Title of Invention	METHOD AND APPARATUS FOR DYNAMICALLY CHECK WEIGHING
Abstract	Disclosed are a method and an apparatus for dynamically checking the weight of objects (18a-c) which are guided across a weight-sensitive zone (14) of a weighing device (12) at an adjustable conveying rate by means of a conveying mechanism (20a-c). At regular intervals, the weight-sensitive zone (14) supplies individual measured weight values (Ei,.. .,En) from which resultant weight values are derived in a digital evaluation unit (16) by averaging. The evaluation unit (16) comprises a plurality of cascaded averaging filters (24a-e) which have different filter lengths that are varied by a common scaling factor depending on the conveying rate.

Title of Invention

METHOD AND APPARATUS FOR DYNAMICALLY CHECK WEIGHING

Abstract

Disclosed are a method and an apparatus for dynamically checking the weight of objects (18a-c) which are guided across a weight-sensitive zone (14) of a weighing device (12) at an adjustable conveying rate by means of a conveying mechanism (20a-c). At regular intervals, the weight-sensitive zone (14) supplies individual measured weight values (Ei,.. .,En) from which resultant weight values are derived in a digital evaluation unit (16) by averaging. The evaluation unit (16) comprises a plurality of cascaded averaging filters (24a-e) which have different filter lengths that are varied by a common scaling factor depending on the conveying rate.

