DE-102025115903-A1 - RISK ANALYSIS METHOD AND RISK ANALYSIS SYSTEM

DE102025115903A1DE 102025115903 A1DE102025115903 A1DE 102025115903A1DE-102025115903-A1

Abstract

The aim is to clarify an element that is normally executed to reduce risk in an emergency. In a risk analysis procedure, a risk analysis system defines a variety of normal-time objective functions, each representing a normally achievable goal. The input to a predetermined function, which outputs a predetermined value in response to the input, is a predetermined equation comprising, as first-order terms or factors, a variety of explanatory factors, each multiplied by a first contribution degree related to the relevant explanatory factor. Furthermore, the risk analysis system defines an emergency-time objective function, which is an objective to be achieved to reduce risk. This function outputs a value in response to the input, where the input is a predetermined equation comprising, as first-order terms or factors, a variety of normal-time objective functions, each multiplied by a second contribution degree related to the relevant normal-time objective function. Furthermore, the risk analysis system outputs the second contribution levels in the emergency time objective function by machine learning from second sample data relating to the emergency time objective function, and extracts a prioritized normal time objective function based on the second contribution levels, which is to be optimized with priority from the normal time objective functions.

Inventors

Ken Naono
Mika TAKATA
Tsunehiko Baba
Keita Mizushina
Ken Sugimoto
Hiroaki Masuda
Mayuko Ozawa

Assignees

HITACHI, LTD.

Dates

Publication Date: 20260513
Application Date: 20250424
Priority Date: 20241112

Claims (14)

A risk analysis procedure executed by a risk analysis system that performs a risk analysis to reduce a risk occurring in an emergency, wherein the risk analysis procedure comprises: by a processor of the risk analysis system, defining a plurality of normal-time objective functions, each of which is an objective function representing a normally achievable goal and has as input to a predetermined function that outputs a predetermined value in response to the input, a predetermined equation comprising as first-order terms or factors a plurality of explanatory factors, each multiplied by a first contribution degree relating to the relevant explanatory factor; defining an emergency-time objective function, which is an objective function representing a goal to be achieved in order to reduce the risk, as a function that outputs a value in response to the input, wherein the input is a predetermined equation comprising as first-order terms or factors the plurality of normal-time objective functions, each multiplied by a second contribution degree relating to the relevant normal-time objective function; and Outputting the second contribution grades in the emergency time objective function by machine learning from second sample data relating to the emergency time objective function, and extracting, based on the second contribution grades, a prioritized normal time objective function to be optimized with priority from the normal time objective functions.
Risk analysis procedure according to Claim 1 , furthermore, encompassing: by the processor, outputting the first contribution grades in the prioritized normal-time objective function by machine learning of first sample data relating to the prioritized normal-time objective function, and extracting the explanatory factors corresponding to a predetermined number of first contribution grades, in descending order of the first contribution grades.
Risk analysis procedure according to Claim 2 , furthermore comprehensively: by the processor, managing the explanatory factors and the countermeasures to reduce the risk in conjunction with each other; and issuing the countermeasures that correspond to the extracted explanatory factors.
Risk analysis procedure according to Claim 2 , where the processor outputs the first contribution levels through machine learning for each sampling of the first sample data.
Risk analysis procedure according to Claim 3 , wherein the processor, together with the countermeasures, outputs priorities of the countermeasures that correspond to the explanatory factors, the priorities depending on the magnitude of the predetermined number of first contribution levels.
Risk analysis procedure according to Claim 3 , where the risk is an infection worsening during an infection pandemic, each of the explanatory factors are data relating to a health condition of each individual, each of the normal-time objective functions represents a relationship between an infection situation of each individual's disease and each explanatory factor, and the emergency-time objective function represents a relationship between a situation of infection worsening during the infection pandemic and each normal-time objective function.
Risk analysis procedure according to Claim 3 , where the risk is a delay in the delivery of a product in the event of an earthquake, each of the explanatory factors are data relating to an operation of each factory that produces a component that forms the product, each of the normal-time objective functions represents a relationship between the number of days of delay in the delivery of each component supplied by each factory and each explanatory factor, and the emergency-time objective function represents a relationship between the number of days of delay in the delivery of the product in the event of the earthquake and each normal-time objective function.
A risk analysis system that performs a risk analysis to reduce a risk occurring in an emergency, wherein a processor of the risk analysis system defines a plurality of normal-time objective functions, each of which is an objective function representing a normally achievable goal and has as input to a predetermined function that outputs a predetermined value in response to the input; a predetermined equation comprising as first-order terms or factors a plurality of explanatory factors, each multiplied by a first contribution degree relating to the relevant explanatory factor; and an emergency-time objective function that is an objective function representing a goal to be achieved in order to reduce the risk, as a function that outputs a value in response to the input, wherein the input is a predetermined equation. is, which comprises as first-order terms or factors the multitude of normal-time objective functions, each of which is multiplied by a second contribution degree relating to the relevant normal-time objective function, and outputs the second contribution degrees in the emergency-time objective function by machine learning from second sample data relating to the emergency-time objective function, and on the basis of the second contribution degrees extracts a prioritized normal-time objective function to be optimized with priority from the normal-time objective functions.
Risk analysis system according to Claim 8 , wherein the processor outputs the first contribution degrees in the prioritized normal-time objective function by machine learning from first sample data relating to the prioritized normal-time objective function, and extracts the explanatory factors corresponding to a predetermined number of first contribution degrees in descending order of the first contribution degrees.
Risk analysis system according to Claim 9 , wherein the processor manages the explanatory factors and the countermeasures to reduce the risk in conjunction with each other, and outputs the countermeasures that correspond to the extracted explanatory factors.
Risk analysis system according to Claim 9 , where the processor outputs the first contribution levels through machine learning for each sampling of the first sample data.
Risk analysis system according to Claim 10 , wherein the processor, together with the countermeasures, outputs priorities of the countermeasures that correspond to the explanatory factors, the priorities depending on the magnitude of the predetermined number of first contribution levels.
Risk analysis system according to Claim 10 , where the risk is an exacerbation of infection during an infectious pandemic, each of the explanatory factors are data relating to a health condition of each individual, each of the normal-time objective functions represents a relationship between an infectious situation of each disease of each individual and each explanatory factor, and the emergency-time objective function represents a relationship between a situation of exacerbation of infection during the infectious pandemic and each normal-time objective function.
Risk analysis system according to Claim 10 , where the risk is a delay in the delivery of a product in the event of an earthquake, each of the explanatory factors are data relating to an operation of each factory that produces a component that forms the product, each of the normal-time objective functions represents a relationship between the number of days of delay in the delivery of each component supplied by each factory and each explanatory factor, and the emergency-time objective function represents a relationship between the number of days of delay in the delivery of the product in the event of the earthquake and each normal-time objective function.

