JPS6288936A - Randam sampling device - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a new and improved method and method for randomly selecting a predetermined number of samples of an article from a preselected batch or population of articles. Regarding equipment. The article being monitored may be one that is manufactured in quantity and in which it may be desirable to inspect the level of quality from time to time.

For purposes of disclosure, the term "batch or population" is employed to refer to a whole group of articles meeting a particular specification, e.g. The terms "random and sample" are used to refer to objects selected for a batch or population, and the term "random and sample" refers to a sample in which each item in a batch or population has an equal chance of being selected as a sample. means.

[Prior art]

In many industries that utilize mass production techniques, it is customary to remove manufactured articles from the assembly line at intervals as the article progresses down the assembly line. These goods are
Tested to determine its suitability for sale and consumption. Ideally, each article would be tested. However, 100% testing is practically or commercially impractical due to the time and expense that would be unacceptable. Furthermore, it is unnecessary in any case to perform a comprehensive test. This is due to statistically appropriate sampling.
This is because it is unlikely that any problems will occur. Also, if a problem with the manufactured article appears with more than an acceptable level of rats, this is a clear indication that some action must be taken with the gold manufacturing process to alleviate the problem.

Such sampling is particularly important in the food and beverage processing industry to ensure that articles intended for human consumption are suitable for this consumption. however,
Block sampling techniques have been used in almost every industry where a relatively large number of articles are manufactured and it is impractical to inspect the articles on each vehicle.

However, sampling of articles need not only be done to determine the quality of the majority of articles of the entire batch or population on which testing is performed. in fact,
The sampling needs to be random for it to be effective.

As has been said, machinery behaves repetitively, and while this is a desirable property for many functions, it is an undesirable property when the objective is to obtain randomness of behavior. be. Specifically, selecting items at uniform intervals during the manufacturing process may result in a pattern that closely matches a pattern in the population of items and may not be an accurate representation of the final product. . For example, since most filling and packaging equipment operates on a fixed dimensional basis, the selection of sample articles at regular intervals may result in successive samples being filled by the same station on a filling machine or closed by the same station in a sealing machine. You may end up obtaining goods. In this way, if we only focus on one aspect of the manufacturing process,
Problems that may actually exist at other stations with the machine may also be overlooked.

Next, prior art related to the present invention will be discussed. Hrabak et al., U.S. Pat. No. 3,151,237, issued September 29, 1964, discloses a conventional method of the type useful for collecting a certain amount of a particular variable in order to control the quality of that variable. It represents a technological invention. Although this patent is not concerned with the selection of random numbers,
A statistical product control system for determining the statistical distribution of certain quality characteristics of manufactured articles is disclosed.

Wallace et al., U.S. Pat. represent. If there is a problem, the problem will be fixed. That is, the offending items are released. The WallBc6 patent discloses an apparatus for determining the orientation of an article as it travels along a conveyor, correcting the article if it is not properly oriented, and releasing it for another attempt. Similar to the Hrabak patent, W.
There is no discussion or discussion of random number selection in the Allace patent.

Weaver US Pat. No. 4,944.051, issued March 16, 1976, is the closest known prior art to the concept of the present invention. Weaver utilizes a pair of counters that work together with a switching mechanism to achieve a degree of randomness. However, unlike the system adopted in the present invention, W@ave
r seeks to achieve a result in which the random selection is spread evenly over the entire manufacturing process. That is, only one sample is selected from a batch of a predetermined size, thereby making a sample of a preselected percentage of the total at the end of the manufacturing process.

Each time the first counter recirculates, a new batch begins, for example after 50 to 100 articles have passed through the process.

Although Weaver's system achieves a certain level of randomness, in modern industry, especially in the food processing industry,
It is not possible to obtain the degree of randomness that is considered necessary to provide the product quality that today's consumers demand.

The patents mentioned above are to be respected and on the face of it,
Although each patent discloses an advance in the state of the art at the time of grant, the present invention is claimed to be a significant improvement over these known devices. It was precisely in recognition of the need and state of the art that the invention was conceived and put into practice.