Full Text	FIELD AND BACKGROUND OF THE INVENTION The invention relates to a method for dynamically checking the weight of objects which are conveyed by means of a conveying mechanism at an adjustable conveying rate over a weight-sensitive zone of a weighing mechanism wherein the weight- sensitive zone supplies individual measured weight values at regular intervals, from which resultant weight values are derived through averaging in a digital evaluation unit. The invention also relates to an apparatus for dynamically checking the weight of objects, comprising a weighing ihechanism with a weight-sensitive zone, a conveying mechanism which conveys the objects at an adjustable conveying rate over the weight-sensitive zone of the weighing mechanism, wherein the weight-sensitive zone supplies individual measured weight values at regular intervals, from which a digital evaluation unit connected downstream derives a resultant Weight value by averaging. Methods and apparatus of this type are disclosed by DE 103 22 504 Al. This document discloses "control scales" and a method for the adjustment and operation thereof. Control scales are understood to be a weighing mechanism to whose weight-sensitive zone objects are fed more or less continuously by means of a conveying mechanism in order to be weighed there. The weighed objects are then transported away by the conveying mechanism and possibly sorted according to the weighing result. A typical field of use of such control scales is the final checking of nominally identical objects. An example thereof is the final filling quantity checking of cans of preserved! food. A fundamental problem of such systems lies in finding a satisfactory compromise between weighing accuracy and weighing speed. Such systems are also typically operated in an industrial environment with severe interfering influences. A typical configuration involves, for example, the feeding of the objects by means of a fast- moving conveyor belt which passes the objects to a separate conveyor belt zone supported on the weight-sensitive zone of the weighing mechanism, said conveyor belt zone subsequently passing the objects, following weighing, to a further conveyor belt section. In systems of this type, the weighing signal is overlaid with significant interfering influences firstly from the movement of the conveyor belt, secondly from the only partial contact of the object with the conveyor belt section supported on the weighing mechanism on entry and exit, and thirdly from other vibrations in the industrial environment. It has therefore proved useful, instead of a single measurement value, to record a plurality of individual measurement values for an object and, by means of suitable averaging, to derive a resultant weight value. In the document cited, averaging is carried out over a particular section of the sequence of individual measurement values. Within the context of a pre-setting procedure wherein a plurality of objects are weighed with varying choice of the section, the optimum position and length of the averaging section is found by "automatic" experimentation. This section selection is then maintained for the subsequent checking operation of the system. A disadvantage of the known method is the lack of flexibility in relation to changes in belt speed. Variations of belt speed frequently occur in practice during industrial operation. They can arise from variations of both technical and personnel-related origin in the conveying rate during feeding of the objects. In order to maintain operation, the conveying rate over the weight-sensitive zone of the weighing mechanism must be exactly matched to the conveying rate for feeding. In the known apparatus, it is necessary to carry out a new pre-setting procedure for every change of conveying rate, and this is associated with significant time and therefore also cost disadvantages. OBJECT OF THE INVENTION It is an object of the present invention to develop control scales of the aforementioned type and a method for check weighing of the aforementioned type in order to ensure better adaptation to varying conveying rates. SUMMARY OF THE INVENTION According to the invention, this is achieved in conjunction with the preamble of claim 1 in that the evaluation unit has a plurality of cascaded averaging filters of different filter lengths and the filter lengths are varied depending on the conveying rate and a common scaling value. The aim is also achieved in conjunction with the preamble of claim 8 in that the evaluation unit has a plurality of cascaded averaging filters of different filter lengths and filter length variation means, which vary the filter lengths by a common scaling value depending on the conveying rate. Particularly advantageous embodiments of the invention are the subject matter of the dependent claims. The features, effects and advantages of the inventive method and the inventive apparatus will now be discussed together. The invention is based on the recognition of essential properties of "averaging filter cascades." An averaging filter cascade is understood to be a series of averaging filters each of which converts a number of sequential input values, determined by the "filter length," into an average value and outputs this value as an input value for the next filter. Essentially two variants thereof are favorable. In a first variant, the sequence of individual measurement values is subdivided into subsections of the filter length and, for each section, an average value is calculated and output. The number of values input into the subsequent filter is decremented relative to the number of values input to the previous filter by a factor depending on the filter length. In the second variant, the average values are each calculated in a moving window over the filter length. This means that the number of calculated average values approximately corresponds to the number of individual values input into the filter. With suitable choice of the filter lengths in the cascade, dominant interference frequencies can be very reliably filtered out. The particular choice of filter lengths is a complex undertaking depending on the individual case, but is fundamentally known to a person skilled in the art. It is acknowledged as a fundamental property of a filter cascade of this type that the underlying form of the impulse response, i.e., the transfer function of the filter cascade is essentially dependent only on the relationships of the filter lengths of the individual filters to one another. Variation of the filter lengths without changing their relationships to one another can change the position and width of the transfer function on the frequency axis, but not the fundamental form thereof. This special property is made use of in the present invention. The invention is based on the fact that, with a change in the conveying rate, the main interference frequencies caused by the conveying motion change accordingly. If, for example, the conveying rate is increased, the interference frequencies are displaced toward higher frequencies. The converse is the case for a reduction in the conveying rate. The invention proposes, in place of a complete resetting of the filter cascade, on a change in the conveying rate, the filter lengths of the cascade should be adapted without changing their relationships to one another. In other words, the filter lengths of the cascade are scaled using a common scaling value. In particular, an inversely proportional dependence of the scaling value on the conveying rate has proved to be a suitable form of dependence. An essential advantage of the method is the rapid and flexible adjustment to conveying rate variations, even if said variations occur only for a short time. This flexible adjustability enables automation of the speed adjustment. To this end, the conveying rate is measured at regular intervals by a speed sensor and a measured speed value is transferred to the evaluation unit for corresponding setting of the scaling value. Speed sensors, which can be included in the control scales for this purpose, are also