The CHAPTER 8. PRINCIPLES of CONSTRUCTION
MODERN MIS
The application of modern means of circuitry (integral gauges, chips, microprocessors etc.) radical image changes principles of construction MIS. Besides methods the reasonable determination of information streams suffices allow essentially to reduce their redundancy. It allows to deliver a problem about possible maximum transposition of processing of a measuring information to a place of it shaping. That is to proceed to conveyor processing of a measuring information in the distributed informational - measuring system. As a whole such system consists of the following main bodies: systems of primary converters (gauges), devices of the collection and primary information processing, means of secondary information processing, control units and monitoring, intercouplers with other systems of object, accumulators of an information. The construction of such systems in this or that branch of human activity can essentially differ. We shall try to consider this problem in area, where requests to an equipment most rigid. Namely, in the field of an air and space-rocket mechanical engineering. Naturally, such systems can be applied and on other mobile kinds of engineering (automobiles, courts etc.).
Except for general requests presented to MIS (exactitude and reliability of an information), in conditions of the space application to MIS, as well as to all onboard equipment, such
requests as are presented:
- Minimum of weight-dimensional performances;
-
Stability to transhipments, impacts, vibrations;
- Stability of work in
conditions of heavily radio electronic parasites;
- Work in a broad band of
temperatures;
- High reliability and maintainabity;
-
High ergonomic properties.
All it generates before the developer a mass of problems. I hope, that some ideas stated in the present work, will help with their solution.
So, briefly we shall consider main groups of means which are included in MIS. To a him it is possible to refer a totality of primary converters (gauges), means of the collection and primary
information processing, means of secondary processing, control facility and monitoring, accumulators of an information and communication facilitieses with systems of object.
1. The totality of primary converters (gauges) can be divided into three main groups:
-
Gauges, in which any other functions, except for transformation of a data-in to an electrical signal basically it is impossible, for example strain gauges;
-
Gauges, in which it is possible to place some electronic components permitting to form test signals to keep datas for linearisation etc.;
-
Gauges, in which it is possible to make primary information processing, including, such gauges can be, as a matter of fact, other onboard means and systems.
2. As the gauges are placed on all object, it is meaningful to group them in kinds and localization on object and to connect to blocks of primary information processing under the radial scheme. Blocks, as they allow to realize the complicated enough protocols of exchange, to unit among themselves and other computing devices of a system through a parallel data link, for example of MANCHESTER - 2. Thus, general scheme of a system is possible to present as the scheme of a Fig. 54.
Into a system enter:
- Blocks of primary information processing from gauges, in which what or the information processing can not be made basically, for example strain gauges;
- Blocks of primary processing of a measuring information from gauges, which design allows to make some functions connected with information processing;
- Special gauges. Usually such gauges, in which is made the primary information processing also is present an output on a data link in a system;
- Blocks of interface with other sources of an information, for example with regular systems (course systems, inertial systems of navigation etc.), and also with other measuring systems. Besides it can be the systems of telemetry communication. And also executive systems (for example, systems of automatic control, autopilots etc.);
- The various accumulators of an information, including emergency registrars can enter into a system;
- The information from all sources arrives in the central computing device, in which the secondary information processing is made, and the outcomes it are removed on the device of indication, in control systems of object, on telemetry, accumulators etc.;
- If MIS not regular, in it the device of indication and management can enter.
3. As it was marked above, the gauges can be classified as three groups - a tensometrical type, gauges with built-in electronic components and special gauges.
To gauges such as tensometrical type it is possible to refer all gauges, in which there is only sensing element, where under an operation a measured parameter varies what or electrical performance (resistance, capacity, arises EMF etc.). What either means of information processing or means of interface with a system in such gauges cannot be placed. It is necessary or to use for these functions the separate specialized block, and then it will be on our classification the specialized gauge, or, grouping such gauges (if they need) to be connected them to the block of data acquisition. As a whole such block will have a structure general for such blocks, and to differ by the circuits of interface with gauges. The primary information processing on gauges of a tensometrical type can be made because of of nameplate data on gauges of the given type have certain average static and a dynamic characteristics.