[Summary of the invention]

To this end, methods and apparatus are provided for randomly selecting and removing articles from a series of moving articles. The operator selects the total number of batches of articles for which a random selection is to be made, as well as the total number of samples of articles to be taken from that batch. A series of random numbers, each representing the incremental advancement of a series of articles, is generated when the reset button is actuated by the operator. This operation is performed by interrupting a very long sequence of numbers held in the microcomputer's memory and thereby setting a starting point. Sensors are provided to detect the articles as they pass through the scheduled stations, and at Mli1 each article associated with a randomly generated number is removed from the advancing article train. Information relating to the number of batches, the number of sandals, and the number of articles still passing through the scheduled station immediately before the next article scheduled for removal is displayed. The method and apparatus of the present invention can also be used to remove random samples from ongoing substitutability trade flows.

Before elaborating further on dud hi itself, it is deemed advantageous to first set out in general terms the concepts of patch size, sample size, randomness, and the occurrence of dud hi.

Initially, the term "batch" was defined as any group of articles that fit into a particular manufacturing process.

In the extreme, for any manufacturing facility that makes a particular product, the final batch number will be the total number of that product that has ever been or will always be made at that facility. However, for quality control purposes it may not be possible to consider more than a very small proportion within a full-size batch. In particular, information regarding product quality needs to be available quickly enough so that it can influence the outcome of the manufacturing process. Thus, the user will want to limit the batch being tested to a number of articles that is large enough to produce meaningful results, but not so large that the response time becomes unacceptably long.

When trying to set up a crystal control procedure, there is a predetermined, well-known mathematical relationship that allows one to specify what percentage of a product will be defective according to a certain predetermined test. Once the percentage of acceptable units is identified, a sampling procedure can be developed to set the sampling percentage. For example, to take a very simplistic example, if it is desired to obtain a product with no more than 1% defects for shipment to the consumer, then 125 samples are randomly selected from each batch of 1500 articles. selected and tested, failures in no more than five of this group would be considered acceptable. Edited by J. M. Juran, MCGraw.
-Quality Control CI-A/-Handbook published by HIII Book Company
Control Handbook) Refer to standard table MIL-8TD-105D in Part 3.

Of course, it should be understood that randomness is not an absolute concept, and that what should be pursued is an ideal state. Generally speaking, the closer you get to the ideal state, the more expensive it will be.
Ru. The present invention has been successful in obtaining highly relevant randomness at a reasonable price.

Now, once a large number of "random numbers" have been generated, the randomness of the selection can be evaluated in terms of "perfect randomness." For simplicity, if we take the numbers 0 to 9, 1. If they are randomly generated, any number (e.g. 3) will occur at a rate that gives a sum of 1/10, or 10%, of the sum of the numbers produced. . In other words, tower numbers in the range 0-9 have an equal (10%) probability of occurrence. This measured value will be expressed by the following formula.

Here, 0 is equal to the actual number of occurrences of the number I, and e is equal to the expected number of occurrences of the number i. It is usually accepted as a criterion that this formula must be true because it is larger than 27 and smaller than 190. Prent Ice-Hall I
nc, published in 1972Kennedy 1M1cha
el and Solmon z Mart in B.
Eight Statement PL/C (PL/ZERO) Plus P LloNE (EightStatement(
PL/C (PL/ZERO) pILII PVONE
) J6 Teru. The result is a set of 0 to 9 in the vertical table or
On the axis, the number of times a particular number is selected can be indicated in t) IMi by slotting it on the horizontal or Y-axis. If a hypothetical die were rolled for a very long time and the results were plotted, the results would be a flat or horizontal distribution curve.

This is because any of those numbers has an equal chance of being chosen after many choices.

The way the inventor arrived at the random sample of the present invention was to achieve as close as possible the ideal state of die rolling, aiming for independence from multiple influences.

However, as a practical matter, it is very difficult to obtain truly random numbers in simple electronic machines. Very few electrical phenomena are random.

However, they usually pose the danger of being cyclical at some points and at other times. For this reason, it is common to work with pseudorandom numbers.

These are numbers taken from very long sequences of numbers that can be obtained by arithmetic operations. One example is to take the ascending power of 75, modulo 65557, subtract 1, then 6
Divide by 5556.