known to a person skilled in the art, as are the requisite techniques for setting-up the evaluation unit, which can be carried out through the automated programming of digital filters in a data processing system. An important step in preparing the sequence of the inventive method is the initial choice of the filter lengths for a given normal conveying rate. This is usually carried out empirically, since the interference frequencies which need, in each case, to be filtered out are strongly dependent on the individual environment, the normal conveying rate, the object sizes and weights, etc, In order to simplify this empirical setting-up process, it is proposed that in order to make a choice of initial setting of the filter lengths, the individual measured weight values for a representative object are stored in a circular buffer and averaging is carried out on the stored values repeatedly by the evaluating unit and with iteratively varying filter lengths until the resultant weight value agrees with the actual weight value of the object. In other words, this means that the weighing of a plurality of objects with different filter settings is carried out virtually, in that the new uptake of individual weight values of many objects is simulated by the frequent repetition of the individual measured weight values for a single object that are stored in the circular buffer. This simulation with varying filter settings is repeated until the resultant weight value agrees with the (known) actual weight of the object. The concept of "agreement" is naturally understood here to be "agreement within given tolerance rules," wherein the tolerance rules must be adapted to the respective requirements of the individual case. In many cases, starting from a previously tried, or standard, pre-set filter length configuration, it may be sufficient during iterative variation of the filter lengths as part of the adjustment process, to keep the relationship of said filter lengths to one another constant. In other words, this means that in such cases, the adjustment procedure is limited to finding the initial scaling value, in particular, one, from which the conveying rate-dependent variations during operation are derived. In a development of the inventive method, it is provided that the same underlying principle is also applied to the adjustment of the filter configuration to different object sizes and particularly to their extent in the conveying direction. This is achieved in that the filter lengths are varied by a common scaling value depending on the longitudinal length of the objects in the conveying direction. For the purpose of automation, it is favorable if the longitudinal length of an object in the conveying direction is detected by a length sensor and is transmitted to the evaluation unit for corresponding setting of the scaling value. This enables the use of control scales not only for pure monitoring tasks in relation to nominally similar objects, but also for weight-dependent sorting of objects of different sizes. BRIEF DESCRIPTION OF THE DRAWINGS Other features and advantages of the invention will now be disclosed in a detailed description, making reference to the drawings, in which: Fig. 1 is a schematic representation of a set of control scales. Fig. 2 is a schematic representation of the inventive scaling principle. Fig. 3 is a schematic representation of a preferred filter adjustment procedure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Fig. 1 shows a schematic representation of a set of control scales 10. The control scales 10 comprise a weighing mechanism 12 with a weight-sensitive zone 14 and an attached evaluation unit 16. The evaluation unit 16 can be configured, in particular, on the basis of a microprocessor. Control scales of this type typically comprise a display and operating unit, although said unit is not shown in Fig. 1. Objects 18a-c, for which check weighings are to be carried out, are conveyed to and away from the weighing mechanism 12 via a conveyor belt 20, comprising a plurality of sections 20a-20c, The central conveyor belt section 20b in Figure 1 is supported on the weight-sensitive zone 14 of the weighing mechanism 16. It follows therefrom that an object (18b in Fig. 1) situated on the conveyor belt section 20b can be weighed by the weighing mechanism 12. As indicated by the movement arrows 22, the object 18b moves with the conveyor belt in the conveying direction during the weighing. During the period in which the object 18b is situated on the conveyor belt section 20b, the weight-sensitive zone 14 of tine weighing mechanism 12 which can, in particular, comprise an A/D converter, supplies a sequence of n individual measured weight values El, E2,... En. This sequence of individual measured weight values represents a temporally varying measurement signal in which the measurement value depending on the weight of the object 18b is overlaid with signals which are attributable to the aforementioned interference variables. In order to rid the signal of the interference signals, as shown schematically in the upper part of Fig. 2, the signal is passed through a cascade of averaging filters. In the embodiment shown, the filter cascade comprises a sequence of five averaging devices of different filter lengths connected behind one another. Each averaging device comprises a shift register which is able to store a number of results values which corresponds to the filter length. As soon as the register is full, an average value is generated from the stored individual values and is output as a first output value. Each new input value in the shift register pushes out the respective oldest stored value and initiates the calculation of a new average value from the individual values currently stored in the register and then outputs this value as the next output value. The resulting sequence of output average values from the first averaging device is read into the second averaging device which operates on the same principle, but can have a different filter length. In this way, the values run through the entire filter cascade so that at the end thereof, a filtered sequence of resultant weight values, identified as G, is produced. Alternatively, it is also possible to configure the filter cascade such that a single resultant weight value is the result. This can be achieved, for example, in that a sequence of values leaving the last filter is grouped together,, far example, average & in that a single value is selected from the output sequence of the last filter or in that the filter cascade does not operate according to the principle of the "rolling window" described above, but batch-wise, thus reducing the number of values to be passed on in each filter stage. The individual filter stages, which are identified in Fig. 2 with the reference signs 24a-e are shown Schematically as blocks of different lengths, symbolizing their different filter lengths. It has been assumed, in particular in Fig. 2, that the filter lengths of the individual filter stages 24a-e are in the ratio to one another of 3:2:4:5:1 and this is indicated in Fig. 2 by the symbols Z3, 12, IA, E5, £1. Expressed in absolute values, for example, filter length gradations of 12:8:16:20:4 values to be averaged can represent a favorable choice in practice. Typically, the most even possible distribution of zeros in the frequency response is aimed for. The actual further use of the resultant weight value(s) G needs to be adapted to the requirements of each individual case. For example, a target weight of the object 18b can be taken as Achieved if a pre-set weight threshold value is overshot and undershot by a sequence of resultant weight values G a particular number of times. In cases where an individual resultant weight value G is calculated, the weight value can be compared with one or more pre-set weight threshold values and a subsequent sorting system (not shown in the figures) can be switched. The special use of the resultant weight value(s) G is not part of the subject matter of the present invention. The lower part of Fig. 2 shows