EP0158192B1 - Measurement data acquisition method for a plurality of measurement points - Google Patents

Description

In the method according to DE-PS 30 05 746, the measurement data, which are continuously generated at a large number of measurement points, are recorded and processed by a central data processing system in a multi-digit textile machine. The query speed is increased in that several decentralized data processing devices are provided between the central data processing system and the large number of measuring points. Only a limited number of measuring points are assigned to these decentralized data processing devices. The decentralized data processing devices take over the query (scanner) and the temporary storage of the data.

On the one hand, this is a complex solution. On the other hand, this solution increases the query speed and the query frequency. However, there remains the disadvantage that only the instantaneous values of the measurement at the time of the query are recorded. As a result, only random values are determined and evaluated that do not allow a reliable statement regarding the process sequence and the product quality achieved.

DE 31 42 468 A1 describes an independent, portable data receiving unit equipped with a measuring sensor, which is connected to an evaluation unit alternately for data acquisition with any measuring point and then for data transmission by means of a plug connection to be made for this purpose. During the connection to the measuring point, the continuously occurring measured values are tapped and temporarily stored. A data reduction memory is provided, but there is no sensible selection of the data to be stored and transmitted.

A similar measurement results from the essay "Harvey A. Goldstand, SOFTWARE TECHNIQUES FOR A MICROPROCESSOR-BASED DATA ACQUISITION DEVICE", published in a report on the international conference "AUTOTESTCON" on automatic testing, San Diego, USA, Nov. 1978. and evaluation principle for the measuring points occurring on a vehicle.

The invention solves the problem of ensuring continuous quality monitoring of the process at a large number of measuring points with similar measured values which are queried and processed by a data acquisition system at regular time intervals. For this purpose, only the extreme values of the continuously occurring measured values are determined within each query period and the output signal is the The measuring point is available until it is queried by the central data acquisition system. The signals provided can be deleted after being queried and recorded by the central data processing system. By restricting the query to the extreme values that have occurred over time, it is possible to make a complete statement regarding the quality of the process monitored by the measurement.

A further simplification and acceleration of the evaluation of the measurement results can be achieved in that from the outset a certain range of permissible measured values is defined at each measuring point, and error signals are generated when these are exceeded.

A preferred embodiment of the invention provides that the mean value of the measurement results is also determined and made available for periodic polling. The mean value of the measurement results alone and together with the extreme values has a very high value for the quality. For this reason, it is preferably provided that the extreme values are determined as the deviation of the peak values of the measured values from the mean value.

If permissible bandwidths for the fluctuation of the mean value and permissible bandwidths for the extreme values determined in this way as a deviation from the mean value are determined in this method, a further improvement in the reliable, time-gap-free quality statement can be obtained from the only three periodically queried measured values. In this case, the mean value can be made available for the periodic query very simply by providing the current measured values that are obtained via a low-pass filter.

Instead of storing the measured values as analog signals and making them available for periodic polling, as previously provided, in a further development of the invention they can also be provided in digitized form, which results in advantages for further processing.

This is done by dividing the measuring range into measuring stages, generating a permanent output signal for each measuring stage that is reached or exceeded by the continuously occurring measured values at a point, and that the pending output signals are queried at predetermined time intervals by the central data acquisition system and be recorded.

With this method it is also not possible to record the time course of the measured variables. However, it can be recorded which measured variables occurred in the query period. This allows the extreme values of the measured parameter that occurred in the query period to be recorded. This allows a statement about the course of the measured values in the query period. The amount of data to be transferred can be reduced if only the extreme values are saved and recorded during the query.

After query and detection, the output signals are advantageously deleted again by the central data processing. According to this, further information about the course of the measured data and the quality of the process can be obtained by evaluating the measured data of several successive query periods. The length of the query periods can be selected to be very short, since the measurement data are already available in digital form. By shortening the query periods, extreme monitoring and evaluation is practically equivalent to continuous monitoring and evaluation. Nevertheless, the scanning (scanning) frequency can lag behind that which was necessary according to the previous methods of complete measurement data acquisition.