This approach yields a series of numbers in the range 0≦N<1, which can then be multiplied by the patch size and used to obtain the appropriate sampling number. The example below is a partial representation of this technique (75 of (75)”=75 0
0075 Hit011291 1(75)”=5
625 05625 Qf1858154
43(75)'=421875 28655 1
114371948 219(75)'=31640
625 51791 (17902526395
(75)=2375046875 17642 L
! 691803 135 To complete this example,
Continue operationally for a series of 65,556 total numbers. Number 6
5556 is 16'@ of 2, and this is a convenient number utilized by electronic hardware for computational operations. To elaborate, the term "modulo" is a mathematical term that simply represents a series of successive subtractions of n until Togane becomes smaller than the size of the divisor. Therefore, the numbers in the second column represent the remainder from the numbers in the first column υ, which is the number 65557 from the numbers in the first column.
Continuing the sequential subtraction of , the number becomes smaller than 65537. Of course, in the first two cases, the numbers in the first column are too small to have this remainder. Therefore, one number in the first column of a row has simply been moved to the second column. To further ensure randomness, a number 65537 is chosen which is one greater than 65536 (ie 216). therefore,
This is a prime number, and the resulting remainder is very unlikely to result in a regular repeat number. Next, as an operation presented in the third column of the example, the number 1 is subtracted from the numbers in the second column to ensure that all the resulting numbers are within the range of the computer. In this connection, the number 65536 is 2N
B, and recall that this is a convenient number for electronic hardware.

Since we have adopted power calculations to obtain such a long series of numbers, we can insert the sequence at any point to obtain a suitable set of numbers. Of course, as one can easily imagine, such technology could not be implemented until the advent of relatively inexpensive microcomputers. Now, if we use a single random event, such as the length of time a human presses a reset button, to select a starting point for a sequence, the resulting sequence will be random, and we cannot define acceptable randomness of the selection. Guaranteed to maintain.

Next, the concept of the random sampling device of the present invention is generally as follows. The microcomputer waits for the reset button to be pressed down. During this time, the computer is continuously and cyclically performing, or at least able to utilize, the random number calculations described above. When the button is pressed down, the front panel display is tested to find out how many samples are needed and what size batches are to be sampled from. Once these data are determined, the microprocessor picks up where in the sequence of numbers it happened when the button was pressed and begins listing the desired samples in its memory.

Pseudo numbers are multi-digit decimal numbers in the range 0<N<l, so
The microprocessor multiplies the first number by the number in the batch to obtain the number of samples to be taken. The resulting number will, of course, be the closest total number in the 0-batch range.

This procedure is repeated again until a sufficient number of samples have been accumulated in memory. At that time 1. Kyuji are paid in order of ascending order. Once this is completed, the microprocessor begins counting the articles until the first number of samples is reached.

When the first number of samples is reached, at this point the packages are discharged from the conveyor. The next sample number is then retrieved from memory and the microprocessor maintains the article count until the next sample number is reached. This process continues until all desired samples have been taken and then the microprocessor starts the next batch as programmed or stops until reset by the operator.

Thus, the sampling arrangement of the present invention, similar to the Weaver patent procured above, can perform sampling by selecting one sample from a batch of 50 or (for example) by selecting one sample from a batch of 4973. In selecting 79 samples from a batch, or any other number of samples from any other patch size, within the design, the sample can be performed much more satisfactorily from a statistical point of view.

In addition, the sampling device of the present invention ensures that whatever conditions are on the conveyor are acceptably random.

Users of the sample device of the present invention may also enjoy other benefits. In this regard, it should be noted that many large producers of consumer goods require their suppliers to include "mask samples" in their shipments. This mask sample will be inspected for defects prior to approval for shipment. Typically, during a truckload of 100,000 units, a box of 100 units (cans, bottles or cups) is packed separately. This “Mask Sample J”
are collected manually by the manufacturer according to preset instructions from the supplier. This is an expensive system and difficult to manage. The novel random sampling system of the present invention can easily accomplish this task automatically and in an organized, economical, and accurate manner.
The mask sample can be removed immediately. In this way, producers and suppliers benefit from the invention and can pass these benefits on to consumers, including savings reflected in lower prices.

These and other advantages of the invention will become apparent from the following description with reference to the drawings.