schematically a mechanism for adaptation to changes in the conveying rate. In the example shown, it is assumed that the conveying rate v accelerates from a Starting conveying rate v0 by a factor 1.25 to v = 1.25 x vO. This speed change is preferably detected by speed sensors (not shown in the figures) and transmitted to the evaluating unit 16. This then changes the filter lengths of the filter stages 24a-e. The change takes place for all the filter stages 24a-e to the same extent, i.e., with the same, preferably linear, dependency, although non-linear dependencies can also be realized. In the example shown in Fig. 2, a particularly advantageous dependency of the scaling value on the acceleration factor of the conveying rate, specifically an inversely proportional dependency, is realized. The result, as shown in the lower part of Fig. 2, is that the lengths of the individual filter stages 24a-e are each shortened absolutely, although their relative relation to one another is maintained. This means that the filter lengths are in the ratio of 3:2:4:5:1 to one another, as before which, given identical objects 18a-c, essentially leads to the same weighing result, symbolized by the resultant weight value G. This is the consequence of the fact that a linear scaling of the filter lengths of a cascade does not essentially change the underlying form of the transfer function of the filter cascade, but only influences the length and width thereof. Fig. 3 shows schematically a preferred method for adjusting a starting configuration of filter lengths in the cascade. For this purpose, the individual measured weight values El, E2,..., En of an object 18b are initially generated in the above described manner and stored in a circular buffer with n storage places. The stored sequence of values is then fed anew into the filter cascade, wherein for each repetition step, the lengths of the individual filter stages 24a-e are varied, which leads to different resultant weight values or sequences of values G, G'/ G", G"1,... In other words, the actual weighing of a plurality of objects 18a, 18b, 18c,--- is replaced by the repeated filtration of the sequence of values measured once for the object 18b. As soon as the resultant weight value(s) of the (known) object 18b have the expected and desired values, the adjustment process can be concluded and the filter configuration found can be adopted for subsequent operation in the manner described above. Naturally, the embodiments discussed in the description and illustrated in the drawings represent only exemplary embodiments of the present invention for illustration purposes. A broad spectrum of possible variants is available to a person skilled in the art in light of the present disclosure. In particular, the number and configuration of the filter stages of the cascade can be adjusted to the individual case. It is also possible to use cascades with different sections, of which only one or a few sections follow the above explained variation principle and one or more other steps remain constant regardless of the conveying rate or the object size. The latter is particularly meaningful if it is known that interfering influences that are independent of weight and size overlay the measurements. WE CLAIM: 1. A method for dynamically checking the weight of objects (18a-c) which are conveyed by means of a conveying mechanism (20a-c) at an adjustable conveying rate over a weight-sensitive zone (14) of a weighing mechanism (12), wherein the weight-sensitive zone (14) supplies individual measured weight values (Ei,...,En) at regular intervals, from which resultant weight values are derived by averaging in a digital evaluation unit (16), characterized in that the evaluation unit (16) comprises a plurality of cascaded averaging filters (24a-e) of different filter lengths and the filter lengths are varied by a common scaling value depending on the conveying rate. 2. The method as claimed in claim 1, characterized in that the scaling value depends on the conveying rate in inversely proportional manner. 3. The method as claimed in one of the preceding claims, characterized in that the conveying rate is measured at regular intervals by a speed sensor and a measured conveying rate value is transmitted to the evaluation unit (16) for corresponding setting of the scaling value. 4. The method as claimed in one of the preceding claims, characterized in that in order to select an initial setting of the filter lengths, the individual measured weight values (El,..., En) of a representative object (18b) are stored in a circular buffer and the average value calculations are repeated by the evaluation unit (16) and carried out on the stored values with iteratively varying filter lengths until the resultant weight values ((G, G1, G", G'") match the actual weight of the object (18b). 6. The method as claimed in one of the preceding claims, characterized in that the filter lengths are varied by a common scaling value depending on the longitudinal extent of the objects (18a-c) in the conveying direction. 7. The method as claimed in claim 6, characterized in that the longitudinal length of an object (18a-c) in the conveying direction is detected by a length sensor and is transmitted to the evaluation unit (16) for suitable setting of the scaling value. 8. An apparatus for dynamically checking the weight of objects (18a-c), comprising a weighing mechanism (12) with a weight-sensitive zone (14), a conveying mechanism (20a-c) which feeds the objects (18a-c) at an adjustable conveying rate over a weight-sensitive zone (14) of the weighing mechanism (12), wherein the weight-sensitive zone (14) supplies individual measured weight values (E1,...,En) at regular intervals, from which a digital evaluation unit (16) derives resultant weight values (G) by averaging, characterized in that the evaluation unit (16) comprises a plurality of cascaded averaging filters (24a-e) of different filter lengths, and filter length variation means, which vary the filter lengths by a common scaling value depending on the conveying rate. 9. The apparatus as claimed in claim 8, characterized in that the scaling value depends on the conveying rate in inverse proportional manner. 10. The apparatus as claimed in one of the claims 8 to 9, characterized in that the apparatus comprises a speed sensor for regular recording of a conveying rate value and for transmitting said conveying rate value to the evaluation unit (16), wherein the evaluation unit (16) is adjusted to change the scaling value accordingly after transmission of an updated conveying rate value. 11. The apparatus as claimed in one of the claims 8 to 10, characterized in that the evaluation unit (16) comprises a circular buffer in which the individual measured weight values (Ei,..., En) of a representative object (18b) can be stored and the evaluation unit (16) is also configured, in order to make a choice of initial setting of the filter lengths, to repeat averaging on the stored values with iteratively varying filter lengths until the resultant weight values (G, G', G", G'") agree with the actual weight value of the object. 12. The apparatus as claimed in claim 11, characterized in that the iterative variation of the filter lengths enables the relationship thereof to one another to remain constant. 13. The apparatus as claimed in one of the claims 8 to 12, characterized in that the filter length variation means vary the filter lengths depending on a linear extent of the objects (18a-c) in the conveying direction by a common scaling value. 14. The apparatus as claimed in claim 13, characterized in that said apparatus comprises a length sensor for recording a length value of an object (18a-c) and the transmission thereof to the evaluation unit (16), wherein the evaluation unit (16) is configured to change the scaling value accordingly each time after the transmission of an updated longitudinal extent value. 5. The method as claimed in claim 4, characterized in that during the iterative variation of the filter lengths, their relationship to one another remains constant.