If the features of application and design performances of the gauge allow to place in it electronic components, that is basis think of that what functions these components can to execute. From the point of view of a system approach it would be desirable, that the replacement of the gauge for want of it to breakage or removal on dates did not require modifications in a software MIS.
In this connection, in the gauge should be reprogrammable the storage with an electrical rewriting of an information of small volume. There, in process testing of the gauge with the help of of appropriate automated means, the data record about to performance as factors designed on techniques circumscribed is made above.
It is desirable, that in the gauge there was a source of a test signal for monitoring serviceability of the gauge.
The gauge should have interface to the block of data acquisition with the standardized protocol permitting. To give the command to the gauge on issue of a control measurement, a datas and current information. The connection is desirable for realizing under the single-wire circuit with transfer of a discrete digital code, that allows to receive an acceptable guard from parasites.
For want of of sufficient degree of standardization of such gauges, the electronic circuits can be executed with a high degree of integration, have extreme small weight and dimensions high reliability.The structure of such gauge is shown in a Fig. 55.
Fig. 55 |
Into the scheme of the gauge enter: the primary converter and source of a test signal connected to an analog-to-digitial converter (ÀÖÏ), with which information arrives on a converter of a parallel code in sequential, whence, through the circuit of interface, on the communication line with the block of data acquisition. Normal voltage on the communication line - high. For want of to submission of this voltage from the block of data acquisition on the gauge, it, as the signal, acts on the control circuit. For want of it the sequence of work of the gauge is started, for want of with which the test signal, a datas of factors, and then a measured parameter are transmitted to a system.
4. The blocks of data acquisition are of the same type and can differ by the circuits of interface with gauges and programs of primary processing. The block diagram of such block is shown in a Fig. 56. Common for such blocks are the site of the processor consisting of the processor, RAM and ROM, communication center with the interface of a system and site of a feed. The communication center with gauges should meet to features of a collection of gauges, which are connected to it. The software of the processor noted in a ROM, executes tasks on a realization of monitoring of gauges, deriving from them of an information about service faktors (including factors of process of scaling), deriving of an information about current significance of parameters, primary processing of a measuring information on techniques stated above. The inquiry of gauges is made in the correspondence with algorithm loaded from the central computing device, where the frequencies of inquiry of gauges, list of interrogated gauges and parameters of their primary processing are specified.
Fig. 56 |
|
Fig. 57 |
The requests for deriving of an information about parameters arrive from the central computing device. From it the commands on testing of gauges and block arrive.
5. The central computing device has a structure shown in a Fig. 57. Design feature of the central computing block is the availability of a remote memory block. In case if MIS is developed as regular, it can and be absent, as the software in this case does not vary and can completely be placed in a ROM. However if the system is applied in conditions, when the problems executed MIS vary (for example for want of tests of object), such block is necessary.
The remote memory block is intended for two purpose. At first, in it the collection of the programs defining specificity of work MIS in conditions of the concrete task is placed. For want
of it the modes of their primary processing, algorithms of secondary processing are determined a collection of measured parameters, the algorithm of interaction with an operator, list of
fixed parameters etc. is determined during fulfilment of the task, outcomes of full processing of measurements also are noted in a remote memory block. After fulfilment of the
task a remote memory block is removed and arrives for further processing and analysis to the appropriate experts. Structurally remote memory block can be made as a memory block
on cylindrical magnetic domains.
ROM of the
central computing block, except for the general system programs, should have in itself the package of a software for support of secondary processing, in particularly programs of calculation
of transcendental functions. The theoretical fundamentalses of construction of such programs are stated above.
6. Except for circumscribed blocks, in a system can enter a number of devices of the coordination with regular systems of object, with telemetry etc. In the theoretical plan, on our sight, the special problems in their development of problems do not arise, and on them we shall not stay.
The separate interest represents a problem on design realization of the indicator. Especially if the system is not regular and is applied, for example, to tests of mobile objects (for example, flight vehicles). In this case to the indicator the special requests are presented. It should be small-sized and allow to install it in various conditions.
