SELECTION AND JUSTIFICATION OF PARAMETERS IN FORECASTING AND ASSESSING THE QUALITY OF ASYNCHRONOUS MOTORS

O. P. MURAVLEV, O. L. RAPOPORT

(Presented by the scientific seminar of the Departments of Electrical Machines and General Electrical Engineering)

Systematic improvement of the quality of industrial products is an important condition for increasing the efficiency of social production. The Directives of the 24th Congress of the CPSU emphasize that raising quality is one of the principal tasks of Soviet industry. For electrical machines of serial and mass production, this task is the most pressing, since their quality is connected above all with the reliability of the machines themselves and of the mechanisms driven by them.

Improving the quality of general-purpose asynchronous motors is associated with solving a large number of problems both at the design stage and at the stages of production and use. Among the quality-control problems listed in (1), the following should be singled out as those that must be solved first of all: establishing the existing level of quality, forecasting the level of quality, and quality planning.

Under the condition of a systems approach, solving the stated problems will require a quantitative assessment of quality. Therefore, there arises the need to create a methodology for the quantitative assessment of the quality of asynchronous motors.

Any method of quality assessment depends to a significant degree on the adopted classification of quality indicators. It is necessary to identify certain requirements imposed on the system of indicators. These are, first of all, exhaustive completeness in covering all characteristics of asynchronous motors. If this requirement is considered in practical terms, then, with an infinite set of parameters, it is necessary to select the minimally possible number of them, the system of which would describe the quality of the motors with sufficient accuracy. A second, no less important requirement is the possibility of quantitatively expressing the quality indicators. For asynchronous motors, several types of such classifications may be proposed, relating the parameters according to one or another attribute. Proceeding from this, we considered several systems of parameters combined according to the degree of responsibility for the operation of an asynchronous motor. They are divided into four groups:

Group 1—critical indicators, whose exceeding of permissible values causes motor failure, and in some cases creates a danger to the lives of maintenance personnel;

Group 2—major indicators, whose nonfulfillment sharply reduces the quality of the motors, causing their rapid breakdown;

Group 3—secondary parameters, which affect quality, but a motor with unsatisfactory parameters of this group can operate;

Group 4—low-impact parameters—these are all those whose fulfillment or nonfulfillment has little effect on the quality of motor operation.

The difficulty of using this type of classification lies in the need to determine the weight of individual parameters and in the impossibility of describing many of them quantitatively.

Another type of classification is the division of parameters according to degree of subordination. In this case the general goal is set as follows: improving the quality of asynchronous motors on the basis of eight generalized indicators, including: 1) technical parameters; 2) design parameters; 3) technological parameters; 4) parameters of standardization and unification; 5) reliability parameters; 6) economic parameters; 7) parameters of patent protection; 8) parameters of technical aesthetics and ergonomics.

Their fulfillment, in turn, depends on parameters of the second level; that of the second level, on parameters of the third, and so on. A classification constructed in this way represents a tree-like structure, and it must be considered by separate levels.

The advantage of such an approach is the targeted coverage of the maximum possible number of indicators, with automatic ranking according to the principle of subordination. Such a system can encompass all stages through which the motor passes, from design to production of the finished product.

The difficulty of applying it, as in the first system, lies in the impossibility of quantitatively expressing many indicators. At the same time, to assess the importance of parameters by levels, expert evaluation involving a large number of specialists is necessary.

A common shortcoming of both parameter systems is that they do not make it possible to find relationships among individual parameters or to evaluate them.

An analysis of existing classifications of indicators, as well as of methods for combining them into a comprehensive indicator of the quality level, has shown that at present only a differentiated assessment of quality is possible at various stages from design to consumption. Common to all stages is the establishment of the qualitative level in the present-day development of asynchronous motors, the identification of trends in this development, and the determination of ways to improve quality. On this basis, a generalized assessment of the quality of the motor as a whole can be obtained.

One of the important directions of progress in technology at the present stage of development is minimization: reduction of material consumption and dimensions, and improvement of reliability. These trends are also characteristic of asynchronous motors, since the reliability, dimensions, and weight of the latter affect the analogous indicators of the machines and mechanisms driven by them. They largely determine the qualitative condition of electric motors, the steady improvement of which, characteristic of technology, for them results in a periodic process connected with progress in the metallurgical, chemical, and other branches of industry.

The demand for asynchronous motors and the requirements for their quality outpace the existing state of the art. To meet these requirements, new modifications appear, with increased manufacturing accuracy (high-precision asynchronous motors), increased reliability, and so on. In general, experience shows that the work carried out in electrical machine engineering to improve quality is not sufficiently systematic or comprehensive. The problem of quality embraces a wide range of issues, often so contradictory that it is difficult to choose

select optimal ways of improving it. It is still often possible to observe cases where a considerable increase in expenditures associated with improving parameters is not accompanied by any appreciable effect for the consumer.