Full Text

FIELD AND BACKGROUND OF THE INVENTION
The invention relates to a method for dynamically checking the weight of objects which are conveyed by means of a conveying mechanism at an adjustable conveying rate over a weight-sensitive zone of a weighing mechanism wherein the weight- sensitive zone supplies individual measured weight values at regular intervals, from which resultant weight values are derived through averaging in a digital evaluation unit.
The invention also relates to an apparatus for dynamically checking the weight of objects, comprising
a weighing ihechanism with a weight-sensitive zone,
a conveying mechanism which conveys the objects at an adjustable conveying rate over the weight-sensitive zone of the weighing mechanism,
wherein the weight-sensitive zone supplies individual measured weight values at regular intervals, from which a digital evaluation unit connected downstream derives a resultant Weight value by averaging.
Methods and apparatus of this type are disclosed by DE 103 22 504 Al. This document discloses "control scales" and a method for the adjustment and operation thereof. Control scales are understood to be a weighing mechanism to whose weight-sensitive zone objects are fed more or less continuously by means of a conveying mechanism in order to be weighed there. The weighed objects are then transported away by the conveying mechanism and possibly sorted according to the weighing result. A typical field of use of such control scales is the final checking of nominally identical objects. An example thereof is the final filling quantity checking of cans of preserved! food.
A fundamental problem of such systems lies in finding a satisfactory compromise between weighing accuracy and weighing speed. Such systems are also typically operated in an industrial environment with severe interfering influences. A typical configuration involves, for example, the feeding of the objects by means of a fast- moving conveyor belt which passes the objects to a separate conveyor belt zone supported on the weight-sensitive zone of the weighing mechanism, said conveyor belt zone subsequently passing the objects, following weighing, to a further conveyor belt section. In systems of this type, the weighing signal is overlaid with significant interfering influences firstly from the movement of the conveyor belt, secondly from the only partial contact of the object with the conveyor belt section supported on the weighing mechanism on entry and exit, and thirdly from other vibrations in the industrial environment. It has therefore proved useful, instead of a single measurement value, to record a plurality of individual measurement values for an object and, by means of suitable averaging, to derive a resultant weight value. In the document cited, averaging is carried out over a particular section of the sequence of individual measurement values. Within the context of a pre-setting procedure wherein a plurality of objects are weighed with varying choice of the section, the optimum position and length of the averaging section is found by "automatic" experimentation. This section selection is then maintained for the subsequent checking operation of the system.
A disadvantage of the known method is the lack of flexibility in relation to changes in belt speed. Variations of belt speed frequently occur in practice during industrial operation. They can arise from variations of both technical and personnel-related origin in the conveying rate during feeding of the objects. In order to maintain operation, the conveying rate over the weight-sensitive zone of the weighing mechanism must be exactly matched to the conveying rate for feeding. In the known apparatus, it is necessary to carry out a new pre-setting procedure for every change of conveying rate, and this is associated with significant time and therefore also cost disadvantages.
OBJECT OF THE INVENTION
It is an object of the present invention to develop control scales of the aforementioned type and a method for check weighing of the aforementioned type in order to ensure better adaptation to varying conveying rates.
SUMMARY OF THE INVENTION
According to the invention, this is achieved in conjunction with the preamble of claim 1 in that the evaluation unit has a plurality of cascaded averaging filters of different filter lengths and the filter lengths are varied depending on the conveying rate and a common scaling value.
The aim is also achieved in conjunction with the preamble of claim 8 in that the evaluation unit has a plurality of cascaded averaging filters of different filter lengths and filter length variation means, which vary the filter lengths by a common scaling value depending on the conveying rate.
Particularly advantageous embodiments of the invention are the subject matter of the dependent claims.
The features, effects and advantages of the inventive method and the inventive apparatus will now be discussed together.
The invention is based on the recognition of essential properties of "averaging filter cascades." An averaging filter cascade is understood to be a series of averaging filters each of which converts a number of sequential input values, determined by the "filter length," into an average value and outputs this value as an input value for the next filter. Essentially two variants thereof are favorable. In a first variant, the sequence of individual measurement values is subdivided into subsections of the filter length and, for each section, an average value is calculated and output. The number of values input into the subsequent filter is decremented relative to the number of values input to the previous filter by a factor depending on the filter length. In the second variant, the average values are each calculated in a moving window over the filter length. This means that the number of calculated average values approximately corresponds to the number of individual values input into the filter. With suitable choice of the filter lengths in the cascade, dominant interference frequencies can be very reliably filtered out. The particular choice of filter lengths is a complex undertaking depending on the individual case, but is fundamentally known to a person skilled in the art.
