JP2001351057A - Predictive system and method and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a predictive system, a predictive method and a recording medium for grasping a process which may lead to dangerous work in working as the chain of changes in a time series working procedure or in working environment, time-sequentially analyzing the causal relation of these changes and predicting the dangerous work. SOLUTION: The decomposition of a working procedure to be analyzed and the arrangement of the sequences are performed (step 101), the causes of a disaster are extracted, the causal relation of these causes is arranged (step 102), the working processes are connected with the causes of the disaster (step 103), and the connected state is expressed by a Petri net (step 104). Simulation concerned with the work is performed by using the Petri net (step 106), a working method, a working environment or the like is improved (step 108) and the processing is repeated until prescribed safety is satisfied.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a prediction system, a prediction method, and a recording medium on which a program for operating a computer as a prediction system is recorded. It is.

[0002]

2. Description of the Related Art Hitherto, prediction of dangerous work in construction work is mainly based on empirical judgment based on past disaster cases, and prediction by a quantitative method is rarely performed. Therefore, the prediction of dangerous work or disaster often depends on the experience of the worker at the time of work. In addition, in fields other than construction, analysis methods for analyzing the cause of a disaster and evaluating the degree of danger were sometimes applied.

[0003]

However, such an analysis method does not take into account the fact that the configuration or state of the system to be analyzed changes, and the work environment changes with time as in the case of construction work. However, there is a problem that it is difficult to analyze work safety in consideration of the change. In addition, it is difficult to predict a dangerous work throughout the entire work in response to such a change in the work environment even in the risk prediction based on the experience of the worker.

The present invention has been made in view of such a problem, and an object of the present invention is to consider a process leading to a dangerous work as a chain of changes in a work procedure and a work environment in a time series, and a causal result thereof. An object of the present invention is to provide a prediction system, a prediction method, and a recording medium for predicting dangerous work by analyzing relationships in a time series.

[0005]

According to a first aspect of the present invention, there is provided a work extracting means for decomposing and extracting a work process in time series, a factor extracting means for extracting a disaster factor, It is characterized by comprising: a creating means for creating a Petri net using the work process extracted by the extracting means and the factors extracted by the factor extracting means; and a predicting means for predicting a danger in the work using the Petri net. Is a prediction system.

In a second aspect of the present invention, a work extraction step for decomposing and extracting work steps in chronological order, a factor extraction step for extracting a disaster factor, and a work step and a factor extraction step extracted by work extraction means. A prediction method comprising: a creation step of creating a Petri net using factors; and a prediction step of predicting danger in work using the Petri net.

In a third aspect of the present invention, a work extracting means for decomposing and extracting work processes in a time series, a factor extracting means for extracting a disaster factor, a work process and a factor extracted by the work extracting means are extracted by the work extracting means. A recording of a program for operating a computer as a part or all of a prediction system including a creation unit for creating a Petri net using factors, and a prediction unit for predicting danger in work using the Petri net. Medium.

[0008]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a schematic configuration diagram of hardware of a prediction system 1 according to an embodiment of the present invention, and FIG. 2 is a flowchart illustrating a processing procedure of the prediction system 1. The work to be analyzed by the prediction system 1 is a construction work.

As shown in FIG. 1, the prediction system 1 includes a computer 3, a work database 5, a disaster cause database 7, a CD-ROM 9, and the like. Computer 3
Performs a process of predicting dangerous work, disaster, and the like using a Petri net described later.

The work database 5 holds data relating to work, such as work procedures and their context. The disaster factor database 7 holds data relating to disasters that may occur during work, factors of dangerous work, and the like. A petri net is created using the data stored in the work database 5 and the disaster factor database 7, and is processed by the computer 3.

The CD-ROM 9 records a program for danger prediction processing performed by the computer 3 and the like. Next, the process of danger prediction by the prediction system 1 shown in FIG. 2 will be described.

The work procedure to be analyzed is disassembled and the context is arranged (step 101), and a dangerous work that can occur therein, that is, a factor of a disaster is extracted, and a causal relationship between the work and the cause is organized ( Step 102).