![]() |
Photo 1 |
![]() |
Photo 2 |
Example of such indicator can be the device shown on a Photo 1. It allows to include MIS, to make it monitoring at any time, with the help of of switch and button "MODE" to select the necessary program of measurements and processing. For want of it on the display the entry conditions of realization of experiment are removed. The operator (pilot-verifier) results object in the correspondence to specific conditions and presses a button "START-UP". On the display the assigned current modes and, after ending experiment are displayed, on the display the information about successfulness or not successfulness of experiment and (in case is removed if the experiment is executed successfully) generalized outcomes of secondary processing. The following mode etc. is further selected. The mounting conditions of the indicator can be various. Including, can fasten on thigh of the pilot, as shown in a Photo 2.
7. Except for purely MIS, for maintenance it of work some auxiliary systems are necessary. In particular, means for testing of gauges, means of automation of preparation MIS to concrete
tasks, and also means of the automatic analysis of outcomes of measurements.
The means for
testing of gauges can form as an automated job have a structure shown in a Fig. 58.
|
Fig. 58 |
|
Fig. 59 |
In the correspondence with the specific program of testing, from the personal computer assigned magnitude of a standard parameter through the interface unit ÓÑ1 gives also source of a standard parameter ÝÒ, whence parameter gives on the gauge Ä. From the gauge the information arrives through the interface unit ÓÑ2 in the computer. After removal of datas on all range of a a parameter, the datas are treated, and obtained factors are writed in the gauge. For each type of the gauge there are appropriate mathematical means and subsystems of sources of standard parameters. The connection them to the computer can is made manually, and can and with the help of of switchboards.
8. The means of preparation MIS and automated analysis can be incorporated as a job of the engineer (ÐÌÈ). The structure ÐÌÈ is shown in a Fig. 59. Into it enter: the personal computer of the PC, interface unit with remote memory (ÂÏ) ÓÑ1, interface unit with an accumulator of an information (ÍÈ) ÓÑ2, and also other necessary means (for example, communication facilitieses with telemetry, with means of outobjects measurements etc.).
In ÐÌÈ the appropriate software including is formed:
- Data base about
objects of the given kind;
- Data base about the given type of object;
-
The sample programs of measurements;
- The shell of programming ÈÈÑ;
-
Handlers of an information with ÂÏ together with an information from accumulators of an information, datas of telemetry and outobjects measurements;
-
The programs of the automated analysis of outcomes of measurements with allowance for informations on objects of appropriate kinds and types;
-
The programs of preparation of the necessary documentation etc.
The shell of programming MIS allows to select a necessary collection of experiments, to introduce in them appropriate datas (initial modes, lists of gauges, frequency of their inquiry etc.), to note them in ÂÏ, to make modelling (especially on parameters connected with safety), documentary to issue the task for experiment etc.
9. The circumscribed structure technical and software MIS allows rather essentially to reduce terms and costs to realization of works connected with measurements. It is achieved
following.
At first, the object can completely be equipped
with all collection of gauges and blocks of MIS. During realization of researches, under each experiment the population of gauges and handlers of an information is determined program. Due to
this, the process of preparation of object to realization of experiments is sharply reduced on time and expenditures of labour.
Secondly, the process of realization of measurements and processing can be made simultaneously in real time, that allows sharply to reduce time and expenditures of labour to information
processing.
Thirdly, the collection of experiments of
realization, executed at the given stage, of works (for example, in the given test flight) can be programmed with some redundancy. Besides the dialogue form of dialogue of an operator
with MIS allows it strictly enough to maintain parameters of experiment and to know it outcomes. As a whole, it allows essentially will increase an overall performance of an operator and
maximum qualitative to execute the task, that also essentially reduces terms and costs to realization of works.
In - fourth, the determination of frequencies of inquiries on the above indicated techniques allows to ensure metrological validity, and, therefore, reliability of outcomes of measurements.
In - fifth, the account of an amount of gauges on techniques circumscribed above (in particular, of tensometrical gauges), allows to determine the metrologically reasonable
amount of gauges, that provides reliability of an obtained information and, probably, to reduce their amount up to really necessary.
In - sixth, the automation of processes of preparation MIS to experiments, realization of experiments both information processing in real time and analysis after realization of experiments
provides a possibility of a concentration of all works in one hands, hands of the expert working on ÐÌÈ. It sharply reduces losses of time usually originating from mismatching and from
absence of mutual understanding, if many experts participate in work. In this connection, there is a possibility sharply to reduce an amount of staff, time to fulfilment of works, and
consequently also of costs of them.
