Factory Steam Prediction Using Operational Data Instead of Flow Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting the usable amount of steam in a factory are limited by the need for steam pressure and flow velocity measurements, which can be difficult to obtain, leading to inefficiencies in energy management.
Innovation Solution
A gas amount prediction method using a learning model to calculate the usable steam amount based on past operation data, considering steam loss due to heat dissipation, and utilizing regression analysis or autoregressive moving average to predict the generation and use rates of steam.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If steam pressure and flow velocity measurements are used to predict usable steam amount, then prediction accuracy is improved, but measurement difficulty and device complexity increase
Solution Approach 1:
The patent introduces a learning model as an intermediary that indirectly predicts usable steam amount through readily available operational parameters (steam generation amount, atmospheric pressure, temperature, humidity) rather than directly measuring difficult-to-obtain steam pressure and flow velocity. This mediator translates easily measurable data into accurate predictions without requiring complex measurement systems.
Solution Approach 2:
The patent replaces the mechanical measurement system (pressure sensors, flow velocity meters) with an information processing system (learning model using regression analysis or autoregressive moving average). This substitution eliminates the need for complex physical measurement devices while achieving comparable or superior prediction accuracy through data-driven modeling.
2Reliability
If steam pressure and flow velocity measurements are implemented, then usable steam amount can be predicted, but device complexity and measurement requirements increase
Solution Approach 1:
The learning model serves multiple functions: it predicts usable steam amount, adapts to different climatic conditions through selective learning, and works with various input parameters (steam generation amount, atmospheric pressure, temperature, humidity). This multi-functional approach replaces multiple specialized measurement devices with a single versatile prediction system.
Solution Approach 2:
The system uses readily available factory operational data and public meteorological data to perform predictions without requiring external measurement infrastructure. The learning model learns from historical operation data and autonomously improves prediction accuracy, eliminating the need for complex external measurement systems.
3Measurement precision
If correction coefficients based on steam pressure or flow velocity are used, then steam loss correction is improved, but ease of operation deteriorates
Solution Approach 1:
The learning model incorporates feedback mechanisms by continuously learning from historical operation data and adjusting prediction parameters. The model uses past relationships between steam generation, environmental conditions, and actual steam usage to automatically refine correction factors, eliminating the need for manual correction coefficient calculations based on complex measurements.
Solution Approach 2:
The patent changes the parameter set from difficult-to-measure physical parameters (steam pressure, flow velocity) to easily obtainable parameters (steam generation amount, atmospheric pressure, temperature, humidity). This parameter transformation maintains prediction accuracy while dramatically simplifying data acquisition and system operation.
Data Source
AI summary
A gas amount prediction method predicts an amount of gas generated in a factory, and includes a generation amount calculation step of calculating a generation amount of actually usable gas by using a learning model that has learned a relationship between a generation amount of the gas and a use amount of the gas in past operation data.