The general description above and the detailed description below are intended to be illustrative only and not limiting.

[Explanation of Examples]

The present invention will be described below with reference to the drawings, with reference to preferred embodiments.

Reference is made to the drawings and in particular to FIG. FIG. 1 generally illustrates a random sampling device 20 embodying the principles of the present invention.

In accordance with the invention, there is provided an apparatus for randomly selecting articles for removal from a train of articles advancing past a scheduled station, comprising the following means. That is, the apparatus includes a first selection means operable to set the total number of batches of articles from which a random selection is to be made and a total number of samples of articles to be taken from the batch. a number generator providing a very long number sequence, and operative to interrupt the number generator externally and set a starting point in the number sequence provided by the number generator; and, in response to operation of the reset means, a series of random numbers within the magnitude of the patch size set by the first selection stage, the number representing the gradual advancement of the column of articles. random number generating means for generating a number; random access memory means for storing in memory the series of random numbers generated by the generating means; and sensing the presence of each article as it passes through a scheduled station. sensor means;
and selecting means operable to remove from the array of articles articles related in turn to each randomly generated number stored in said random access memory means.

Referring to FIGS. 1-3, as embodied here,
The random sampling device 2o is a suitable conveyor 2.
4 is used in conjunction with a series of articles 22 that progress in a conventional manner over the same. For purposes of this invention, when the conveyor is stationary,
It makes no difference whether the articles are moved relative to the conveyor or whether the conveyor itself moves and the articles 22 travel with it. Importantly, a sample station, such as a sample collection container 26, is positioned in contact with the conveyor 24 in some suitable manner and in a suitable position to receive a selection of articles 22. The article 22 is moved with respect to.

As shown in more detail in FIG. 2, a first selector 28 located on the front panel of the main controller 29 serves to set the total number of batches of articles 22 from which a random selection is made. The first selector 28 is illustrated in FIG. 2 as a thumbwheel switch 30, which may be any suitable device for this purpose. Each switch is 1 due to multiple digits.
It represents one digit. Random sampling device 2
Operator 0 sets the first selector 28 to represent the total number of batches of articles for which a random selection is to be made.

Similarly, a second selector 32 is located on the front panel of the control device 29 and is operative to set the total number of samples of articles 22 to be taken from the batch determined by the first selector 28 . It is the operator's choice to determine the number of samples to be taken, and the selector 32 includes a plurality of thumbwheel switches 34 on the dovetail. Thus, each switch represents a single digit of a multi-digit number.

Thus, the operator can use well-known mathematical relationships and standards of quality control to determine how large a sample is desired from the scheduled size, and this information can be used prior to the operation. The random sampling device 20 is loaded.

Referring to FIG. 3, the random sampling device 20 is connected to a microcomputer 36.

This is illustrated as a single element for simplicity. In practice, the microcomputer 56 is a series of interconnected integrated circuits comprising a number generator and an interface between passive elements for receiving information and active elements for performing physical operations. Typical examples of the microcomputer 56 include a Z80-CPU microprocessor integrated circuit and a Z80A-PIO programmable input/output circuit.
The output may be a combination of integrated circuits. These are manufactured and sold by Zllog Inc., California.

As seen in FIG. 3, the first and second selectors 28
and 32, respectively.
8 and 40 to the microcomputer 36.

Further explaining the operation of the random sampling device 20, as mentioned above, the microcomputer is equipped with a number generator that provides a very long number sequence.

The number sequence can be generated according to the above example and may be the number sequence obtained in the third column of the above example. As mentioned above, these numbers are preferably numbers in the range 0 to 1, so that when multiplied by the batch size determined by the operator,
The dummy multiplier is an integer not larger than the intended size of the batch.

The element of the device 20 that actually serves to initiate the operation of the random sampling device 20 is a reset button 42 located on the front panel of the main controller 29 (see FIGS. 2 and 3). The reset button communicates with the microglocomputer 36 via m14s. By operating the reset button 42, the number generator section of the microcomputer 36 is interrupted externally, setting the starting point of the number sequence provided by the number generator. The reset button 42 is actuated as desired by the operator. However, it should be understood that while button 42 can be actuated by a human, it can be actuated by any other suitable external force or mechanism. However, when external forces are applied by some external mechanism, whether mechanical, electrical, chemical or other forms of energy, some of the random behavior guaranteed by biological use can be avoided. I may lose it. Thus, the reset button 42
serves to set a fortuitous starting point for the series of numbers presented in the fifth column of the example above.

