Machine Risk Evaluation Using Main and Subsidiary Accident Factors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current risk management systems at construction sites struggle to accurately assess and differentiate between various types of accidents, such as accidental contact, overturning, and falls, due to the complexity of information from sensors, leading to inefficient analysis and potential oversight of near-miss events.
Innovation Solution
A risk management system that classifies accident factors into main and subsidiary factors, using sensors to measure machine and environmental parameters, calculates evaluation values for these factors, and integrates risks to determine the occurrence risk, recording parameters when the integrated risk exceeds a threshold, thereby enhancing the accuracy of accident analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the number of sensors is increased to capture more information about accidents and near-miss events, then the ability to grasp accidents and near-miss events improves, but the amount of information to be analyzed becomes enormous making specification and selection of necessary information difficult
Solution Approach 1:
The system extracts only the necessary information from the enormous amount of sensor data by identifying key parameters specific to each accident type. For contact accidents, it extracts position and velocity information; for overturning accidents, it extracts posture information such as machine body inclination; for fall accidents, it extracts height information. This extraction principle resolves the contradiction by filtering out unnecessary information while retaining critical data for accident detection.
Solution Approach 2:
The system segments the analysis process by dividing accidents into distinct types (contact, overturning, fall) and applying specific analysis methods for each type. This segmentation allows the system to handle different accident scenarios with tailored approaches, reducing the overall complexity by breaking down the comprehensive analysis into manageable, type-specific segments.
2Loss of information
If all collected information is saved for a long period, then the completeness of data for analysis is improved, but the storage capacity is limited making it difficult to save all information
Solution Approach 1:
The system extracts and stores only the necessary information for risk analysis by identifying key parameters relevant to each accident type. Instead of saving all sensor data, it selectively stores position, velocity, posture, and other critical parameters when accidents or near-miss events are detected. This extraction approach ensures data completeness for analysis while respecting storage capacity limitations.
Solution Approach 2:
The system performs preliminary analysis to identify when accidents or near-miss events occur before storing data. By detecting accident conditions in advance and then storing only the relevant information surrounding these events, the system prepares the necessary data for analysis without requiring long-term storage of all collected information.
3Measurement precision
If various types of information including posture information and terrain information are utilized to determine the occurrence of accidents, then the accuracy of accident determination is improved, but it becomes more difficult to properly consider occurrence factors and accurately compute the degree of danger
Solution Approach 1:
The system segments the complex determination process by creating separate determination methods for each accident type. For contact accidents, it uses position and velocity-based methods; for overturning accidents, it uses posture-based methods involving machine body inclination; for fall accidents, it uses height-based methods. This segmentation simplifies the consideration of occurrence factors by applying tailored approaches to each accident type rather than attempting a unified complex analysis.
Solution Approach 2:
The system applies local quality by using different determination methods and information types suited to each specific accident type. Instead of applying a uniform complex analysis to all accidents, it uses contact determination for contact accidents, posture determination for overturning accidents, and height determination for fall accidents. This localized approach improves accuracy while reducing overall complexity.
4Measurement precision
If the determination of near-miss events is made strictly to avoid false positives, then the accuracy of near-miss detection is improved, but there is a possibility that significant near-miss events are overlooked
Solution Approach 1:
The system segments near-miss detection into type-specific determinations (contact near-miss, overturning near-miss, fall near-miss) with appropriate criteria for each type. This segmentation allows the system to apply strictly appropriate thresholds for each accident type rather than a single strict threshold, improving both accuracy and completeness by capturing significant near-miss events that meet type-specific criteria.
Data Source
AI summary
The present invention intends to provide a risk management system that can accurately extract information necessary for analysis of a variety of accidents in which a machine is involved. For this purpose, a controller calculates a main risk that is the degree at which a main factor of an accident contributes to the occurrence of the accident on the basis of an evaluation value of the main factor, and calculates a subsidiary risk that is the degree at which a subsidiary factor of the accident contributes to the occurrence of the accident on the basis of an evaluation value of the subsidiary factor. Furthermore, the controller calculates, as the occurrence risk of the accident, an integrated risk that has a value equal to or larger than the main risk and increases or decreases at a degree lower than the degree of increase or decrease in the subsidiary risk, and, when the integrated risk has exceeded a predetermined threshold, causes a recording device to record parameters measured by measuring devices in a certain time period including a clock time at which the integrated risk has exceeded the predetermined threshold.