System for building and deploying inference model

A computer-implemented method of constructing one or more inference models for use in remote condition monitoring of one or more first assets of a first fleet, the method comprising:

• providing a memory module and a processor communicatively coupled with the memory module;
• providing a server module communicatively coupled with the processor and a model builder application being an executable computer application configured to execute one or more tasks when run on the processor;
• providing at least one of a configuration file being stored in the memory module and a compiled configuration database being stored in the server module, wherein one or both of the configuration file and the compiled configuration database (i) is directly editable and readable by a user and (ii) comprises a model configuration data containing information and being customized to the one or more first assets and the one or more inference models;
• wherein the model configuration data is defined as a plurality of structural and performance characteristics of the one or more first assets;
• dynamically configuring, via the processor, the model builder application based on the status of the one or more first assets and the one or more options selected by the user;
• receiving, by the model builder application, the model configuration data directly from the configuration file and/or compiled configuration database;
• and
• operating the model builder application in a learning mode according to the received model configuration data, wherein in the learning mode operating the model builder application comprises:
• receiving historical operational data corresponding to a plurality of operating states of across the one or more first assets of the first fleet and a plurality of fault conditions for the first asset;
• calling one or more processing algorithms, wherein the one or more processing algorithms is determined by the model configuration data;
• and
• constructing, by using the received model configuration data, the received historical operational data, and the one or more processing algorithms, the one or more desired inference models;
• wherein during operation of the learning mode, the information is contained within the model configuration data, and the processing algorithms are integrated within the model builder application, the information and the processing algorithms thereby being separate from each other.