The operations extracted in steps 101 and 102 are combined with the disaster factors (step 103), and they are represented by a Petri net (step 104). Petri nets are an effective tool for modeling and evaluating systems. A detailed description of Petri nets will be given later.

Next, the state transition probability in the Petri net is determined (step 105), a simulation is performed on the work using the Petri net, and the disaster occurrence probability of each work, the time until the disaster occurs, and the like are calculated ( Step 10
6) When the predetermined safety is satisfied (step 107), the simulation is terminated.

If the predetermined safety is not satisfied in step 107, that is, if dangerous work or disaster can occur with a high probability, the work method and work environment are improved (step 108), and the points of improvement are considered. Then, remodeling is performed by the Petri net (step 104), and the processing is repeated until predetermined security is satisfied.

Next, the processing of the system shown in FIG. 2 will be described by taking specific work as an example. Here, the dismantling work of the beam formwork is taken as an example of the work to be analyzed.

(Step 101) The work procedure of the target work is disassembled, and the context is rearranged. For example, in the case of dismantling work of a beam formwork, work is extracted as follows, and the work order is arranged.

Operation 1: Ascend and descend from the slab to the scaffold. Operation 2: A nail removal operation is performed. Operation 3: Perform small transport work. Operation 4: The panel is peeled off using a bar. Operation 5: Peel the panel by hand. Task 6: Move from the scaffold onto the slab.

From the above-mentioned work, the place where the work is performed, the name of the work, the state change due to the work, the time required for the work, and the like are extracted and arranged. Data on the extracted work can also be held in the work database 5.

(Step 102) Next, a disaster factor that can occur in the target work is extracted, and a causal relationship such as why it occurs and what happens when it occurs is arranged. For example, when “there is an unnecessary material at the foot” during the work on the scaffold, the worker gets on the unnecessary material. The feet slip on unnecessary materials, causing a loss of balance. It loses balance and falls on the slab.
It may be a near miss or a serious disaster depending on the condition of the fall. These extracted and arranged factors and causal relationships are also stored in the disaster factor database 7 or the like.

(Step 103) The work procedure extracted in step 101 is combined with the disaster factor extracted in step 102. For example, the above-mentioned “there is an unnecessary material at the foot” and the dropping is a dangerous work that can occur after climbing to the scaffold because the work place is on the scaffold.

(Steps 104 to 108) Next, the processing result of step 103 is represented by a Petri net. FIG. 3 is a diagram modeling a petri net for risk prediction in the dismantling work of the beam formwork. Before giving a detailed description of the Petri net model shown in FIG. 3, the Petri net model will be described.

FIG. 4 is a diagram showing elements constituting a Petri net. Petri Net is a place 33-1, 33-
2, arc 37-1, 37-1, transition 35,
It consists of a token 39. Places 33-1 and 33-
2 indicates a condition, and transition 35 indicates an event or a state change.

The arc 37-1 indicates the connection from the place 33-1 to the transition 35, and the arc 37-2 indicates the connection from the transition 35 to the place 33-2. The token 39 is placed in the place 33-1 and moves to the place 33-2 by a transition 35, that is, an event. The occurrence of an event or state change is called "a transition is fired". FIG. 5 is a diagram showing the state of the Petri net after the transition 35 has fired.

There are the following rules for the firing condition of the transition 35. 6, 7, and 8 are diagrams for explaining the firing conditions. As shown in FIG. 6, in order to fire the transition Ta, a token must be present in both of the input places Pa and Pb. That is, when both the conditions Pa and Pb are satisfied, the transition Ta fires, and the event Ta occurs.

The firing condition shown in FIG. 7 shows a case where transitions Ta and Tb can occur with a certain probability. This probability is called a state transition probability. FIG. 8 shows that the state shifts to the place Pa when any one of the transition Ta and the transition Tb is fired.