Description

BACKGROUND OF THE INVENTION 1. Field of the invention The present invention relates to a risk analysis method and a risk analysis system. 2. Description of the state of the art For example, the JP-2022-149246-A A technology to ensure that, in a power supply system equipped with a power generation unit that includes a fuel cell, the unit stops generating power not through an emergency stop process, but through a normal stop process that includes a temperature reduction step, in the event of a predicted earthquake of a specified magnitude within a certain timeframe. By clarifying a countermeasure to be taken in this way in an emergency, it is possible to mitigate the risk of such an emergency as a deterioration of the power generation unit. SUMMARY OF THE INVENTION However, the state of the art described above only indicates the countermeasure to be taken in an emergency, such as performing normal stop processing instead of emergency stop processing under certain conditions in the event of an earthquake, and does not clarify any element that should normally be performed to reduce the risk in an emergency. The present invention was made in view of the situation described above and has as one of its objectives the clarification of an element that should normally be carried out in order to reduce the risk in an emergency. As one aspect of solving the problem described above, a risk analysis procedure is provided, executed by a risk analysis system. This procedure performs a risk analysis to reduce a risk that arises in an emergency. The risk analysis procedure involves, through a processor of the risk analysis system, defining a multitude of normal-time objective functions, each representing a normally achievable goal. The input to a predetermined function, which outputs a predetermined value in response to the input, is a predetermined equation. This equation comprises, as first-order terms or factors, a multitude of explanatory factors, each multiplied by a first-order contribution level related to the relevant explanatory factor. The risk analysis procedure further includes, by the processor, defining an emergency-time objective function, which is an objective function representing a target to be achieved to reduce risk, as a function that outputs a value in response to the input. The input is a predetermined equation comprising, as first-order terms or factors, the multitude of normal-time objective functions, each multiplied by a second contribution grade related to the relevant normal-time objective function. The risk analysis procedure further includes, by the processor, outputting the second contribution grades in the emergency-time objective function by machine learning from second-sample data related to the emergency-time objective function, and extracting, based on the second contribution grades, a prioritized normal-time objective function to be optimized with priority from the normal-time objective functions. According to the present invention, it is possible, for example, to clarify an element that is normally to be carried out in order to reduce the risk in an emergency. BRIEF DESCRIPTION OF THE DRAWINGS 1A is a diagram illustrating a configuration of a selection system for a most prioritized countermeasure in normal time according to a first embodiment;1B is a diagram illustrating a hardware configuration of the selection system for a most prioritized countermeasure in normal time according to the first embodiment;2 is a flowchart to illustrate the selection processing of a highest-priority countermeasure in normal time according to the first embodiment;3A is a view illustrating a selection screen of a most prioritized countermeasure in normal time according to the first embodiment;3B is a view illustrating the selection screen of a most prioritized countermeasure in normal time according to the first embodiment;3C is a view illustrating the selection screen of a most prioritized countermeasure in normal time according to the first embodiment;3D is a view illustrating the selection screen of a most prioritized countermeasure in normal time according to the first embodiment;3E is a view illustrating the selection screen of a most prioritized countermeasure in normal time according to the first embodiment; and4 is a diagram illustrating a normal-time objective function group, an emergency-time objective function, and explanatory factors according to a second embodiment. DESCRIPTION OF PREFERRED EXECUTION FORMS A description of embodiments of the present invention will now be given with reference to the drawings. [First embodiment] In a first embodiment, in order to achieve the goal of preventing the risk of an infection worsening in an infected person in the event of an infectious pandemic, countermeasures are sought and presented that are normally taken by people before the occurrence of the infectious pandemic. In the first embodiment, an emergency time objective function G(x) to be minimized is