JPH057654B2 - - Google Patents

Description

[Detailed description of the invention]

(Industrial Application Field) The present invention relates to a method and apparatus for non-destructively testing the internal pressure of sealed containers such as canned goods, and in particular cans housed in cardboard boxes can be inspected from the outside without opening the box. TECHNICAL FIELD This invention relates to testable methods and devices. (Prior art) Canned beverages and food, except for special cases such as carbonated drinks, are generally filled under reduced pressure at a gauge pressure of about 30 cmHg to extend the shelf life of the inner container. If there is a leak through a hole, the vacuum will drop or the tank will become pressurized to positive pressure due to the inflow of outside air or the generation of carbon dioxide or hydrogen gas due to deterioration of the contents. Something happened. Therefore, a normal vacuum can has a concave lid surface, while a defective can has a flat or convex surface. Conventionally, various methods of non-contact detection of such expanding cans while a single can is transported on a conveyor have been put into practical use, and at the same time, methods for detecting cans that have been packed in boxes have also been used. The possibility of testing as is is disclosed in Japanese Patent Publication No. 59-25170. This conventional method measures the degree of unevenness of the can lid as a change in the inductance of the electromagnetic coil.
Since it is a magnetic material, it has no electromagnetic properties, so it is possible to inspect cans stored in cardboard boxes. The principle of this measurement will be explained with reference to FIG. An eddy current distance meter 3 is placed at the closest position on the lid surface 2a of the can 2 being conveyed straight on the conveyor 1, and the distance from this distance meter 3 to the lid surface is measured at multiple points A, B, and C. , distance b from the center B of the lid surface, and distances a and c from points A and C that are spaced apart by a certain distance in opposite directions from the center B, and the degree of unevenness of the lid surface is h, and the arithmetic processing unit 4 Then, calculate h = b - a + c / 2 ... (1). The larger h is, the greater the depressurization is, and this is determined to be a good product. If h is small, it is determined to be a defective product, and the next sorting is carried out. It emits an output signal d to the process. The positions of A, B, and C in these calculation processes are determined by the photoelectric sensor 5 that detects that the can 2 has come to the measurement position.
and the signal from the conveyor movement amount sensor 6,
It is done on a fixed basis. Furthermore, even in the case of cans housed in cardboard boxes, the outer edge of the box is detected by a photoelectric sensor, and the position of each can within the box is not detected. Therefore, if there is a disorder in the arrangement of cans in the box,
Errors occurred due to deviations in measurement positions for individual cans, and there was a large risk of misjudgment. (Problems to be Solved by the Invention) Now, in order to more specifically understand the necessity of the present invention, we will explain the arrangement of cans inside a box that occurs when the degree of unevenness of can lids is measured by the conventional method. The situation of measurement errors due to disturbances will be explained. FIG. 4 shows a schematic diagram of the arrangement of cans seen from the top of a wrap-around cardboard box, which is most commonly used for packaging canned goods. The thickness of the cardboard is approx.
2.5mm, and the length of the inside flap 7 at the glue stop at the top and rear end of the box is only equivalent to one and a half cans, so when packing in 5 rows, 3 of the rows are made of cardboard in the front and back directions. Looseness equal to one sheet of paper 8,
8', the can can move that much,
This will cause a deviation in the measurement position. In other words, since the box remains closed and the measurement timing is determined at the tip of the box without knowing the position of the can inside, this looseness remains as it is.
This will cause a 2.5mm deviation in the measurement position. Furthermore, since the hardness of corrugated paperboard tends to change depending on the humidity, the disorder in the arrangement of cans in the box is also influenced by the climate. Diameter 53 used for fruit juice drinks, coffee drinks, etc.
