JP3333254B2 - Rock judgment method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for judging a rock. More specifically, the present invention relates to a rock judging method which is very useful for judging rock in industries such as civil engineering and mining.

[0002]

2. Description of the Related Art Generally, mining sites such as dam construction, tunnel excavation, bridge construction, etc., and mining sites for resource exploration, resource excavation, etc., generally perform such construction and resource excavation. As the first step, it is necessary to investigate the characteristics of the rocks that make up the rock at the construction site and the site where the resources are excavated. By investigating the rocks that make up the rock mass in this way, the most appropriate construction method and resource excavation method are determined.

[0003] It is essential to determine the name of a rock as a fundamental requirement in investigating the characteristics of such a rock. If the determination of the name of the rock is erroneously made, Substantial defects remain in the subsequent construction and resource excavation, and therefore, the accuracy of rock name determination is required. In general, such a method of determining the name of a rock is performed by an expert who actually goes to a construction site or a resource excavation site, removes the rock from the outcrop with a tool such as a rock hammer, and visually inspects or magnifies the loupe. Observe the characteristics of the rock, such as the color, pattern, gloss, texture, cracking, cracks, hardness, and appearance of the rock, using the tools and feel, etc., and determine the name of the rock from the results based on the results are doing.

However, in such a method of determining a rock name relying on an expert, there is a problem that the presence of an expert is an absolute condition, and no one can easily determine a rock. Therefore, when conducting construction work and resource excavation, etc., only simple tools such as rock hammers and loupes and observation using human senses such as sight, touch, smell, etc. The development of a system that can quickly determine the name of a rock without the assistance of experts has been urgently needed.

The present invention has been made in view of the above circumstances, and solves the drawbacks of the conventional rock judging method that relied on experts, so that anyone can easily and accurately judge the name of a rock. It is an object of the present invention to provide a new rock judgment method that can perform the method .

[0006]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention solves the above-mentioned problem by asking questions indicating the characteristics of rock.
Rock judgment to determine the type of unknown rock based on inference
Is a fixed method, and the question
Rule-based reasoning for less ambiguous questions
And apply fuzzy inference to questions containing ambiguity.
And apply rule-based inference to
Answers to a small number of questions and the confidence of the answers
Is set in advance for each rock type, and for each selected answer
Calculate the final confidence from the answer confidence and use it for fuzzy inference.
Fuzzy rules for questions containing ambiguity
And the membership function corresponding to the fuzzy rule and
Is set in advance for each rock type, and each selected fuzzy
Calculate goodness-of-fit from membership rules function
Then, a representative value of the integrated goodness of fit is calculated from each goodness of fit,
Confidence and fuzzy calculated by rule-based reasoning
Combined with the representative value of the integrated fitness calculated by inference
The rock certainty, and then calculate the rock certainty for each rock type.
The rock type with the highest value
It provides a rock determination method <br/> characterized by determining the type of knowledge rock.

That is, in the present invention, the color index, the pattern, the color tone, the gloss, the texture, the crack, the crack, the hardness, the shape, the size of the grains, the shape of the grains, the crystal, the state of the crystals, For the answers to the question items indicating the characteristics of each rock, such as the color of the vein, the hardness of the white vein, the state of the stripes and layers, the color of the net, etc., the answer certainty expressed as a number from 0 to 1.0 The rock name is determined by the membership function and the rule-based reasoning function and the fuzzy reasoning function.

[0008] This rule-based inference function has an ambiguity.
Used for no small question, fuzzy inference function is used for questions that feel the ambiguity. For example, without the low ambiguity color effects questions rocks, using rule-based reasoning function. However, for a granular pattern of rock, for example, there is ambiguity in selecting sesame salt, granular, agglomerated, and the like, and such inference uses a fuzzy inference function.

[0009] Rule-based inference is an inference method in which the answer certainty given in advance for the answer to each question is calculated using a combine function. For example, question 1
Let a be the answer certainty factor of question 2 and b be the answer certainty factor of question 2, the combine function F (a, b) is expressed by the following equation.

