy margin for all types of loads.

When the instantaneous value of a factor goes beyond this area, a product failure occurs:

If the process cannot be predicted based on an analysis of the product itself, then such failures are called sudden failures. They are of two types: a) if the cause is an accidental change in the signal, then such a failure is unstable (self-eliminating). When the signal returns to normal, the performance of the product is restored (sudden failure); b) if the failure occurred as a result of an abrupt increase in load, then such a failure is called stable (sudden catastrophic failure).

If the parameters change under the influence of aging and wear processes monotonously - this leads to a gradual failure. Such failures can only be stable and are divided into two types: a) if the cause is a change in the PES, the failure is called gradual parametric; b) if the product cannot withstand the rated external load, a sharp violation of the product’s performance occurs - a gradual catastrophic failure.

Combined failures, they are both unstable and stable.

Breakdowns running in.

3.2. A formalized description of ASU structures.
The structure of the system is the composition of its elements, a list of the functions it performs and the interaction of the elements in the performance of these functions.

The primary functional-technical description (PFTO) will be called the usual description (technical documents, diagrams, verbal description, etc.)

PFTO is quite complete, i.e. contains all the necessary information about the structure of the system. The language of PFTO is informal. To study reliability, you need to go to a formalized description of the system.

All possible languages ​​for describing structures can be divided into two groups:

Analytical (formulas and tables).

Graphic (graph languages).

A graph is a diagram representing the structure of a system; we will call them reliability-functional diagrams (NFS).

NFS - a certain graphic image that displays the elements of the circuit and its functions, and that allows using a set of formal rules for an arbitrary set of states (operability or failure of all elements) to uniquely determine the state of the system for each of its functions.

If the state of the system is Zj, then Zj = {x1, x2, ..., xn}

n– number of system elements

xi– element of the system (indicates the state of the i-th element)

xi = 1 - performance;

xi = 0 - failure.

- system elements - functions

Arrows indicate the interaction of elements in the execution of functions.

In the graphs, nodes of higher and lower ranks are distinguished: the highest (lowest) rank, if all the edges connected with it are graphs directed only to it (from it)

To perform function 17, the operability of one of the elements 13-16, elements 11.7.5 or 11.7.6 is necessary.

To perform function 3, the operability of elements 4 and 5 or 7 and 6 is required. If the element is in a failure state, then the corresponding node in the graph is closed.

Types of connection elements.

1) Independent functioning of elements

Each of the n elements performs its function independently of the operation of other elements, i.e. contains top nodes and lower rank nodes.

2) Serial connection of elements.

All elements are involved in the performance of certain functions and each of them is necessary.

n + 1 - th node includes the implemented function.

3) Parallel connection of elements

In a reliable sense, a function is executed if at least 1 of the n elements of the system is in working condition.

In the case when the structure under consideration is not part of a more complex system and functions independently, the graph is supplemented with n + 1 nodes of a higher rank, assuming that this node is absolutely reliable.

4) Two-function system.

It includes two elements and two functions, moreover, both elements are involved in the execution of one of them, and the other has only one.

This system is displayed by a graph in which one node (1) corresponds to two nodes (11 and 12), and only one of them 12 has the highest rank. When analyzing the reliability of this system, it is understood that the nodes. Corresponding to the same element have not only the same reliability characteristics (distribution of random variables), but also the implementation of these random variables.

Majority system ("2 of 3").

A three-element system that performs one function and maintains operability with the operability of any 2 elements.

Bridge systems.

Widely distributed in energy systems. Easily fit into the NFS graphs.

Using all of these structures, any system can be described. NFS graphs display only the reliability properties of structures in relation to the functions they perform. On the basis of NFS graphs, graphical displays of the properties of complex multifunctional systems with respect to their effectiveness can be built. The effectiveness of a function over a fixed time interval is usually proportional to its duration. Therefore, we introduce the concept of specific efficiency of the i-th functionei, then the efficiency of the entire system:

Еi = Σyit ei (3.11)

yit is a binary variable that takes the value of unity if at the time the tth function is executed and 0 otherwise.

Equation (3.11) is valid only in those cases when the performance indicators used in the system allow summation (economic systems).

ei (i = 1, ..., m) is a reliably efficient system circuit (NES).

3.3. Characteristics and reliability indicators of asu.
Characteristic features of ACS:

As a rule, they have complex and redundant structures, therefore, a complex form of distribution of uptime, which is difficult to reduce to one or two parametric mathematical models.

They are complex multichannel and multifunctional products. Moreover, the role of individual functions in the general task of ACS can be different.

For some products, the overall (integral) efficiency is significant in a certain time interval.

Their integral effectiveness can be considered as the sum of the efficiencies of all the functions they perform.

They are designed for continuous operation, during which multiple failures and restoration or replacement of almost all system components are possible.