A petri net model for danger prediction of the beam formwork dismantling work shown in FIG. 3 will be described using the above-described petri net and its firing conditions. In FIG. 3, P indicates a place, and indicates a work name and a state of the work.
In addition, T is a transition, and indicates a factor of a state change. The numbers in the figure indicate the working time (minutes).

The token in the place P1 moves from above the slab (P1) to ascending / descending (P2) when the transition T1 "start of ascending and descending to the scaffold" is fired, and when the transition T2 "end of ascending and descending to the scaffold" is fired. On the scaffold (P
Go to 3). Here, the movement of the token indicates that the work is performed and the state changes.

Next, the transition T3 "work start" is ignited, and a nail pulling (P4) operation is performed.
4 The small transport (P5) work is performed by the ignition of "start small transport".
Next, the transition T5 “start peeling of panel by burl” is fired, and panel peeling (P6) operation is performed.

At this time, the work of the place P6 is performed smoothly, and one of the transition T6 “starting the peeling of the panel by hand” and the transition T10 “bar comes off” occurs with a predetermined state transition probability. sell. The state transition probability is determined in consideration of the work environment and the like (step 105) and is input. It is also possible to make this state transition probability into a database and hold it.

When the transition T6 "Panel peeling by hand is started" fires, the panel is peeled by hand (P7), and when the transition T7 "Panel peeling end" fires, the panel is placed (P8) and the next work is performed. If there is a part, a token is newly given, the next construction part is moved (P11), and the work is started.

If there is no next construction site, the worker gets off the scaffold (P9), and returns to the slab when the transition T9 "end of scaffolding elevation" fires (P10).

When the above-mentioned transition T10 "bars come off" is fired, there are cases where the work can be continued even if the bars come off (P21) and cases where it receives a recoil (P12). If it reaches place P21, transition T12
The “start of next work” fires, and the panel is removed by hand (P
7).

When a recoil is received (P12) and the stepping work is further performed (P13: a new token is given in this case), a transition T11 "unstable state" is fired, and the balance described below is reached. State (P
16). That is, according to the firing conditions shown in FIG.
With the tokens on both 12 and P13, T11 will be fired, and here, "It is a work on a stepladder and
If you receive a recoil, you will lose balance and fall. "

Next, consider a case where there is an unnecessary material on the scaffold as one of the causes of the disaster. A place P14 is set for the condition that there is an unnecessary material, and a new token is arranged. In this case, one of the transition T13 “the worker gets on the unnecessary material” and the transition T17 “does not get on” is fired at a predetermined state transition probability. If T17 fires, no disaster will occur.

When the transition T13 is ignited, the operator gets on the unnecessary material (P15), and
4 "Step slips" and transition T18 "No slip"
Fires at a predetermined probability. If T18 fires, no disaster will occur.

When the transition T14 is fired, or becomes unstable due to the recoil due to the stepladder operation or the removal of the crowbar, the operator loses his balance (P16), and the transition T19 has a predetermined probability.
"Do not fall" fires. When T19 is fired, the work is continued (P18), and transition T21 “transition to next work” is fired.

When the transition T21 is ignited and the transition to the next operation is performed, the cause of the disaster returns to "there is an unnecessary material at the foot (P14)", and at the same time, the preparation for the next operation is performed (P20). T12 "Start next work"
Ignites. Thus, the process proceeds to the next operation, which is to remove the panel by hand (P7).

Further, the operator loses balance (P1
6) Insufficient handrail given new token (P
When both conditions (19) are added, the transition T15 "falls" is fired with a predetermined probability. T1
When 5 fires, the worker falls from the scaffold onto the slab (P17), and if transition T16 "permissible load" fires at a predetermined probability, a near-miss occurs, and transition T20 "overload on body" occurs. If fires,
It is a "serious disaster."

In step 106, the above-described simulation is performed to calculate the probability and the time of a “close call” or a “serious disaster”.