The invention is described below using an exemplary embodiment.

Fig. 1 shows a measurement record of a measured value without gaps, e.g. the thread tension of a running thread 1 in a textile machine. A part of such a textile machine is shown in Fig. 3.

In Fig. 3 four identical processing points of a multi-digit textile machine with the threads 1 are shown schematically. Each thread 1 is conveyed into a stretching zone by godet 2, guided therein via a heater 3 and drawn off again by godet 4. 5 denotes a thread tension (thread tension) meter. In a winding device, each thread is wound into a bobbin 7 by means of the drive roller 6. Each transducer 5 consists of the sensor 8 and a comparator circuit 9, which is shown in detail in Fig. 4. More on this later.

The measured value of the thread tension shown in FIG. 1 will be kept within certain limits without disturbing the process. Ideally, the measured value would be constant. Fluctuations and, in particular, also short-term fluctuations occur.

It can be seen that such an analog measured value signal can only be detected without gaps if a continuous temporal query takes place. In the case of a large number of measuring points, such as those found in machines for processing or producing a large number of threads, in particular in the textile industry, the continuous recording and central processing not only requires a large computing capacity, but also a hardly feasible expenditure on cabling and switching . Therefore, the measuring points are usually queried in series. Because a certain time is required for each query , it follows from this serial query that a continuous measurement of the measured value is not possible, but that a minimum time is required for the query of a measured value, which is not least determined by the required analog-digital conversion. A sufficiently sharp resolution of the measured value through short scanning times is therefore only possible by limiting the number of measuring points connected to a computer. This also means that short-term fluctuations cannot be recorded, even if such fluctuations may be of technological importance.

According to the invention, a certain range of measured values that occur at all or that are to be recorded at all are initially defined. This bandwidth is shown in Fig. 1 by dashed lines. This range of measured values is defined in the application as a measuring range.

In the exemplary embodiment, this measuring range is divided into measuring stages I, II, III ... VIII, which are each of the same size.

With reference to FIG. 2, it is explained how these measured values which occur without gaps are transformed according to the invention and in what form these measured values are made available for polling at time-predetermined intervals ST1, ST2, ST3.

According to the invention, a defined digital output signal AI to AVIII is assigned to each measuring range I to VIII. An output signal AI to A VIII is called when and as soon as the current measured value passes through the corresponding measuring stage. Furthermore, the output signals AI to A VIII, insofar as they have been called up, are stored. The called output signals are therefore always available for interrogation. All output signals are in the time interval T1 A II to A VIII incurred. These output signals are stored and are therefore available for interrogation at query time ST1. From this it can be seen that very large fluctuations in measured values occurred in the query period. The fluctuation range is even greater in the query period T2, while in the query period T3 there was a much smaller fluctuation.

The values stored in a period T1, T2, T3, ... can now be called up by a central computer at time ST1, ST2 and then deleted. It is accepted that there is no gapless acquisition of the measured value curve. However, the extreme values can be recorded without gaps in time. In particular, by reducing the query times T1, T2, T3, ..., an extensive breakdown of the course of the measured values, for example, periodic fluctuations in measured values, temporal course of the measured value scatter, temporal trend of extreme values, etc., can take place. Here, too, the query times T1, T2, T3 are considerably longer than the query times that are necessary in the previously practiced methods in order to be able to detect the extreme values of the analogue measurement value with sufficient reliability. It also avoids the disadvantage that only the measurement data that occurs at the time of the query are queried or complex intermediate data processing systems have to be available, which in turn can also carry out a query at certain time intervals and at certain times and therefore in particular the extreme measured values and the Measured value peaks can only be recorded randomly and therefore with little informative value. A comparator circuit (FIG. 4) is provided as particularly advantageous for dividing the measuring range into measuring stages. Each measurement value is assigned to a defined step sequence of comparison values and one to the respective measurement step assigned output signal is generated when the measured value and the respective stage of the comparison signal meet the comparison criterion, for example, are the same.