The microcomputer 36 also has a reset button 42
A generator is provided for generating a series of random numbers within the patch size set by the 1 selector 28 in response to the operation of the 1 selector 28. Each random number corresponds to a series of items 2
2 γ舊 [Represents general progress. Each random number represents one of the sample articles selected to be removed from the conveyor 24, and when the random numbers are arranged in a column from lowest to highest, each random number represents one of the sample articles selected to be removed from the conveyor 24. represents the next sample object 0 selected in .

The random number generator then represents a microcomputer part that subtracts the third column error of the above example with the number selected for the batch to obtain the fourth column error of the same column.

Random sampling device 20 includes a random access memory 44 that stores the sequence of numbers generated by microcomputer 36 . Memory 44 serves to store transient data generated during operation of sampling device 20. One representative example of memory 44 is manufactured by National 8emiconduct, located in California.
M M manufactured by orCoypoyBtion
5290 type integrated circuit. The memory 44 utilized in the present invention is preferably bidirectional.

That is, it is capable of receiving, recording and retransmitting information. Further, the memory 44 used in the present invention includes a conductor &amp;
145 supplements the limited data storage of the microcomputer 36, which is electrically connected thereto. In detail, memory 44 provides microcomputer 36 with a list of random numbers as described above.

A suitable sensing device (sensor) 46, such as a proximity sensor, may include:
It is operative to sense the presence of each article 22 passing through sample collection container 26 . One example of a proximity sensor is A11en-Bradlsy jjE located in Wisconsin.
As illustrated in FIG. Each item passed 2
Information regarding 2 is transferred to the microcomputer. Random sampling device b! t20 also operates to remove from the advancing series of articles 22 articles related in turn to each randomly generated number stored in random access memory 44. To this end, program memory 52 provides instructions to microcomputer 36 in the operation of selector@50. In practice,
Program memory 52 is a list of instructions for the microcomputer.

An integrated circuit operating in a manner essential to supporting the operation of the present invention in this regard is manufactured by Intel C.
According to the organization, it is designated as a h-built "2716" type. The "2716" type is a purchased blank integrated circuit chip, but is programmed with information provided by the manufacturer or user that allows control of the operation of the microcomputer 36.

In this application, the chip reads the information set on the thumbwheel switches 30 and 34 (by the operator) and the number generator (one of the functions of the microcomputer 36).
and controls the memory of sequences of random numbers. The primary function of the program memory is to determine when the selector mechanism 50 is activated. That is, the number of articles 22 detected by the sensor 46 is counted, and in cooperation with the random number generator of the microcomputer 36, the number of articles 22 detected by the conveyor 24 is counted.
The selector mechanism reaches the specified number of next articles 22 to be removed from the . Once this has been determined, the memory informs the microcomputer 36 via lead 54 and
The drive circuit 56 is activated through the circuit 7. The drive circuit is
A selector ram 58 , suitably positioned on the conveyor 24 in the vicinity of the sample collection container 26 , is actuated to engage the article 22 and eject the article from the conveyor 24 into the container 26 . The selector ram 58 is pneumatically actuated by an air cylinder 60 supplied with air from a suitable compressor 62 via a suitable valve 64. It will, of course, be appreciated that the selector ram may be operated electrically, for example by a solenoid, or hydraulically or otherwise.

The random sampling device 20 of the present invention, as described above, comprises @1 display means for displaying the total number set for the batch of articles for which random selection is to be made in response to the operation of the first selection means. . As shown in the second figure, adjacent to the first selector 28, a numerical display includes:
66 is provided on the front panel of the control device 29. When the thumbwheel switch 30 is actuated by the operator to indicate the number of articles 22 selected for the batch, this information is visually displayed on the numerical display 66 in a suitable manner. be done. There may be some suitable connection between the display device 66 and the selector 28 depending on whether the display device 66 is mechanically, electrically, or otherwise activated. provided. Display device 6
The correlation between the information displayed on 6 and the information loaded into microcomputer 56 is provided by conductor 67.