It is acknowledged as a fundamental property of a filter cascade of this type that the underlying form of the impulse response, i.e., the transfer function of the filter cascade is essentially dependent only on the relationships of the filter lengths of the individual filters to one another. Variation of the filter lengths without changing their relationships to one another can change the position and width of the transfer function on the frequency axis, but not the fundamental form thereof. This special property is made use of in the present invention.
The invention is based on the fact that, with a change in the conveying rate, the main interference frequencies caused by the conveying motion change accordingly. If, for example, the conveying rate is increased, the interference frequencies are displaced toward higher frequencies. The converse is the case for a reduction in the conveying rate. The invention proposes, in place of a complete resetting of the filter cascade, on a change in the conveying rate, the filter lengths of the cascade should be adapted without changing their relationships to one another. In other words, the filter lengths of the cascade are scaled using a common scaling value. In particular, an inversely proportional dependence of the scaling value on the conveying rate has proved to be a suitable form of dependence.
An essential advantage of the method is the rapid and flexible adjustment to conveying rate variations, even if said variations occur only for a short time.
This flexible adjustability enables automation of the speed adjustment. To this end, the conveying rate is measured at regular intervals by a speed sensor and a measured speed value is transferred to the evaluation unit for corresponding setting of the scaling value. Speed sensors, which can be included in the control scales for this purpose, are also known to a person skilled in the art, as are the requisite techniques for setting-up the evaluation unit, which can be carried out through the automated programming of digital filters in a data processing system.
An important step in preparing the sequence of the inventive method is the initial choice of the filter lengths for a given normal conveying rate. This is usually carried out empirically, since the interference frequencies which need, in each case, to be filtered out are strongly dependent on the individual environment, the normal conveying rate, the object sizes and weights, etc, In order to simplify this empirical setting-up process, it is proposed that in order to make a choice of initial setting of the filter lengths, the individual measured weight values for a representative object are stored in a circular buffer and averaging is carried out on the stored values repeatedly by the evaluating unit and with iteratively varying filter lengths until the resultant weight value agrees with the actual weight value of the object. In other words, this means that the weighing of a plurality of objects with different filter settings is carried out virtually, in that the new uptake of individual weight values of many objects is simulated by the frequent repetition of the individual measured weight values for a single object that are stored in the circular buffer. This simulation with varying filter settings is repeated until the resultant weight value agrees with the (known) actual weight of the object. The concept of "agreement" is naturally understood here to be "agreement within given tolerance rules," wherein the tolerance rules must be adapted to the respective requirements of the individual case.
In many cases, starting from a previously tried, or standard, pre-set filter length configuration, it may be sufficient during iterative variation of the filter lengths as part of the adjustment process, to keep the relationship of said filter lengths to one another constant. In other words, this means that in such cases, the adjustment procedure is limited to finding the initial scaling value, in particular, one, from which the conveying rate-dependent variations during operation are derived.
In a development of the inventive method, it is provided that the same underlying principle is also applied to the adjustment of the filter configuration to different object sizes and particularly to their extent in the conveying direction. This is achieved in that the filter lengths are varied by a common scaling value depending on the longitudinal length of the objects in the conveying direction. For the purpose of automation, it is favorable if the longitudinal length of an object in the conveying direction is detected by a length sensor and is transmitted to the evaluation unit for corresponding setting of the scaling value. This enables the use of control scales not only for pure monitoring tasks in relation to nominally similar objects, but also for weight-dependent sorting of objects of different sizes.
BRIEF DESCRIPTION OF THE DRAWINGS
Other features and advantages of the invention will now be disclosed in a detailed description, making reference to the drawings, in which:
Fig. 1 is a schematic representation of a set of control scales.
Fig. 2 is a schematic representation of the inventive scaling principle.
Fig. 3 is a schematic representation of a preferred filter adjustment procedure.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
Fig. 1 shows a schematic representation of a set of control scales 10. The control scales 10 comprise a weighing mechanism 12 with a weight-sensitive zone 14 and an attached evaluation unit 16. The evaluation unit 16 can be configured, in particular, on the basis of a microprocessor. Control scales of this type typically comprise a display and operating unit, although said unit is not shown in Fig. 1.
Objects 18a-c, for which check weighings are to be carried out, are conveyed to and away from the weighing mechanism 12 via a conveyor belt 20, comprising a plurality of sections 20a-20c, The central conveyor belt section 20b in Figure 1 is supported on the weight-sensitive zone 14 of the weighing mechanism 16. It follows therefrom that an object (18b in Fig. 1) situated on the conveyor belt section 20b can be weighed by the weighing mechanism 12. As indicated by the movement arrows 22, the object 18b moves with the conveyor belt in the conveying direction during the weighing. During the period in which the object 18b is situated on the conveyor belt section 20b, the weight-sensitive zone 14 of tine weighing mechanism 12 which can, in particular, comprise an A/D converter, supplies a sequence of n individual measured weight values El, E2,... En. This sequence of individual measured weight values represents a temporally varying measurement signal in which the measurement value depending on the weight of the object 18b is overlaid with signals which are attributable to the aforementioned interference variables.