Using a steel lid with a thickness of 0.22 mm, and setting the measurement points A, B, and C at intervals of 13 mm, the error due to the measurement position deviation in the calculation result of the degree of unevenness using equation (1) is calculated based on the gauge pressure of a normal product of 35 cmHg. We investigated the defective product and the defective product with a gauge pressure of 0 cmHg. The results are shown in FIG. The horizontal axis shows the magnitude of the deviation of the measurement position, and the vertical axis shows the value of h calculated using equation (1) using the output voltage of the arithmetic processing unit 4 in mV units. Curve e represents a normal product, and curve f represents a defective product. As is clear from the relationship between these two curves, if a defective product has a measurement position of ±4 mm, there is a risk of erroneously determining it as a normal product. In addition, intermediate defective products that do not reach a completely non-vacuum state and have a gauge pressure of about 15 cmHg follow a value between the two curves. It is clear that with borderline defective products like this, there is a risk of misjudgment due to a measurement position deviation of ±3 mm. The object of the present invention is to provide a method and apparatus for testing the internal pressure of a sealed container, which can perform accurate measurement and discrimination even when there is a disorder in the arrangement of cans in the box, and which reduces the risk of misjudgment. be. Another object of the present invention is to provide a method and apparatus for testing the internal pressure of sealed containers arranged at irregular intervals without opening the boxes. Furthermore, another object of the present invention is to provide a method and apparatus for inspecting the internal pressure of a sealed container, which can inspect containers of different sizes without adjusting position detection. (Means for Solving the Problem) The sealed container internal pressure inspection method of the present invention provides a method for inspecting the internal pressure of a sealed container at a point in the center of the lid of the sealed container that passes directly under the distance sensor.
When each point reaches directly below the distance sensor, the distance from the distance sensor to each point is measured at three points: a point and two points spaced apart by a certain distance with one point in the center as a symmetrical point. , in a sealed container internal pressure inspection method that measures the degree of unevenness of the lid surface based on the difference between the average measured value at two distant points and the measured value at one point in the center, the measurements at the three points are considered as one set. Measure the distance multiple times at slightly different positions on the line including the three points above, select the minimum value among the degrees of unevenness calculated for each set, and compare the minimum value with the preset normal internal pressure. The above object is achieved by a configuration characterized in that the internal pressure is determined to be good when the minimum value is larger than the standard value by comparing the degree of unevenness of the sealed container with a standard value. . In addition, the sealed container internal pressure testing device to realize such an inspection method is placed at a fixed position above where the approximate center point of the sealed container lid surface of the conveying device that transports the sealed container passes, and is connected to the measurement start signal. Therefore, there is a distance sensor that continuously detects the distance to the lid surface of the sealed container at regular intervals, a register that temporarily stores the distance information that is sequentially input from the sensor, and a register that temporarily stores the distance information that is sequentially input from the sensor, and a distance sensor that continuously detects the distance to the lid surface of the sealed container at regular intervals. A selection circuit that selects and reads out a set of measured values at three separate points, an arithmetic circuit that calculates the degree of unevenness from the read out set of measured values, and an output of the arithmetic circuit. a register that selects and temporarily stores the minimum value within the range, a numerical value setter that sets the degree of unevenness of a sealed container with normal internal pressure as a reference value, and a minimum value stored in the register that selects and temporarily stores the minimum value. and a reference value set in the numerical value setter, and if the minimum value is larger than the value, the internal pressure is determined to be good. (Function) By the way, in Fig. 5, the error that occurs due to the measurement position deviation is in the direction where the value of h is larger, regardless of the direction of the deviation, for both good and defective products, and when there is no deviation, It can be seen that it is the minimum value. Therefore, by performing measurements many times while consciously moving the measurement position little by little, calculating the value of h from the measured values many times, and selecting the smallest value among them, the measurement position deviation can be corrected. A value of h corresponding to the case without is obtained. The present invention is based on the above-mentioned idea and automatically calculates the degree of unevenness at the center of the can lid to inspect the internal pressure of the can. For continuous execution like this, one set of distance sensors installed at positions corresponding to the three points to be measured is used to obtain one set of measured values, and one set of measured values is obtained at regular intervals as the relative movement with the container moves. It is also possible to obtain one set of measured values and perform calculation/judgment. However, such a method requires three distance sensors and complicates the equipment, so one sensor measures at regular intervals and one set of three measurement values is selected from among the measured values. It is more convenient and practical to perform calculation discrimination using Although it is possible that the can arrangement within the cardboard box may be misaligned in the lateral direction, in reality, the conveyor that conveys the cardboard box is provided with a fixed guide rail along the conveyor, so the cardboard box may be misaligned during transportation. The box itself does not shift horizontally, and the structure of a cardboard box, such as a 30-piece box normally used for beverage cans, has a single horizontal board and does not loosen, as shown in Figure 4.