The random sampling device of the present invention, as described above, comprises second display means for displaying a set total number of samples of articles to be taken from the batch in response to operation of said second selection means. As shown in FIG. 2, a numerical display 68 of the number of samples to be selected is provided to provide visual information correlating to the information previously supplied to thumbwheel switch 34 by the operator. used. Numerical display * As in the case of 66,
A numerical display 68 is suitably connected to the second selector 52 to represent information entered into the system by the operator. Correlation between the information presented on display 68 and the information loaded into microcomputer 36 is likewise provided by conductor 69.

In the random sampling device of the present invention described above, the random number generating means operates to arrange the random numbers in a similar manner. As embodied, a preferred function of the program memory 52 is that the sampling device is programmed to operate with the numbers stored in the random access memory arranged in a sequential manner.

Thus, the microcomputer 36 is activated to energize the drive circuit 56 as each successive article 22 to be discharged into the sample collection container 26 reaches a position relative to the selector mechanism 50.

As mentioned above, the random sampling device of the present invention has the following features:
A fifth display means is provided responsive to said sensor means and said selection means for displaying the number of articles still passing through the scheduled station immediately before the next sample scheduled for removal from the advancing article train. Depending on the embodiment, as shown in FIGS. 2 and 3, a numerical display means 70 is also provided on the front of the control device 29. Display μ
70 is responsive to operation of sensor 46 and sensor mechanism 50 to determine the number of articles remaining before the next sample is selected for removal by selector column 58 as determined by random access memory 44. indicate. Sensor 46 activates display 70 after sampling device 20 enters operation and information is provided to displays 66 and 68.

Thus, the display 70 shows a changing display that subtracts as the article passes and is updated at the next random interval immediately after the sample is removed. For clarity, consider the example below in which a list of random numbers is configured. That is, random numbers such as 107.362.753 are considered.

These were the numbers generated when the reset button 42 was depressed. Until the first article passes sensor 46, display 70 displays "107".

This is the period before the first sampling. After the first article passes sensor 46, display device 70 displays "106
”, and after the second item, it becomes “105”, and so on. When display 70 shows "0", the first sample is removed from conveyor 24, at which point display 70 is updated to show "225". "225
J is the interval to the next sample (i.e. 362-107). When this process is returned ab and the 562nd article is selected to be the second sample, the display 70 will display "391" (i.e. yss-sb2).
has been updated to show the following. In this way, the observer
When a 0 is between the samples being taken, you will be notified that it is truly working, and you can reconfirm that a sample is indeed about to be taken. Of course, the decreasing count value presented on display 70 informs the observer of where the next sample will actually take place.

Program memory 52 cooperates to organize the information output for display readings 70 into numerical sequences at appropriate times. As with displays 66 and 68, the particular type of display is not essential to the invention and may be any suitable type sufficient for purposes of the invention. Similar to displays 66 and 68, the correlation between the information presented on display 70 and the information loaded into microcomputer 36 is provided by conductor 71.

In the random sampling device of the present invention described above,
The selection means includes means for counting, in response to operation of the sensor means, the number of articles removed from the advancing column of articles from the starting point of the patch, and a display for displaying this information in response to the counting means. Have the means. As embodied, microcomputer 56 includes a counter portion responsive to article sensor 46 and selector mechanism 50 to count the number of substances discharged from conveyor 24 into sample selection container 26 . As the process continues, the microcomputer commands the appropriate display device It72 to provide a visual indication of the number of samples already received in container 26. As with displays 66.6 and 70, the correlation between the information presented on display 72 and the information loaded into microcomputer 36 is provided by conductor 73.

The above description was focused on the discussion of random sampling of southern-limited articles 22. However, substitutable substances,
Thus, it will be recognized that 9 samples of material that are not neatly packaged or packaged in boxes, cans, bottles, or other containers are highly desirable. Substitutable substances are usually in particulate or liquid form, ie, of such a nature that one sample or part is indistinguishable from another sample or part. See Figure 4 for sampling of this type of substitutable material.