For example, the calculation of the probability of a near-miss event will be described below. Here, the probability of getting on the unnecessary material of the scaffold (ignition of T13) is “a”, and the foot slips (T
The probability that “14” will be fired) is “b”, the probability that the bar will come off and become unstable due to the recoil (T11 fires) is “c”, and the fall from the scaffold is an allowable load (T16 fires). Probability "d", falling from the scaffold (T15)
Assume that the probability is “e”.

In consideration of the firing conditions shown in FIG. 6 to FIG. 8, when the probability of a near-miss event is calculated from the above probability, the result is “(a × b + c) × d × e”.

As a result of the above simulation,
If the probability of a “serious disaster” or “near-miss” is greater than a predetermined threshold or the like, in the Petri net shown in FIG. The work environment is changed as described above to improve (Step 108).

For example, in the operation shown in FIG. 3, when the burl comes off in the "peeling off the panel with a burl" (P6), it may receive a recoil and lead to a fall accident. Therefore, when this probability is large, it is possible to improve the work by taking measures to prevent the burl from coming off or to cancel the panel peeling work using the burl, and to perform the simulation again.

As described above, the work is improved, the state transition probability is changed, places and transitions are added and deleted, and the Petri net shown in FIG. 3 is reconstructed.
By repeating the simulation, if the probability of a “serious disaster” or “near-miss” becomes smaller than the threshold value, the dangerous work prediction process ends.

Next, the case where the time until a disaster is calculated in step 106 will be described. FIG. 9 is a diagram illustrating a result of calculating an average time until a “serious disaster” is obtained under different working conditions.

Using the Petri net shown in FIG. 3, a simulation leading to a “serious disaster” as an end condition is repeated (repeated 100 times in the example shown in FIG. 9), and the average time required to reach the “serious disaster” is calculated. Ask. For example, the work condition 1 is “when using a stepladder and a scaffold”, and the average time until the “serious disaster” is “t1”.

Next, the simulation is repeated for "when a stepladder and a scaffold are used and there are unnecessary materials" as work condition 2, and the average time "t2" until the "serious disaster" is obtained.

Here, it can be seen that the work condition 2 has a shorter time to the "serious disaster" than the work condition 1.
In step 107, the work method is changed so as to increase the average time until the "serious disaster".

For example, as work condition 3, for “when a rolling tower is used” instead of “stepladder + scaffolding board”, an average time t3 until a “serious disaster” is calculated. As a result, t3>t1> t2, and it can be seen that the operation safety is higher in "when using a rolling tower" than in "when using a stepladder and a scaffold".
In this way, if a safety level higher than a predetermined level is obtained (Step 1)
07) The simulation is terminated assuming that the work has been improved.

Also in the case of "using a rolling tower" shown in FIG. 9, a simulation is performed by changing the span L and the floor height H. For example, work condition 3 is “H = 3000 (mm), L = 4000 (mm)”, and work condition 4 is “H = 4000 (mm), L = 6000 (m).
m) ”, the average time is t3> t4.

In the above description, the "serious disaster" has been described as an example, but the time required for the "near miss" can be calculated in the same manner. As described above, if the dangerous work prediction system 1 is used, it is possible to calculate in advance the time leading to a “serious disaster” or a “close call”, and to improve the work so as to extend the time.

As described above, by using the present system, dangerous work can be predicted in advance, and the risk factor can be removed. In addition, even when a large change in the work environment occurs during the work, by performing the simulation using the present system, it is possible to predict in advance the dangerous work due to the change in the work environment, and it is possible to guarantee work safety. .

Note that all or part of the processing in FIG. 2 can be performed by a computer, and the program is stored in a CD-RO.
M and the like. In this embodiment, the work of dismantling the beam formwork has been described as an example, but the present system can also perform danger prediction of any construction work or work other than the construction work. In particular, since various works are combined in an actual construction work, it is very effective to predict dangerous work before the work using the present system.

[0055]

As described in detail above, according to the present invention, a process leading to a dangerous work is regarded as a chain of changes in work procedures and work environment in a time series, and a causal relationship thereof is analyzed in a time series. And can predict dangerous work.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of hardware of a prediction system 1 of the present invention.