Each comparator circuit consists of eight comparators 10, which are connected with their one input 11 to resistors 12 connected in series. The second input 13 is acted upon by the output signal of the sensor 8.

The comparators are designed so that they emit an output signal A I to A VIII as soon as the value at input 13 reaches the voltage value at input 11. If each resistor 12 is dimensioned such that it represents a measuring stage, each output signal A I to A VIII corresponds to a measuring stage I to VIII.

The circuit can also be selected so that the measuring stages overlap slightly. Here, for. B. can be prevented by interlocking the measuring stages that the signals of both measuring stages are stored for a measured value that is in the overlap area of two stages. In any case, only the highest value is saved. For this purpose, the comparators are connected to a switching logic, not shown in detail, by means of which the output signals A I to A VII are deleted as soon as the next higher signal is called.

As further shown in FIGS. 3 and 4, each comparator circuit is connected to a memory device 14 in which the respective output value AI to A VIII is stored. At the beginning of the memory, therefore, the output signals AI to A VIII are present, insofar as they were called up in a query period, and are fed to a parallel / series converter 15. The function of the parallel / series converter is to convert the signals AI to A VIII, which are present in parallel and simultaneously, into a pulse train, which can be fed to the computer 17 via a single line 16.

The query times are specified by the computer 17 (FIG. 3). At each point in time, the computer outputs a coded signal chain, possibly for addressing, via line 19 via a connected parallel / series converter 18, through which one of the memories 14 is addressed and the stored output signals A I to A VIII of a measuring point are queried. After data has been retrieved via line 16, the computer outputs an erase signal coded to the individual memory via line 19, so that the stored output signals of this measuring point are erased and released for the following query period.

It is furthermore proposed as preferred that not the entire detection range of the transducer and also not the entire range of the occurring measured values is evaluated, but that a range of measured values is selected from the occurring measured values that is representative of proper operation. Due to this limitation of the measuring range covered by the invention, measurement values outside the operating range are excluded from the evaluation from the outset, thereby further simplifying or - with the same use of technical means - a refinement of the accuracy within the selected measuring range without using the same technical means .

5 explains the measuring method according to the invention when analog output signals are provided at each measuring point. The abscissa is a common time axis for all diagrams. Some query periods are labeled S1-S4 on the time axis. The ordinate of diagram I shows an example of a possible course of the measuring voltage U of a sensor. The measuring voltage represents the current continuously recorded measured values. Diagram II shows the mean U _{average of} the measuring voltage on the ordinate. According to the invention, this mean value is generated by passing the measured value over a low-pass filter. Diagram III shows the formation, provision and query of the maximum value U _max in its chronological course.

Diagram IV shows in the same way the time course, the formation, provision and query of the minimum value U _{min of} the measuring voltage. The measuring voltage is related to a reference voltage in diagrams II and IV. The voltage zero or the mean value are particularly suitable for this. Extreme value formation and provision is discussed in more detail below:
For example, in the query period S3, the measuring voltage is initially essentially constant in a period a and is equal to the mean value. The extreme values U _max and U _min therefore also correspond to the mean value in this period. In the following time period b, however, there is a sudden and brief increase in the measuring voltage U, as shown in diagram I. Because of the short-term nature of this extreme value, this fluctuation has no influence on the mean value. However, the maximum value U _max follows the increase in the measured value up to the extreme value and then assumes this extreme value. As shown in diagram III, this extreme value then remains during the query period S3 and is queried by the central data acquisition system at the end of the query period and then deleted again, so that the value held ready at the measuring point falls back to the reference value.