In order to rid the signal of the interference signals, as shown schematically in the upper part of Fig. 2, the signal is passed through a cascade of averaging filters. In the embodiment shown, the filter cascade comprises a sequence of five averaging devices of different filter lengths connected behind one another. Each averaging device comprises a shift register which is able to store a number of results values which corresponds to the filter length. As soon as the register is full, an average value is generated from the stored individual values and is output as a first output value. Each new input value in the shift register pushes out the respective oldest stored value and initiates the calculation of a new average value from the individual values currently stored in the register and then outputs this value as the next output value. The resulting sequence of output average values from the first averaging device is read into the second averaging device which operates on the same principle, but can have a different filter length. In this way, the values run through the entire filter cascade so that at the end thereof, a filtered sequence of resultant weight values, identified as G, is produced. Alternatively, it is also possible to configure the filter cascade such that a single resultant weight value is the result. This can be achieved, for example, in that a sequence of values leaving the last filter is grouped together,, far example, average & in that a single value is selected from the output sequence of the last filter or in that the filter cascade does not operate according to the principle of the "rolling window" described above, but batch-wise, thus reducing the number of values to be passed on in each filter stage.
The individual filter stages, which are identified in Fig. 2 with the reference signs 24a-e are shown Schematically as blocks of different lengths, symbolizing their different filter lengths. It has been assumed, in particular in Fig. 2, that the filter lengths of the individual filter stages 24a-e are in the ratio to one another of 3:2:4:5:1 and this is indicated in Fig. 2 by the symbols Z3, 12, IA, E5, £1. Expressed in absolute values, for example, filter length gradations of 12:8:16:20:4 values to be averaged can represent a favorable choice in practice. Typically, the most even possible distribution of zeros in the frequency response is aimed for.
The actual further use of the resultant weight value(s) G needs to be adapted to the requirements of each individual case. For example, a target weight of the object 18b can be taken as Achieved if a pre-set weight threshold value is overshot and undershot by a sequence of resultant weight values G a particular number of times. In cases where an individual resultant weight value G is calculated, the weight value can be compared with one or more pre-set weight threshold values and a subsequent sorting system (not shown in the figures) can be switched. The special use of the resultant weight value(s) G is not part of the subject matter of the present invention.
The lower part of Fig. 2 shows schematically a mechanism for adaptation to changes in the conveying rate. In the example shown, it is assumed that the conveying rate v accelerates from a Starting conveying rate v0 by a factor 1.25 to v = 1.25 x vO. This speed change is preferably detected by speed sensors (not shown in the figures) and transmitted to the evaluating unit 16. This then changes the filter lengths of the filter stages 24a-e. The change takes place for all the filter stages 24a-e to the same extent, i.e., with the same, preferably linear, dependency, although non-linear dependencies can also be realized. In the example shown in Fig. 2, a particularly advantageous dependency of the scaling value on the acceleration factor of the conveying rate, specifically an inversely proportional dependency, is realized. The result, as shown in the lower part of Fig. 2, is that the lengths of the individual filter stages 24a-e are each shortened absolutely, although their relative relation to one another is maintained. This means that the filter lengths are in the ratio of 3:2:4:5:1 to one another, as before which, given identical objects 18a-c, essentially leads to the same weighing result, symbolized by the resultant weight value G. This is the consequence of the fact that a linear scaling of the filter lengths of a cascade does not essentially change the underlying form of the transfer function of the filter cascade, but only influences the length and width thereof.
Fig. 3 shows schematically a preferred method for adjusting a starting configuration of filter lengths in the cascade. For this purpose, the individual measured weight values El, E2,..., En of an object 18b are initially generated in the above described manner and stored in a circular buffer with n storage places. The stored sequence of values is then fed anew into the filter cascade, wherein for each repetition step, the lengths of the individual filter stages 24a-e are varied, which leads to different resultant weight values or sequences of values G, G'/ G", G"1,... In other words, the actual weighing of a plurality of objects 18a, 18b, 18c,--- is replaced by the repeated filtration of the sequence of values measured once for the object 18b. As soon as the resultant weight value(s) of the (known) object 18b have the expected and desired values, the adjustment process can be concluded and the filter configuration found can be adopted for subsequent operation in the manner described above.