APPENDIX 1. THE REVIEW of THEORETICAL BASES DIGITIZATIONS
The problem of digitization is widely covered in the literature. It is considered, that the first basic works in this area were work of Kotelnicov V.À. [1] , left in 1933ã. and irrespective of it the article [2], written by Ê.Shannon and published in 1949ã. The article of the Shannon was written because of works of the Whittaker “ to Function the theories, represented by distribution, of an interpolation ” [34] left in a 1915, and also [35].
In the mathematical plan, problem of function restoration by separate significances and restoring it with the help of interpolations of a beginning to be decided in 18 century in work of the Laplace, of a Newton etc. [7]. The rough development of the theory both practice of communication and information as a whole has confronted with necessity of further development of methods of determination of sampling rate, optimum interpolation both stochastic performances of digitization and interpolation. Thereof, the number of attempts was undertaken to decide this problem both in the general, fundamental plan, and for concrete practical application.
Such works, as the article Æåëåçíîâ Í.À have appeared. “ A Research of orthogonal expansions of stochastic signals with a unlimited spectrum ” [36] left in 1957ã., work Õëèñòóíîâ Â.Í. “ A Fundamentals of digital electric engineering ” [37] etc. Have appeared a number of attempts, by and large to improve main methods, for example in such work, as [38,39,40,41 etc.]. However practice of creation and use MIS does not give the basis to consider, that there are simple enough, convenient and effective methods of account MIS.
The criterions of determination of sampling rate can be grouped together as follows:
1. Frequent criterions, for want of which sampling rate is selected in view of frequency of a
function spectrum;
2. Correlation criterions installing connection between an interval of digitization and an interval of an function autocorrelation;
3. Quantum criterions
of readout [42] applaud to a determined model of a signal and installing dependence of intervals between readout from significance of a stage of quantization on a level and steepness (first
derivative) signal;
4. Criterion of the greatest deviation concluded in choice of such interval of digitization, that a residual between function with the limited number of derivatives and
Lagrang polynomial did not exceed the certain magnitude.
1.
The frequent criterions are based to the theorem of readout. In Russia it is known more as the Êîòåëüíèêîâ theorem (in the primary source the Theorem II ), which states: “
Any function F ( t ) consisting from frequencies from 0 up to f 1 , it is possible continuously to transmit with any exactitude with the help of of numbers of the following one
after another through 1/2 f 1 s. Really, measuring magnitude F ( t ) For want of t = n /2 t 1 ( n - integer), we shall receive: F ( n /2 t 1 ) = D n
w
1 (8) {numbering of the formulas under the primary source}.
As all members of a row (1) for this significance f will convert in zero except for member with k = n , which, as it can be received, uncovering indefinitency, will be equaled D n w 1 . Thus, through each 1/2 f 1 s. We can place next D k . Transmitting these D k on queue through 1/2 f 1 s, we can till a him agrees equ. (1) termwise to restore F ( t ) with any exactitude. ”
This theorem is justified by the Theorem I , Which states: “ Any function F ( t ), consisting from frequencies from 0 up to f 1 per. in s, it is possible to
present by a row:
(1),
Where - integer,
- constants dependent from
.
And on the contrary, any function
represented number (1) consists only from frequencies from 0 up to
per. in s.
THE PROOF {is resulted with some insignificant reductions}
Any function , Satisfying to a condition Derihle (final number of maximas, minima and point of discontinuities on any final section) and integrable in
limits - ¥ Up
to + ¥ , that
always in the electrical engineer takes place, the Fourier can be represented by an integral.
(2).
And:
(3),
As consists from frequencies from 0 up to
, that
and
for want of
and consequently
can be represented according to equ.2 so:
(4)
Functions and
on a plot
can be represented by Fourier serieses, and these
numbers, at our desire, can consist of one cosines or one sines, if we for period take
( See. Ñìèðíîâ - “ a Rate of maximum mathematics ”, ò. II
èçä.1931ã., page 385 []).