[0010]

(Equation 1)

However, the combine function is "-1.0 to 1.0"
Since the confidence is in the range of “0 to 1.0”, each value is handled by the combine function after conversion by confidence × 2−1. On the other hand, when fuzzy inference is used, the degree of fitness of the membership function is similarly obtained from the fuzzy rules set in the knowledge base, and is synthesized to calculate a representative value. In addition, the values calculated from these inferences are integrated to calculate the rock certainty of each rock.

Further, in the present invention, once the inference is completed, a detailed question may be asked again and the inference may be performed again. By performing this secondary question, the type of rock can be determined more accurately. The present invention also has a knowledge explanation function and / or a knowledge correction function in addition to the inference function. In this case, knowledge modification is allowed only to experts. By doing so, the accuracy of the knowledge can be maintained.

Hereinafter, the present invention will be described in detail with reference to Examples.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS In the following, in the case of targeting rhyolite as an unknown rock, an inference method will be described by exemplifying the contents of a rhyolite knowledge database device. When the rule-based inference function is used, for example, rock classification frames of rhyolite in the knowledge database device can be shown in Tables 1 and 2.

[0015]

[Table 1]

[0016]

[Table 2]

As shown in Tables 1 and 2, the classification frame includes questions such as a rock name, a color index, and a pattern, answers to the questions, and a response certainty of 0 to
Described by a number less than 1.0. For example, a fuzzy inference function is used because there is ambiguity in a question such as a granular pattern of a rock. The fuzzy rules of the rhyolite at that time are shown in FIGS. 1 and 2, the condition part membership functions thereof are shown in FIGS. 3 to 5, and the conclusion part membership functions thereof are shown in FIG.

Such rock classification frames, fuzzy rules, and membership functions are registered in advance for all registered rocks, and inference of the type of unknown rock is performed from these knowledge database devices. . Therefore, an example of the rule-based inference function and the fuzzy inference function in the case of applying the observation result of an unknown rock to rhyolite is shown. (I) Rule-based reasoning For example, if the color index of an unknown rock is observed as a neutral color,
Applying it to rhyolite, the answer confidence is 0.7 from Table 1 above. Similarly, when the color of an unknown rock is observed as white, the answer certainty of rhyolite is 0.7. When these values are subjected to certainty calculation by the combine function described above, the following is obtained.

[0019]

(Equation 2)

Next, when the color tone of the unknown rock is observed as gray as the target answer certainty factor, the answer certainty factor for rhyolite is 0.5. Assuming that the calculation result is a and the answer certainty factor is converted by the conversion method described above and the result is b, the confidence calculation is performed as follows.

[0021]

(Equation 3)

Similarly, in the same manner, the color tone is dark green (rhybite has a certainty of answer of 0.5), and the manner of cracking is none (the rhyolites have a certainty of 0.)
7), Cracks: (rhyolite response certainty 0.5), Hardness: comparable to iron (rhyolite response certainty 0.5), Geography: strata, alternate layers (rhyolite response certainty 0.3) Is calculated, the calculated value is 0.64. From this result, the final certainty factor by the rule-based inference is as follows.

[0023]

(Equation 4)

(II) Fuzzy Inference Function The fuzzy inference function infers using a membership function of conditions and conclusions set by fuzzy rules.
For example, it is assumed that the answers of FIG. 7 are obtained for the questions of the granularity, gloss, and texture of the unknown rock pattern. Applying this to rhyolite, assuming that the fuzzy rule of rhyolite is as shown in FIGS. 1 and 2 as described above, the calculation of the degree of conformity of the conclusion part in each fuzzy rule with respect to this fuzzy rule , And the calculation of the representative value of the integrated fitness are performed in two stages. 8 and 9 exemplify a part of a calculation example of the fitness.

Next, as shown in FIG. 10, the membership function of the conclusion part obtained by each fuzzy rule is synthesized, and the largest value thereof, that is, the center of gravity of the figure connecting the outside of the synthesized function is obtained. And use it as the representative value of the integrated fitness. In this case, the representative value is 0.62. (III) Calculation of rock confidence The representative value (horizontal axis 0.62) obtained by the fuzzy inference described above is
It means the degree of rhyolite likeness and corresponds to the certainty factor in rule-based reasoning.

Therefore, the combination of the two uses a combine function, and converts the certainty factor a = 0.82 obtained from the rule-based inference and the representative value b = 0.62 obtained by the fuzzy inference by the above combine function to obtain the rock certainty factor. obtain. The calculation is as follows.