Lateral displacement of the can is difficult to occur and can be ignored in practical terms. Therefore, by considering only the deviation in the transport direction, it is possible to perform a practically accurate internal pressure test. (Example) Hereinafter, an example of an apparatus for determining pass/fail by selecting the minimum value from a large number of measured and calculated values will be shown, and the sealed container internal pressure testing method of the present invention will be specifically explained. Embodiment 1 FIG. 1 shows a case where a single can is inspected, and the can 101 to be inspected is standing on the conveyor 102 and the arrow 1
It is transported in the direction of 03. Lid 101 of can 101
The eddy current sensor coil 10 is located near the upper value of a.
4 is arranged, and a preamplifier 105 obtains a voltage signal _S1 of 0 to 10 V for a distance of 0 to 10 mm from the can lid. When the photoelectric sensor 106 detects that the can has arrived at the fixed position, the pulse transmitter 107 attached to the drive roller shaft of the conveyor generates pulses every 0.4 mm of conveyor movement, and the timing circuit 1 is activated.
From 08 onwards, the data acquisition command signal S ₂ is generated while substantially the entire diameter of the can lid passes under the eddy current sensor 104 . The voltage signal S ₁ is converted from analog to digital converter 1
09, each pulse of the command signal _S2 is converted into a 12-bit digital value _S3 . The digital value _S3 is stored in a memory block _110M0 in which 96 12-bit memory units are arranged.
are temporarily stored sequentially from _M95 . That is, the shape of the can lid 1011a is set to 0.4 in the memory block 110.
This means that a section of 38mm is memorized with a mm pitch. Using this temporarily stored measurement value, the degree of unevenness h in equation (1) is calculated for multiple points at the center of the lid surface. First, point A in Fig. 3 is M ₀ , and point B, which is 13 mm away from M ₃₂ , corresponds to a memory unit of 0.4 mm, so to be precise,
The points are 12.8mm apart. Furthermore, read M ₆₄ as point C at the same distance using switches SW ₁ , SW ₂ , and SW ₃ ,
The digital values S ₄ , S ₅ , and S ₆ are calculated by substituting them into a, b, and c of equation (1). That is, the adder 111 calculates the sum of S ₄ and S ₆ , and the divider 112 divides the sum by the number 2.
Calculate the second term on the right side of equation (1) to obtain S ₇ . The first subtractor 113 completes the subtraction of (S ₈ −S ₇ ), that is, the calculation on the right side of equation (1), and the result corresponds to h on the left side.
Get _S8 . The calculation processing up to this point is the same as the conventional technology,
All calculations are performed using 12-bit digital values. Next, minimum value selection, which is a feature of the present invention, will be explained. First, before starting the calculation of the above equation (1), all bits are given a binary code of "1" as the minimum value that can be stored in the memory 114 for storing the minimum value, and the output signal of the memory 114 is S ₉ and h above
The subtracter 115 calculates (S ₉ −S ₈ ) between the signal S ₂ of the calculation result, and if the result is positive, the output S ₁₀ of the subtracter 115 becomes a logical value "1" and opens the gate 116. Signal S ₈ passes through the gate (S ¹¹ ) and enters memory 1
Rewrite the contents of 14. If it is negative,
_S10 becomes "0", gate 116 remains closed, and the memory is not rewritten. Once the series of arithmetic processing up to this point has been completed, the switches _SW1 , _SW2 , and _SW3 are switched to _M1 , _M33 , and _M65 , respectively, and the same arithmetic processing is repeated, and the memory units are switched one by one in the same way. The same process is carried out one after another up to the last M ₃₁ , M ₆₉ , and M ₉₅ . The end of these calculations is based on the distance data to the first can lid.
The switch 117 is actuated in response to the reading end signal _S12 indicating the end of the reading of 96 pieces. Therefore, when the changeover switches SW ₁ , SW ₂ , and SW ₃ reach the memories M ₃₁ , M ₆₉ , and M ₉₅ and all processing is completed,
The minimum value among them is stored in the memory 114. When the arithmetic processing for the contents of the memory block 110 is completed as described above, the data read end signal _S13 is sent from the switch 117, and the gate 1118 is opened to open the memory block 110.
Read the minimum value which is the content of and obtain the signal _S14 .
This value corresponds to h, which is the degree of unevenness of the lid surface without deviation of the measurement position. The subsequent processing is the same as in the conventional technology, and the numerical value setter 1 sets a predetermined discrimination level.