(5à)
(5á)
Let's enter the following labels:
(6)
Then the formulas (5) can be copied so:
(7)
Substituting (7) in (4), after some transformation and integration we shall receive equ. (1). ”
The Êîòåëüíèêîâ theorem has played a significant role in development of the theory of communication and information. It is obvious for want of spectral submission of a signal. It advantage is the methodological character. From it follows not only sampling rate, but also method of restoring by a number Êîòåëüíèêîâà and method of restoring with the help of of ideal filter. However creation of telemetry systems has raised the question about an exactitude of transfer of a signal. Here again becomes obvious, that, despite of it usefulness in application to systems of communication, by which a request of metrology are not presented, in application to MIS it simply is not acceptable, as in this area are exhibited a number of defects in it. It has caused controversy and attempts to improve this theorem [etc.]. Some total of a controversy has bringed À.À.Õàðêåâè÷ [43], then the number of the publications on this theme sharply has decreased. For want of it the judgement was ratified, that the Êîòåëüíèêîâ theorem is the basic theorem in the field of the theory of impulse communication, which is, however, approximate statement.
In the foreign literature it is accepted to name the theorem of readout similar to the Êîòåëüíèêîâ theorem , as the theorem of the Shannon. It was published in a 1949 in the
article “ Communication for want of availability of a noise ”. This theorem sounds so: “ the Theorem 1. If the function does not contain frequencies above W gc, it completely is
determined instantaneous significances in moment, posed from each other on s. ”
The proof the following:
Let there is a spectrum
, then:
(2) (numbering of the formulas under the primary source) as
equally to
zero outside of a band W .
Having puted , we shall receive
In an interval on the right it is possible to find out n factor of expansion on harmonicses with period from - W Up to + W ,
It means, that readout
determine factors the Fourier in expansion
. ”
Is further proved, that the function can completely be restored on readout if apply impulse of a kind: .
(6)
Then Is determined as
. (7)
In the proof the Shannon refers to work of the Whittaker “ the Theory of interpolational functions ” left in 1935ã. Is conformable to these theorems and so-called barrier Nikawest. In modern issuings the conclusion of the Êîòåëüíèêîâ theorem on the Shannon is resulted, as a rule, [43,44 etc.]. More often Êîòåëüíèêîâ (Shannon) theorem is resulted generally without the proof [45,46,47.48 etc.].
As was mentioned above, in the theorem of readout are inherent a number of defects, which do not allow to consider it point, and consequently and aplicable for of examination
MIS and their account. These defects the following [45,32,3,43,41,47,49 etc.]:
-
Limitlessness of spectra of actual signals;
- Impossibility of creation of an
ideal filter;
- Complexity of accounts for want of restoring of function by the
members of a Êîòåëüíèêîâ rows;
- The theorem does not allow to
determine stochastic performances of errors for want of digitizations;
- The
additional errors are introduced that it is not obviously possible to consider samples on all axes of time from - ¥ Up to + ¥ ;
-
Even for want of acceptance of the theorem. Readout as approximate statement, the interval between readout is uncertain, as the boundary frequency is accepted to some extent arbitrary;
-
Duality of factors of a number, which can not simultaneously have significances of function in the certain moment and factors of a number of expansion.
Besides as will be shown in APPENDIX 2, the theorem of readout will not be agreed stroboscopic effect.
There are basis to consider, as the conclusion of the theorem of readout is not deprived of defects. Really, in a
conclusion of the Êîòåëüíèêîâ theorem the formula 5à leans on the formula from the Ñìèðíîâ book [50] , which is resulted for a
case history of an essence of harmonic expansion. Hereinafter it is replaced with more point kind of expansion the Fourier:
, which is resulted in all sources on a
Fourier analysis [51,5,11,52,53,54 etc.]. And, as a rule, underlines the special place of factor
. For example [51]: “ …
it is necessary, however, to underline the special position of a parameter
, Included in expansion only by half ”. The introduction in the formula
5à of half of constant component does impossible further transformation and conclusion of the theorem as a whole.
