Development of an algorithm based on GMDH for predictive maintenance based on environmental parameters and usage pattern of the system.

Customer

Leading Medical Equipment Manufacturer in Japan

Business Situation

Customer was facing difficulty to plan maintenance activities and replacement of major parts like X-Ray tubes in advance. The customer wanted to have a system that could predict when such replacements may be required for better planning

Solution

Development of an algorithm based on GMDH for predictive maintenance based on environmental parameters and usage pattern of the system.

Details of the solution:

• Analyze impact of environmental parameters on a X-ray tube’s life (Impact Analysis)
• Predict replacement date of a X-Ray tube based on usage rate (Trend Analysis)
• Adaptive model for trend analysis Learning system for data modeling
• Learning system for data modeling
• Combination of soft computing and statistical methodologies
• A predefined model cannot be designed because of the non availability of previous data
• Neuro-Fuzzy GMDH (Group Method of Data Handling) algorithm for Trend Analysis and Impact Analysis
• GMDH uses information directly from data samples and minimizes influence of apriori author assumptions about results of modeling
• Automatically finds interpretable relationships in data and selects effective input variables

Benefits derived by the customer:

• Better planning for replacement of major equipment parts.
• Better understanding of the impact of environmental parameters and usage on the life of equipment parts.