This FIG. 4 illustrates an apparatus similar to FIG. 1 except that conveyor 24 is shown carrying the citron substitute substance 74 on its way past sample collection container 26. The selector mechanism 50 may be as described above, but a dipper-like member is provided in place of the plunger 58 in order to release the sample mold 78 into the container 26. Additionally, the sensor 46 may be of a nature that senses the gradual advancement of the substitutable material as a function of time from the beginning of the batch in which the random selection is to be made. As another example, the progressive advance may be a function of the cumulative weight of substitutability material after the start of the movement of the patch where the random selection is to be made. As another example, the incremental advance may be a function of the cumulative distance that the substitutable material has moved since the start of the batch movement in which the random selection is made.

In any event, the operation of this random sampling device is similar in all respects to that described above for sampling finite matter utilizing the disclosure presented in FIG. 2, except as noted above.

Although the present invention has been described in detail with reference to preferred embodiments, it is obvious that those skilled in the art can make various changes without departing from the technical idea of the present invention.

[Brief explanation of drawings]

1 is a perspective view of a sampling station along a moving conveyor system utilizing the random sampling method and apparatus of the present invention; FIG. 2 is an enlarged perspective view of the front panel illustrated in FIG. 1; and FIG. FIG. 4 is a schematic diagram illustrating the operation of one embodiment of the invention; FIG. 4 is a diagram illustrating a modified sampling device operative to take random samples of substitutable materials traveling on a conveyor; FIG. FIG. 20: Random sampling device 22: Article 24: Conveyor 26: Collection container 28: First selector 29: Main controller 30: Thumbwheel switch 32: Second selector 34: Thumbwheel switch 36: Microcomputer 42: Reset button 44: Random access memory 46: Article sensor 50: Selector mechanism 52 Niprogram memory 66.68.70.72: Display device FIG, 4°

Claims (8)

[Claims] (1) In an apparatus for randomly selecting articles for removal from an advancing train of articles as the articles pass through a scheduled station, a first step operable to set the total number of batches of articles from which the random selection is to be made; a second selection means operable to set the total number of samples of articles to be removed from the batch; a number generator providing a very long number sequence; reset means operable to externally interrupt the number generator to set a starting point in the number sequence;
random number sequence generating means for generating, in response to operation of said resetting means, a random number sequence that is within the magnitude of the patch size set by said first selection means and each representing a progressive advancement of the article sequence; random access memory means for storing in memory the number sequence generated by the random number sequence generation means; sensor means for sensing the presence of each article as it passes through a scheduled station; an apparatus for random selection of articles, comprising selection means operable to remove articles related in order to a randomly generated number from an advancing column of articles. (2) Random selection according to claim 1, further comprising first display means for displaying a set total number of batches of articles for which random selection is to be performed in response to the operation of said first selection means. Device. (3) The random selection device according to claim 1, further comprising second display means for displaying a set total number of samples of articles to be taken out of the batch in response to the operation of the second selection means. . (4) Claim 1, wherein the random number sequence generation means is operable to arrange random numbers in a sequential manner.
Random selection device as described in section. (5) in response to the sensor means and the selection means;
2. A random selection device according to claim 1, further comprising third display means for displaying the number of articles still to pass through the station scheduled immediately before the next sample scheduled for removal from the advancing article train. (6) A display comprising a counter for counting the number of articles removed from a progressing line of articles since the start of a batch in response to the operation of the sensor means, and a display means for displaying this information in response to the counter. A random selection device according to claim 1, comprising means. (7) In a method of randomly selecting articles for removal from an advancing train of articles, the method includes: (a) establishing the total number of batches of articles for which the random selection is to be made; and (b) the total number of articles to be removed from the latch. Set the number of samples, (c
) Externally interrupting the number generator that provides a very long number sequence to set a starting point in the random number sequence; (e) generating a series of random numbers equal to the total number of samples set in step (b) and each representing a progressive advancement of the article train; and (e) storing in memory the series of random numbers generated in step (d). (f) detecting the presence of each article as it passes through the scheduled station; and (g) relating in turn to each randomly generated number stored in the memory in step (e). A random selection method comprising selecting an item for removal from an advancing line of items. 8. A method as claimed in claim 8, wherein step (e) of storing in a memory the series of random numbers generated in step (d) is carried out in a sequential manner.