In order to carry out measures to improve the indicators of asynchronous motors being manufactured and designed, it is necessary to foresee the effect that should result from this. Forecasting quality becomes an indispensable condition, the basis for planning quality improvement.

The purpose of quality forecasting is to determine rational directions for the further improvement of the quality of electric motors for a definite forecast period of time. It seems to us that the most correct forecast here should be a “feedback” forecast. Such an approach includes direct forecasting, when, on the basis of studying the historical experience of producing asynchronous motors, analyzing the interrelations of their parameters and the tendencies of their development, these are extrapolated into the future. In this way one can obtain forecast estimates of parameters and of the relationships between them. At the same time it is possible to take into account the revolutionary shifts characteristic of the development of asynchronous motors against the background of their monotonic improvement through the perfection of manufacturing technology, optimization, and the use of new materials.

Having a forecast of the type described above and taking it as a goal for a certain point in the future, the so-called normative forecast is carried out, which may be called feedback, from the future to the present. Its essence consists in the fact that, having determined a goal in the future, the most effective ways of attaining it are found, proceeding from the possibilities of contemporary scientific and technical development.

From the accepted sequence of stages in carrying out a forecast, at the present time the task of “passive” direct forecasting is being solved; we shall confine ourselves to considering it. On the basis of a critical examination of a large number of forecasting methods, we have chosen a method based on probability theory and mathematical statistics. The task of probabilistic methods of analysis, which have recently become widely used in technical and other investigations, is to find equations describing the statistical relationships of one parameter with another (in the form of pair correlation) or of one parameter with a group of others (in the form of multiple correlation). The use of probabilistic methods makes it possible, on the basis of studying and generalizing the experience of producing asynchronous motors, to find equations for the relationships among their parameters and, from these formulas, to calculate the most probable values that a certain machine parameter should assume in the future—that is, to forecast their values. Carrying out probabilistic calculations is associated with a definite set of data, which must take into account the specific features of their use for forecasting.

At present, extensive material has been collected for forecasting the quality of asynchronous motors. The largest amount of data has been taken from catalogs and brochures on Soviet series and foreign firms. In all, more than 150 catalogs have been used, as well as all the information available to us from various sources: reviews of the results of the development of electrical machines, standardization materials, reports on scientific research work in electrical-machine building, and works on the prospects for the development and improvement of asynchronous motors by domestic and foreign firms. The catalogs represent 19 countries; they include 46 firms, among them leading firms of the FRG (AEG, Siemens), France (CEM), the USA, Switzerland (Brown-Boveri), and others.

As the initial statistical material for forecasting—

we have adopted parameters reflecting the technical condition of the motors being produced. These are the technical parameters and dimensions indicated in the catalogs.

The advantage of information in the form of technical characteristics is as follows:

1) the physical reality of the characteristic;

2) the realizability of machines with the parameters given in the characteristics;

3) broad publication in informational editions leads to duplication of data and makes it possible to verify them;

4) the manufacturer’s desire to provide the most complete information;

5) the possibility of linking to other forms of information (drawings, diagrams, and others).

The principal advantage of this type of information is its quantitative expression and material realizability. Unlike patents, it describes implemented designs.

Thus, at this stage of the work a simplified-type system of parameters was selected, based on technical characteristics and dimensions. Its advantages are obvious.

The representativeness of the obtained sets of parameters is expressed in the fact that, for the 850 models of asynchronous motors considered, with power from 1 to 10 kW, about 30,000 parameters were obtained.

The first stage of the study is the development of statistical models of the relationships between parameters of asynchronous motors for forecasting possible and most rational values of motor parameters. The search for statistical models is carried out in the form of finding regression equations for the relationships between parameters. Introducing the time factor into them makes it possible to construct dynamic models. On the basis of the latter, general estimates of the forecast relationships between parameters may be given.

In conclusion, it should be noted that quality assessment and its planning for asynchronous motors most naturally follow from quality forecasting.

At the present stage it is advisable to forecast quality by methods of probability theory and mathematical statistics, which have yielded good results in a number of forecasts of products from other branches of industry [3]. Such a method makes it possible to approach the selection of a parameter system unambiguously.

For forecasting the quality of asynchronous motors, a large body of statistical material has been collected, consolidated into special tables and prepared for processing on a digital computer.

REFERENCES

1. O. P. Muravlev, Yu. Ya. Kovalev, A. D. Nemtsev, O. L. Rapoport. Questions of quality management in the manufacture of asynchronous motors. Present collection.

2. L. M. Kostikov. Forecasting the quality of industrial products. Standards and Quality, 1969, Nos. 1, 2.

3. I. P. Kerov. The use of mathematical statistics in processing information on construction and road machines. Moscow, 1969.