As to a conclusion of the theorem on the Shannon , the mentioned duality of factors of a number here has an effect. In essence conclusion consists of repeated application of a
Fourier series to a real-valued function . For want of first transformation the spectrum is received
, consisting from two
component - AFC and FFC. Application of an integral Fourier to
should express as an integration till the Fourier it of the valid and imaginary part as
real-valued functions, that will reduce already in four variables.
Therefore, integral Fourier from
can be taken as:
, that also concludes the theorems of readout of the Shannon impossible.
Really, there is a complex variable, the integral from which takes on a plane, that is till two variables. It is confirmed also by that having a gang of
equidistant readout from some sinusoid, on these readout it is possible to restore infinite number of sinusoids [55]. Proceeding from above-stated it is possible to understand the reason on
which proceeding from the theorem of readout it is recommended to select frequency of inquiries F 0 ³
( 3 ¸ 10)
F max and sometimes and it is more, and recommendations for determination) F max as a rule does not give.
2. The correlation criterion was offered by Í.À.Æåëåçíîâ in work [36] “ A research of orthogonal expansions of stochastic signals with a unlimited spectrum ”,
published in 1957ã. The essence of a method consists that the interval between readout is determined on magnitude of autocorrelated function of a signal. The advantage of a method consists
that it is distributed to signals have a unlimited spectrum. To defects it is possible to refer:
-
Applicability it only to stationary signals, as for non-stationary signals is not stationary magnitude, and, therefore, D Ò, designed for
average
, will result in an error for want of by higher dynamicality of process;
-
determination represents a rather challenge and requires large computing works, and also availability of realization of process, that for want of, as a rule,
is impossible to development MIS;
- From a criterion does not follow stochastic
performances of process of digitization;
- The criterion does not determine an
optimum interpolation.
The Æåëåçíîâ method is close to a Êîòåëüíèêîâ method , as a spectrum and autocorrelated function are connected by the formula of the Wiener - Õèí÷èí [45]. From the point of view of a Æåëåçíîâ criterion more general , while the Êîòåëüíèêîâ theorem describes singular processes [45,56], which, generally speaking, informations do not bear. In connection with the listed defects the Æåëåçíîâ method was not by methodological basis for want of determination of sampling rate in MIS, though in the field of communication this method is effective enough.
3. A quantum criterion of readout [42] for want of uniform interpolations results in irregular digitization, and here will not be considered.
4. The essence of a method of definition of sampling rate on a criterion of the greatest deviation consists that in it the interpolational approach will be realized [3,48,37,5747]. Is
used interpolational polynomial of the Lagrange as:
where ;
.
Errors of an interpolation by polynomials of the Lagrange continuous function
, which has on a section
the limited derivatives of the order
, are determined by the residual member:
where
.
From here, from a condition
, where
- the specific allowable error of digitization, is determined
.
For example, for
;
for ;
generally: , where
factor dependent from
and
, where
- parameter of quality of an approximation.
Main advantage of this method is that it links immediately interval between readout allowable error and parameter of dynamics of function - maximum significance of a derivative of function.
Defects are the following:
-
The method does not allow to determine stochastic performances of digitization;
-
The optimum degree of an interpolating polynomial is not determined;
- The
method not point concerning function differentiable ad infinitum.
6. Except for the indicated methods there are also other [38,39,58], which are variants of the considered above methods. The method is known[57,59] determination of sampling rate on a middle square deviation. The method requires determination of autocorrelated function and stationarity of process. Is close to a Æåëåçíîâ method . In work [38] the method is indicated which uses the approach on a method of the greatest deviation and approaches used in frequent criterions. As a whole defects both stochastic, and frequent methods are inherent in it. The attempts were undertaken [39] to determine an interval of digitization with application of various polynomials of the Lagrange, with use of stochastic properties of function as autocorrelated function. As well as previous works does not introduce essential novelty and has that defect, that does not determine an optimum interpolation, requires a stationarity of process, and also other defects of the methodical plan.
Thus, the review of existing methods shows, that in the present time in the known literature is not resulted of methods of the substantiation of sampling rate, of its stochastic performances and interpolation, connected to it, and extrapolation, there is enough effective for determination of MIS characteristics .(Site "Through thorns to stars". Anatol Grigorenco)
Copyright©2001 Feedback