Reliability is determined by the ability of the system to remain operational under operational conditions for a given time without forced (unscheduled) interruptions.

Maintainability - characterizes the adaptability of the system to the prevention, detection and elimination of failures. This property is important for ACS designed for long-term use with multiple restoration of performance in case of possible failures.

Durability– characterizes the property of the system to maintain working capacity to the limit state.

The durability of the ACS is determined by the factors of moral aging and, therefore, very little depends on the developer. It is determined by the control principles laid down in the system and is very weakly related to the structure of the system, as well as the parameters and characteristics of the components used.

Consider the quantitative characteristics and indicators that make up the reliability of the ACS.

Durability and retention characteristics are the easiest to choose. The generally accepted indicators are taken here: Tsl - service time; Tc - retention time;

To select the characteristics of the maintainability indicators of ACS, we have a special technique.

For multifunctional systems, it is required to set these properties for each of the functions performed separately. Moreover, the characteristics of the reliability components for each function is the distribution of random variables: Ti and Tvii = 1,2, ..., m;

m– number of system functions.

Indicators are the numerical characteristics of these random variables.

If for some function there are several types of failures that differ in their callability or consequences (the value of losses and failure, time to eliminate failures, cost of eliminating failures, etc.), then failure-free and maintainability must be set for each type of failure separately.

3.4. Research Methods and ACU Reliability Assessment
All methods of research and reliability assessment of ACS (as, indeed, of any other objects) are clearly divided into three groups: analytical, experimental and statistical modeling. A special group is combined methods.

Analytical methods make it possible to evaluate reliability and compare various ACS options and find optimal solutions at the earliest stages of design. Analytically, the influence of various factors can be determined, optimal requirements for the reliability of the ACS and its components, optimal (according to the criteria of maximum reliability or efficiency) values ​​of its parameters, maintenance modes, etc. can be found.

An analytical study, as a rule, does not require large expenditures, but for rather complex objects, which are ACS, it involves a significant amount of calculations and time-consuming.

A significant advantage is that the solutions can be obtained in the form of analytical expressions that allow us to study the influence of various factors and find their optimal values ​​in a general way.

The necessary initial data for analytical research (in addition to data on the structure of the ACS) are information on the reliability characteristics of all used components. This is a disadvantage because reliable data on the reliability of components is sometimes impossible to obtain. If the data are not reliable, then the resulting estimates obtained

may differ significantly from the actual numbers. But even in this case, analytical methods allow you to compare the options for the structure of the ACS.

To analytical methods - according to the statement of the problem - the method of statistical modeling is very close. The similarity is that both require data on the reliability of system components. However, the methods for obtaining the results vary significantly.

The method of statistical modeling consists in generating (using special random number generators) random periods of uptime and recovery time of individual components of the ACS and thus “artificial” reproducing the process of functioning of the ACS. If these random number generators have probability distributions that coincide with the distributions of T and Tv corresponding components of the ACS, then the constructed process model will have all the statistical properties of the real process of functioning of the system. With a sufficient duration of this process, statistical estimates of the characteristics and reliability indicators of ACS obtained on its basis can have arbitrarily high reliability.

This method allows the development of standard algorithms and programs suitable for research and reliability assessment of a wide range of ACS.

A positive property of the method under consideration is that, as a result of modeling the real process of ACS functioning, not only the number of reliable characteristics and indicators can be obtained, but also the characteristics of the efficiency of functioning.

The disadvantage of the method (apart from the usual difficulties associated with having a machine, the need to compile and debug an algorithm and a program, and the relatively high cost of the solution) is that the results of the solution are presented not in the form of analytical expressions, but in the form of numerical estimates. In the course of solving the problem on the computer, it is not visible how the influence of individual factors on the final result is manifested. For this, it is necessary to vary the values ​​of individual factors and analyze the many solutions obtained. Moreover, the volume of the “mathematical experiment”, and with it the volume of computer time and its cost increase many times.

It should be noted that this method can play a large role as a means for checking and evaluating the accuracy of the proposed approximate analytical methods. This is due to the fact that for complex structures of modern ACS, accurate analytical methods lead to such cumbersome expressions that their practical application becomes impossible. Therefore, on the one hand, approximate analytical methods are being developed for which the question of estimating the accuracy of approximation is always relevant, and on the other, for many systems at the design stages, the statistical modeling method is the only available method.

Experimental methods are, in fact, the only way to get a final answer to questions about the correctness of the completed system development, the results achieved, and the level of reliability of the created system that is actually achieved.

A significant advantage is that they do not require knowledge of the reliability properties of system components. Moreover, testing the system as a whole - with a thorough analysis of the failures that arise - allows you to clarify the data on the reliability of components in real-life conditions.

The main disadvantage is the very high cost.

An experimental assessment of ACS reliability can be implemented in two versions: 1) organization of special tests and 2) collection of statistical tests