The numerical comparator 120 calculates (S ₁₄ - S ₁₅ ) using the value S ₁₅ of No. 19 and the signal S ₁₄ , and if it is positive, it is a good product, so the resulting signal S ₁₆ is set to the logical value "0".
If it is negative, it is determined that it is a defective product and S ₁₆ is set.
1". The defect signal S ₁₆ generates an output signal S ₁₇ for driving the can removal mechanism via a signal delay circuit 121 and a power amplifier 122 that are synchronized with the movement of the can to be inspected on the conveyor. The device of this embodiment has the ability to automatically correct deviations in the measurement position, so even for cans with slightly different diameters, the photoelectric sensor for detecting the can position can be adjusted without adjustment. In the above embodiment, one memory can be used with a capacity of 0.4 on the can lid surface.
This corresponds to a distance of mm, so the switches SW ₁ , SW ₂ ,
The switching range of the memory unit by SW ₃ is 31 each, which means that it sweeps 12.4mm. Therefore, even if we allow for 2.4 mm of misalignment in the photoelectric sensor's installation, a 10 mm misalignment in the measurement position is allowed. In other words, cans with diameters of up to 20mm can be shared. Specifically, with the photoelectric sensor position adjusted for the standard size ``20'' with a nominal diameter of 53 mm, how much does it take to measure the next larger standard size can with a nominal diameter ``211'' and a diameter of 65 mm? It was possible to use it without any adjustment changes. The above two sizes are often used in common on one line by making slight changes to tool parts or adjustments to guides. Therefore, being able to work at even one location without adjustment is a great advantage in terms of shortening the mold change time and preventing troubles such as adjustment errors. Example 2 Figure 2 shows a case where a large number of cans are packed in a cardboard box and the lid is closed, and the internal pressure of the cans placed inside the box from outside the box is tested on a conveyor with the lid closed. Give an example. When the cans are arranged vertically and horizontally, the arrangement parallel to the direction of conveyor travel is defined as a "column", and the arrangement perpendicular to the direction of conveyor is defined as a "row". Specifically, there are 6 rows and 5 columns, a total of 30 cans.
Explain about canned box packaging. In the figure, the box 201 consists of cans C ₁₁ to C ₆₅ in 6 rows and 5 columns.
The cans are lined up and packed as shown in 201' showing the inside of the box. (The suffix for can No. C is such that the 10th digit is the row number and the 1st digit is the column number.) The box 201 is moved to the arrow 203 by the conveyor 202.
It is transported straight in the direction of. Similar to the number of cans in one row, five eddy current type distance sensors 204-1 to 204-5 are placed on the line connecting the centers of cans in each row, and placed at the nearest fixed position above the boxes as in the previous embodiment. It is located in Rangefinder dedicated amplifiers 205-1 to 205-5, and analog/digital converters 209 to 1 to 209
-5 are connected to the sensors in pairs, and their respective functions and roles are the same as in the embodiment described above, and the same operation is repeated for each can in each row. However, it is not possible to accurately detect cans from outside the box for each row within the box, and only the front and rear ends of the box can be detected. Therefore, the timing circuit 208 requires some ingenuity. First, the photoelectric sensor 206
When the top of the box is detected, 96 pieces of distance data between the distance sensor and the can lid for cans _C11 to _C15 in the first row are read, and the data is stored in memory blocks 210-1 to 210-5.
The details will be described later, but these 1
Calculation of formula (1) for rows of cans is calculation block 21
1, minimum selection is performed in minimum selection block 212. However, since the second row of cans cannot be detected, a pseudo can detection signal is generated in the timing circuit 208 when the conveyor advances by the diameter of one can. That is, using pulses every 0.4 mm from the pulse transmitter 207 attached to the drive shaft of the conveyor, a pseudo can detection signal is generated 6 times every 133 pulses from the box tip detection, and each row is generated for each pseudo can detection signal. 96 data reading pulses at 0.4mm pitch for cans
_S202 is issued to read and store the distance data, and the process is repeated up to the sixth and final line. Now, the distance data for each row is simultaneously stored in parallel in the memory blocks 210-1 to 210-5, but the calculation of equation (1) is performed in one set of calculation blocks 210-1 to 210-5.
11 are shared. Therefore, the changeover switch
SW ₁₀₁ , SW ₁₀₂ , SW ₁₀₃ are for the first row of cans.