Naturally, the embodiments discussed in the description and illustrated in the drawings represent only exemplary embodiments of the present invention for illustration purposes. A broad spectrum of possible variants is available to a person skilled in the art in light of the present disclosure. In particular, the number and configuration of the filter stages of the cascade can be adjusted to the individual case. It is also possible to use cascades with different sections, of which only one or a few sections follow the above explained variation principle and one or more other steps remain constant regardless of the conveying rate or the object size. The latter is particularly meaningful if it is known that interfering influences that are independent of weight and size overlay the measurements.
WE CLAIM:
1. A method for dynamically checking the weight of objects (18a-c) which are conveyed by means of a conveying mechanism (20a-c) at an adjustable conveying rate over a weight-sensitive zone (14) of a weighing mechanism (12), wherein the weight-sensitive zone (14) supplies individual measured weight values (Ei,...,En) at regular intervals, from which resultant weight values are derived by averaging in a digital evaluation unit (16), characterized in that the evaluation unit (16) comprises a plurality of cascaded averaging filters (24a-e) of different filter lengths and the filter lengths are varied by a common scaling value depending on the conveying rate.
2. The method as claimed in claim 1, characterized in that the scaling value depends on the conveying rate in inversely proportional manner.
3. The method as claimed in one of the preceding claims, characterized in that the conveying rate is measured at regular intervals by a speed sensor and a measured conveying rate value is transmitted to the evaluation unit (16) for corresponding setting of the scaling value.
4. The method as claimed in one of the preceding claims, characterized in that in order to select an initial setting of the filter lengths, the individual measured weight values (El,..., En) of a representative object (18b) are stored in a circular buffer and the average value calculations are repeated by the evaluation unit (16) and carried out on the stored values with iteratively varying filter lengths until the resultant weight values ((G, G1, G", G'") match the actual weight of the object (18b).
6. The method as claimed in one of the preceding claims, characterized in that the filter lengths are varied by a common scaling value depending on the longitudinal extent of the objects (18a-c) in the conveying direction.
7. The method as claimed in claim 6, characterized in that the longitudinal length of an object (18a-c) in the conveying direction is detected by a length sensor and is transmitted to the evaluation unit (16) for suitable setting of the scaling value.
8. An apparatus for dynamically checking the weight of objects (18a-c), comprising
a weighing mechanism (12) with a weight-sensitive zone (14),
a conveying mechanism (20a-c) which feeds the objects (18a-c) at an adjustable conveying rate over a weight-sensitive zone (14) of the weighing mechanism (12),
wherein the weight-sensitive zone (14) supplies individual measured weight values (E1,...,En) at regular intervals, from which a digital evaluation unit (16) derives resultant weight values (G) by averaging, characterized in that the evaluation unit (16) comprises a plurality of cascaded averaging filters (24a-e) of different filter lengths, and filter length variation means, which vary the filter lengths by a common scaling value depending on the conveying rate.
9. The apparatus as claimed in claim 8, characterized in that the scaling value depends on the conveying rate in inverse proportional manner.
10. The apparatus as claimed in one of the claims 8 to 9, characterized in that the apparatus comprises a speed sensor for regular recording of a conveying rate value and for transmitting said conveying rate value to the evaluation unit (16),
wherein the evaluation unit (16) is adjusted to change the scaling value accordingly after transmission of an updated conveying rate value.
11. The apparatus as claimed in one of the claims 8 to 10, characterized in that the evaluation unit (16) comprises a circular buffer in which the individual measured weight values (Ei,..., En) of a representative object (18b) can be stored and the evaluation unit (16) is also configured, in order to make a choice of initial setting of the filter lengths, to repeat averaging on the stored values with iteratively varying filter lengths until the resultant weight values (G, G', G", G'") agree with the actual weight value of the object.
12. The apparatus as claimed in claim 11, characterized in that the iterative variation of the filter lengths enables the relationship thereof to one another to remain constant.
13. The apparatus as claimed in one of the claims 8 to 12, characterized in that the filter length variation means vary the filter lengths depending on a linear extent of the objects (18a-c) in the conveying direction by a common scaling value.
14. The apparatus as claimed in claim 13, characterized in that said apparatus comprises a length sensor for recording a length value of an object (18a-c) and the transmission thereof to the evaluation unit (16), wherein the evaluation unit (16) is configured to change the scaling value accordingly each time after the transmission of an updated longitudinal extent value.
5. The method as claimed in claim 4, characterized in that during the iterative variation of the filter lengths, their relationship to one another remains constant.

Documents:

http://ipindiaonline.gov.in/patentsearch/GrantedSearch/viewdoc.aspx?id=cPrnfj0sHr36aHI48FB2/A==&loc=vsnutRQWHdTHa1EUofPtPQ==

« Previous Patent

Next Patent »

Patent Number

271925

Indian Patent Application Number

2414/MUMNP/2009

PG Journal Number

11/2016

Publication Date

11-Mar-2016

Grant Date

10-Mar-2016

Date of Filing

29-Dec-2009

Name of Patentee

SARTORIUS INDUSTRIAL SCALES GMBH & CO. KG

Applicant Address

WEENDER LANDSTRASSE 94-108,37075 GOETTINGEN,GERMANY.

Inventors:

#	Inventor's Name	Inventor's Address
1	KLAUER, ALFRED.	UNTERFELDRING 11, 37083 GOETTINGEN, GERMANY.

PCT International Classification Number

G01G11/04,G01G 19/03

PCT International Application Number

PCT/EP2008/005112

PCT International Filing date

2008-06-25

PCT Conventions:

#	PCT Application Number	Date of Convention	Priority Country
1	102007040301.3	2007-08-24	Germany