In the query period of S3 there are some negative deviations from the mean value and in particular the abrupt and brief drop in the following period b Measuring voltage at c. As diagram IV shows, the minimum value storage U _min also follows this negative deviation of the current measured value U from the mean value, the lowest value being stored in each case and being available for querying at the end of the query period S3.

At the end of the query period S3, the values of the mean value representing the course of the query period as well as the maximum positive and negative deviations therefrom are available for query and evaluation by the central data acquisition system. As an alternative to this, the absolute extreme values can be provided in addition to the mean value.

In this way, for example, the thread tension in a textile machine with a large number of processing points for each thread to be processed can be continuously determined, provided in each query period as maximum, minimum and average value, from the central data acquisition system via a scanner from processing point to processing point polled at intervals and evaluated for a quality statement. Depending on the query values U _medium , U _max and U _min , error alarm signals can be generated. For this purpose, on the one hand, a certain bandwidth is specified for the average thread tension, represented by the output voltage U _medium . If the currently determined mean value leaves this range of permissible thread tensions, an error alarm signal is given and, for example, the processing point concerned is stopped or an error is signaled for this processing point. Likewise, a certain bandwidth is specified for the maximum value as a deviation from the mean value and likewise for the minimum value as a deviation from the mean value. These bandwidths can vary in size.

6 shows a basic circuit diagram for providing the analog extreme values and the analog mean value at each measuring point. The measured value, for example - as in FIG. 3 - the thread tension of a measuring point is continuously detected by the sensor 5. With 14 a memory is designated, which is assigned to each measuring point. The measured values are amplified in the memory and processed to the maximum value U _max , the mean value U _mean and the minimum value U _min and kept ready for interrogation. For this purpose, the memory 14 initially has an amplifier 20 for amplifying the measurement signal. The maximum value is then formed in a peak value meter 21 as a deviation from the mean value and in the inverted peak value meter 22 the minimum value is formed as a deviation from the mean value. 23 shows a low-pass filter for generating the mean value. The other electronic devices in the memory are not identified. Accordingly, the maximum value of the measured values per query period, the running mean value and the minimum value of the measured values per query period are provided at the output of the memory. These values are given to a switching device, a so-called "scanner" 31. Its function is to feed the signals, which are present in parallel and simultaneously, in succession to the computer 17 via an analog-digital converter 32 (A / D converter) and line 16. As already described, the query times are specified by the computer 17. At each time of the query, the computer outputs a coded signal chain for addressing via line 19, through which a measuring point with its memory 14 is addressed and the stored output signals are queried. After data has been retrieved via line 16, the computer outputs a delete signal encoded on the individual memory via line 19, so that the stored maximum or minimum values are deleted. The mean remains. Furthermore, a fault alarm signal device 24 can also be addressed via line 19.

The peak value meters and low-pass filters used for this circuit are known. The peak value meters are explained again using the simplified circuit diagram according to FIG. 7. The peak value meter for determining the maximum value has in particular the diode 25 and the capacitor 26. The capacitor is at a potential of 0 V. A voltage of 0 to 10 V is permitted for the measured value U. Since the diode 25 blocks a current direction, the capacitor 26 is charged up to the maximum value which is reached in the query time, so that this maximum value is stored and remains as the output signal of the query period.

With 27 a switch is designated by which the capacitor 26 is short-circuited. Switch 27 is controlled via line 19 if the maximum value U _max can be deleted after the query.

The inverted peak generator 22 has the diode 28 and the capacitor 29. However, the flow direction of the diode 28 is reversed. The capacitor, on the other hand, is at a voltage of 15 V, which is higher than the highest occurring measuring voltage, which was previously given as 10 V. This means that any reduction in the measured value on the capacitor 29 appears as an increased and permanent voltage gradient, which remains at the output of the memory as the minimum value U _min . 30 again designates a switch by which the capacitor is short-circuited and an erase signal can thus be given via line 19.

The low-pass filter 23 consists essentially of resistors and capacitors that are known.