When the calculation of formula (1) is completed, the memory block 210-2 in the second row is used to calculate the cans in the second row.
can be switched to Similarly, according to the command from the switch switching control circuit 213, the memory block 210-5 in the last column, the fifth column, is switched, and the value of h in equation (1) for the five cans in the first row is calculated. A selection is made of the minimum value among them. These switching and arithmetic processing are performed between the end of reading data on the first line and the start of reading data on the second line. In this embodiment, the data reading interval is 38 mm, and there is an interval of 15 mm until data reading starts for the next can. If the number of boxes to be inspected is 33 boxes per minute (equivalent to 1000 cans per minute for boxes packed in 6 rows and 5 columns), the above 15 mm interval will be approximately 72 msec in time, but it is difficult to calculate using a normal digital integrated device. One step of processing can be performed in less than 1 μsec, so even if we estimate that it will take 480 μs to read all memory elements, and 100 μs to calculate equation (1), open/close switch gates, and transfer numerical values, the total time will be 1 msec. It doesn't even reach that point, so there's plenty of time. In this way, the minimum value of the calculation result of equation (1) for all the cans in the box is stored in the minimum value memory 214. When the calculation for the can in the last row and fifth column, that is, the last can, is completed, the minimum value read signal S ₂₀₃ is transmitted from the switching circuit 213, the gate 215 is opened, and the contents of the minimum value memory 214 are read as S ₂₀₄ . Read out. The subsequent methods of determining pass/fail and rejecting defective products are almost the same as in the previous embodiment, and the signal S ₂₀₅ of the value set in the numerical value setter 216 and the signal S ₂₀₄ are compared by the numerical comparator 217, and the defective products are In this case, "1" is given to the defective memory 217. Finally, when the photoelectric sensor 206 detects the rear end of the box, the timing circuit 2
A box rear end signal S ²⁰⁶ is issued from 08, the gate 219 is opened, and the contents of the defective memory 218 are allowed to pass through.
This signal is passed through the delay circuit 220 and the power amplifier 221 in the same way as in the previous embodiment, and is used as a signal to drive the exclusion mechanism.
Output S ₂₀₇ . Through all of the above operations, if even one defective product is found among the 30 cans in the box, it will be judged as a defective product box and will be removed.Then, the box will be opened and a manual percussion method or visual inspection will be carried out. Eliminate individual defective cans by observing deformation of the lid surface. As shown in Figure 2, orange beverage cans with a diameter of 53 mm and a content capacity of 250 grams are packed in 6 rows and 5 rows, all of which are normal, non-defective products with a reduced pressure of 35 cmHg, and all of which are pinhole-pierced and have a gauge pressure of 0 cmHg. Table 1 shows the results of measurements of 20 boxes of each type of defective product using the apparatus of the conventional method and the method of the present invention. In the table, the value of h indicates the relative value of the degree of unevenness of the can lid surface according to equation (1), and the values for normal products and defective products for each h indicate the frequency of appearance of that h. In the conventional method, the frequency distribution of normal and defective products overlaps, and in order to completely eliminate defective products, if the discrimination level h is set to 16, 51 cans of normal products will be incorrectly classified as defective. become. In Table 1, the total number of boxes is 20, so in some cases, all boxes may be eliminated. Considering workability, h is set as the pass/fail judgment level.
If it is set to 14.5, there will be almost no erroneous exclusion of good products, but there is a risk that more than 2% of defective products will be overlooked. On the other hand, in the method of the present invention, the frequency distributions of good and bad results are completely separated, and the disadvantages of the conventional method do not occur.

【table】

[Table] As another effect, as shown in Figure 6, 3 rows 5
We were able to inspect two columns of boxes wrapped in heat shrink film and connected to form a total of 6 rows and 5 columns.
In such a box, there are two sheets of cardboard between the third and fourth rows, and the thickness of each sheet is approximately 2.5 mm, so the measurement position of the rear box is approximately 5 mm with respect to the front box.