[0027]

(Equation 5)

This value is the rock certainty obtained from rhyolite. The above processing is performed for all the rocks, and the rock name having the highest rock certainty is determined as the type of the unknown rock.

[0029]

As described above in detail, according to the present invention, anyone can easily and accurately determine the name of a rock without relying on an expert.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram illustrating a fuzzy rule in an embodiment of the present invention.

FIG. 2 is an explanatory diagram illustrating a fuzzy rule in the embodiment of the present invention.

FIGS. 3A to 3E are explanatory diagrams illustrating a conditional part membership function in the embodiment of the present invention;

FIGS. 4A to 4E are explanatory diagrams illustrating a conditional part membership function according to the embodiment of the present invention;

FIGS. 5A to 5D are explanatory diagrams illustrating a conditional part membership function in the embodiment of the present invention;

FIGS. 6A to 6D are explanatory diagrams exemplifying a conclusion part membership function in the embodiment of the present invention.

FIGS. 7A to 7C are diagrams showing answers in the embodiment of the present invention.

FIGS. 8A and 8B are diagrams showing examples of calculation of the degree of fit of the conclusion part in each fuzzy rule according to the embodiment of the present invention.

FIGS. 9 (c) to 9 (e) are diagrams showing examples of calculation of the degree of conformity of the conclusion part in each fuzzy rule in the embodiment of the present invention.

FIG. 10 is a diagram showing a calculation example of a representative value of the integrated fitness in the embodiment of the present invention.

──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Katsumi Fujita 3-2-22-2 Nakanoshima, Kita-ku, Osaka-shi, Osaka Inside Kansai Electric Power Company (72) Inventor Katsura Urayama 1-shimanouchi, Chuo-ku, Osaka-shi, Osaka 20-19 Inside New Jec Co., Ltd. (72) Inventor Masayuki Yamada 1-20-19 Shimanouchi, Chuo-ku, Osaka City, Osaka Prefecture Examiner, New Jec Co., Ltd. Kazuhiko Yoshimura (56) References JP-A 1-137011 (JP, A) JP-A-1-313723 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) E02D 1/00

Claims (6)

(57) [Claims] Claims: 1. An inquiry for a question characterizing rock characteristics.
The rock judgment method that judges the type of unknown rock based on the theory
And the less ambiguous quality of the question that characterizes the rock
Apply rule-based reasoning to questions
Apply the fuzzy inference for the questions, including, in the rule-based reasoning, little question things ambiguous
For each rock type
Is set in advance, and is the answer certainty of each selected answer
Calculates the final confidence level for fuzzy inference,
Fuzzy rules and members corresponding to the fuzzy rules
The barship function is set in advance for each rock type, and
Apply the membership function of each selected fuzzy rule.
The fitness is calculated and the fitness of the integrated fitness is calculated from each fitness.
Calculates table values, and calculates final confidence and fuzziness calculated by rule-based inference.
Combined with the representative value of integrated fitness calculated by inference
To determine the rock confidence and calculate this rock confidence for each rock type.
Rock determination method characterized by determining the rock type became rock confidence and kind of unknown rock. 2. The less ambiguity questions include color index, grain
Patterns other than the shape , color tone, cracking method, crack, hardness, birth shape, grain size, grain shape, crystal, crystal state, pulse color, hard white pulse hardness, stripe and layer state, and // About the color of the net
The method of claim 1, wherein the rock is determined .
Law . 3. A questionnaire including ambiguity, granular pattern, gloss, rock determination method according to claim 1, characterized in that for the touch. [4] A question with less ambiguity and the question
For each rock type
You a knowledge database system, the rule-based reasoning with
Claims 1 you, characterized in that used have either 3
One of rock determination method. 5. A question including an ambiguity and the question
Fuzzy rule and corresponding fuzzy rule
Knowledge database with the membership function for each rock type
The over scan device, or rock determination method of claims 1, characterized in Rukoto used in fuzzy inference 3. 6. A knowledge theory for a knowledge database device.
Ability to clarify and / or modify knowledge
The rock determination method according to claim 4 or 5, wherein