FIG. 2 is a flowchart showing a dangerous work prediction process performed by the prediction system 1 of the present invention.

FIG. 3 is a view showing a petri net for dismantling work of a beam formwork.

FIG. 4 is a diagram showing elements constituting a Petri net.

FIG. 5 is a diagram showing a state of a petri net after firing.

FIG. 6 is a diagram showing a firing condition of a Petri net.

FIG. 7 is a diagram showing a firing condition of a Petri net.

FIG. 8 is a diagram showing a firing condition of a Petri net.

FIG. 9 is a diagram showing an average time until a “serious disaster” calculated under different working conditions.

[Explanation of symbols]

 1 Prediction system 3 Computer 5 Work database 7 Disaster factor database 9 CD-ROM P Place (work name or work state) T Transition (state) Cause of change) 33-1, 33-2 place 35 transition 37-1 37-2 arc 39 token

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Kazuhiko Arai 2-9-1-1, Tobita-Shi, Chofu-shi, Tokyo Kashima Construction Co., Ltd. Technical Research Institute F-term (reference) 5B049 BB05 CC02 CC31 DD01 DD05 EE03 EE36 EE41 FF09

Claims (12)

[Claims] 1. A work extracting means for decomposing and extracting work processes in time series, a factor extracting means for extracting a disaster factor, a work process extracted by the work extracting means, and a factor extracted by the factor extracting means. , A prediction unit for generating a Petri net, and a prediction unit for predicting a danger in an operation using the Petri net. 2. The Petri net is a place, an arc,
A transition is represented by a token, a place indicates the work process or a factor, a transition indicates a state change due to the work process or a factor, an arc connects the place and the transition, and the prediction unit determines that the token The prediction system according to claim 1, wherein the danger is predicted by moving the information. 3. When branching from the place to a plurality of transitions, a probability that each transition can occur is set, and when the danger of work is predicted by the prediction unit,
3. The prediction system according to claim 2, wherein danger is removed by changing the probability. 4. The predicting means calculates a time until the danger. If the time until the danger is shorter than a predetermined value, the work process is improved and the time until the danger is reached. 2. The prediction system according to claim 1, wherein is lengthened. 5. A work extraction process for decomposing and extracting work processes in time series, a factor extraction process for extracting a disaster factor, a work process extracted by a work extraction means, and a factor extracted in the factor extraction process. A creating step of creating a Petri net; and a forecasting step of predicting danger in work using the Petri net. 6. The Petri net is a place, an arc,
The place is represented by a transition or a token, a place indicates the work process or a factor, a transition indicates a state change due to the work process or a factor, an arc connects the place and the transition, and the prediction process includes 6. The prediction method according to claim 5, wherein danger is predicted by moving. 7. When branching from the place to a plurality of transitions, a probability that each transition can occur is set, and when a danger of work is predicted in the prediction step,
7. The prediction method according to claim 6, wherein danger is removed by changing the probability. 8. The predicting step calculates a time until the danger. If the time until the danger is shorter than a predetermined value, the work process is improved and the time until the danger is reached. The prediction method according to claim 5, wherein is lengthened. 9. A work extracting means for decomposing and extracting a work process in a time series, a factor extracting means for extracting a disaster factor, a work process extracted by the work extracting means, and a factor extracted by the factor extracting means. A recording medium that records a program that causes a computer to operate as a part or all of a prediction system including: a creation unit that creates a Petri net; and a prediction unit that estimates a danger in an operation using the Petri net. 10. The Petri net is represented by a place, an arc, a transition, and a token, a place indicates the work process and a factor, a transition indicates a state change due to the work process and a factor, and an arc indicates the place and the transition. 10. The recording medium according to claim 9, wherein the prediction unit predicts danger by moving the token in each place. 11. The recording medium according to claim 10, wherein when branching from the place to a plurality of transitions, a probability that each transition can occur is set. 12. The recording medium according to claim 9, wherein said predicting means calculates a time until a danger occurs.