It will shift backward by mm. In addition, in the case of the box shown in FIG. 2, the front end of the box is made of two overlapping sheets due to the folding of the flap, whereas in FIG. 6 there is only one box. Therefore, when the connected boxes shown in Figure 6 are passed through a device that has been adjusted to an accurate measuring position with respect to the boxes shown in Figure 2, the front box will move 2.5 mm forward, and the rear box will move 2.5 mm backward. This causes a measurement position shift of mm. However, with the method of the present invention, the above deviation is less than half of the allowable range of measurement position deviation, so there is no problem at all, and in a test similar to that shown in Table 1 above, it was difficult to recognize the difference from the box in Figure 2. . (Effects of the Invention) As described above, this invention has the ability to automatically correct the deviation of the measurement position, so even if there is a disorder in the arrangement of cans in the box, accurate measurement discrimination can be performed, and erroneous discrimination can be performed. It is clear from the test results of the above-mentioned embodiments that there is little risk of this, and the internal pressure test of the box sealer can be performed without opening the box. Furthermore, the photoelectric sensor for can position detection can be used for cans of different diameters without any adjustment. Furthermore, in the device of the embodiment, the circuit from the analog to digital converter to the delay of the defective product discharge signal is as follows:
A microcomputer may also be used, in which case the program may be activated by a photoelectric sensor. In the embodiment, the degree of unevenness is judged based on the minimum value, but depending on the shape of the surface to be inspected, the judgment may be made based on the maximum value using a similar method.

[Brief explanation of the drawing]

Fig. 1 is an embodiment of the inspection device of the present invention for inspecting a single container, Fig. 2 is an embodiment for inspecting boxed containers, and Fig. 3 is an explanatory diagram of the measurement principle. FIG. 4 is a plan view showing how containers are packed in boxes, FIG. 5 is a graph showing errors in determining the degree of unevenness due to misalignment of measurement positions, and FIG. 6 is a plan view showing how boxes connected with heat-shrinkable film are packed. 1,102,202: Conveyor, 2,101:
Can, 3,104,204-1 to 204-5: Eddy current distance sensor, 4: Arithmetic processing unit, 5,10
6,206: Can position detection sensor, 6,107.
207: Conveyor movement sensor, 7: Inside flap of container box sealing part, 105, 205-1
~205-5: Preamplifier, 108, 208: Timing circuit, 109, 209-1 ~ 209-
5: Analog, digital converter, 110, 210
-1 to 210-5: Memory block, 111:
Adder, 112: Divider, 113, 115: Subtractor, 114, 214: Minimum value memory, 116:
Rewriting gate of memory 114, 117, 21
3: Changeover control circuit of changeover switch SW, 11
8,215: Minimum value read gate, 119,2
16: Discrimination level setter, 120, 217: Numerical comparator, 121, 220: Signal delay circuit, 12
2,221: Power amplifier for can removal mechanism, 210:
Container box, 211: Calculation block, 212: Minimum value selection block, 218: Defective detection signal memory.

Claims (1)

[Scope of Claims] 1. Each point is defined by three points: one point at the center of the container lid of the sealed container passing directly under the distance sensor, and two points spaced apart by a certain distance with the one point at the center as a symmetrical point. When reaching directly below the distance sensor, measure the distance from the distance sensor to each point, and measure the distance from the distance sensor to each point.
In a sealed container internal pressure inspection method that measures the degree of unevenness of the lid surface based on the difference between the average measured value at a point and the measured value at one point in the center, distance measurement in which measurements at the above three points are set as one set is performed as described above. The measurement is performed several times at slightly shifted positions on the line including the three points, and the minimum value is selected from among the degrees of unevenness calculated for each set, and the result is a sealed container having the minimum value and a preset normal internal pressure. A method for inspecting internal pressure of a sealed container, comprising comparing the degree of unevenness with a reference value, and determining the internal pressure as a good product if the minimum value is larger than the reference value. 2. It is placed at a fixed position above where the approximate center point of the lid surface of the sealed container of the transport device that transports the sealed container passes through, and the distance to the lid surface of the sealed container is continuously measured at regular intervals by a measurement start signal. A distance sensor for detecting, a register for temporarily storing distance information sequentially inputted from the sensor, and a set of measured values of three points spaced apart from each other by a certain equal distance from the stored distance information. A selection circuit that selects and reads out a selection circuit, an arithmetic circuit that calculates the degree of unevenness from a set of read measured values, a register that selects and temporarily stores the minimum value of the output of the arithmetic circuit, and a seal that has a normal internal pressure. A numerical setting device that sets the degree of unevenness of the container as a reference value compares the minimum value stored in a register that selects and temporarily stores the minimum value with the reference value set in the numerical setting device, and determines the minimum value. A sealed container internal pressure inspection device comprising a comparison circuit that determines that the internal pressure is good when the internal pressure is larger than the reference value.