It should be noted that the circuit of FIG. 7 differs from the description of FIG. 6 and the flow chart shown in FIG. 6 in that the extreme values are represented here by their absolute values. In order to achieve agreement between FIGS. 6 and 7, there is also the possibility of processing the measured values in the memory 14 in such a way that the extreme values represent the difference between the measured extreme values and their mean value. This can be achieved in two ways: on the one hand, an element can be connected between point 38 and voltage follower 33 and between point 38 and voltage follower 34, which element determines the difference between the current measured value and the mean value U _mean . Such a differential amplifier is indicated at 39 in FIG. 8. On the other hand, the voltage followers 35 and 36 in FIG. 7 can be replaced by a differential amplifier which is connected to the output of the differential amplifier 37 and processes the extreme values U _max and U _min so that they represent the difference between the absolute extreme values and the mean value .

REFERENCE SIGN LISTING

1

thread

2nd

Godet

3rd

heater

4th

Godet

5

Sensor, thread tension meter

6

Drive roller

7

Winding up

8th

Sensor

9

Comparator circuit

10th

Comparator

11

entrance

12

resistance

13

entrance

14

Storage

15

Parallel / series converter

16

management

17th

computer

18th

Parallel / series converter

19th

management

20th

amplifier

21

Crest meter

22

Crest meter

23

Low pass filter

24th

Error signaling device

25th

diode

26

capacitor

27th

shoulder

28

diode

29

capacitor

30th

counter

31

scanner

32

Analog / digital converter

33

Voltage follower

34

Tension follower

35

Tension follower

36

Tension follower

37

Tension follower

38

Point

39

Differential amplifier

Claims (12)

1. Process for central acquisition of the continually arising measured values of a plurality of measuring points which for communicating and evaluating measured values are connected successively and at regular time intervals to a central data acquisition and/or data processing unit,
   characterised in that
   at each measuring point during each time interval the maximum values and minimum values of the continuously measured values are acquired and are each stored as an output signal,
   and that at the preset short time intervals the stored output signals are scanned and acquired by the central data acquisition unit.
2. Process according to claim 1,
   characterised in that
   the stored output signals are erased after scanning and acquisition by the central data processing unit.
3. Process according to claim 1 or 2,
   characterised in that
   a limited band width of measured values is selected as a measuring range from the occurring measured values or from the acquisition range of the measurement transducer.
4. Process according to claim 3,
   characterised in that
   exceeding the selected measuring range is signalled as an error.
5. Process according to one of the preceding claims,
   characterised in that
   a mean value is continually formed from the measured values, stored and scanned.
6. Process according to claim 5,
   characterised in that
   the extreme values are formed through comparison with the mean value and preferably in that an error signal is supplied when the extreme values thus formed exceed a variation preset as the maximum allowable.
7. Process according to claim 5,
   characterised in that
   the extreme values occurring within a scanning period are determined as the maximum variation of the measured values from the mean value and are stored and scanned.
8. Process according to one of claims 5 to 7,
   characterised in that
   the mean value is determined from the currently measured data continually and independently of the scanning periods.
9. Process according to claim 8,
   characterised in that
   the mean value is formed by supplying the continually occurring current measured values by way of a low-pass filter.
10. Process according to one of claims 1 to 9,
    characterised in that
    the extreme values are determined by measuring the peak values of the current measured values.
11. Process according to one of the preceding claims,
    characterised in that
    the entire measuring range of each measuring point is divided into measuring steps (I to VIII),
    that at each measuring point, for measuring steps which are overreached or exceeded by the occurring measured values of a point, an output signal (A I to A VIII) is produced and stored and held in readiness for scanning.
12. Process according to claim 11,
    characterised in that
    step-wise increasing reference values are supplied to a respective comparator and are compared therein to the mean value,
    and that the respective comparator supplies an output signal when the measured value at least approximately reaches or exceeds the reference value.

Images