Neural Network Boiler Control for Waste Incineration

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Waste incineration plant boilers are challenging to control efficiently due to their complexity, requiring human operators and resulting in costly and error-prone processes, with existing methods struggling to achieve optimal combustion efficiency and residue reduction.

Innovation Solution

A trained neural network is used to control the boiler by processing input data from parameters like boiler drum pressure, combustion chamber temperature, and air duct systems, generating output data for control elements such as air dampers, enabling automated or quasi-automated control.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If human operators control the boiler, then the system can handle complex decision-making, but the control becomes costly and error-prone

Engineering Contradiction:
Improvecontrol reliabilityVSAvoidautomation level
Core Design Contradiction:
ReliabilityVSExtent of automation

Solution Approach 1:

The neural network enables the boiler control system to serve itself by automatically processing sensor data and generating control commands without human intervention. The system learns optimal control strategies through training data and autonomously adjusts combustion parameters, eliminating dependence on human operators while improving reliability and reducing costs.

Inventive Principle:
Principle #25Self-service

2Productivity

If conventional control methods are used, then the system is simple to implement, but combustion efficiency and residue reduction are suboptimal

Engineering Contradiction:
Improvecombustion efficiencyVSAvoidcontrol system complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent replaces conventional mechanical control systems with an intelligent neural network-based control system. The neural network processes multiple sensor inputs and generates optimized control commands, achieving superior combustion efficiency and residue reduction despite the increased computational complexity, which is manageable through modern computing resources.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Measurement precision

If more parameters are considered for control, then the control precision improves, but the system complexity increases

Engineering Contradiction:
Improvecontrol precisionVSAvoidparameter measurement complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The neural network serves as a universal control mechanism that can process multiple different parameter types (temperature, pressure, flow rates, oxygen content) through a single integrated architecture. This multi-functional approach allows the system to consider numerous parameters for precise control without proportionally increasing system complexity, as the neural network handles diverse inputs through standardized processing layers.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP3696462B1Method for controlling a boiler of a refuse incineration plant by means of a trained neural network and method for training a neural network for controlling a boiler of a refuse incineration plant
Publication Date: 2021.08.11 UNIPER TECH GMBH
  • EP3696462B1 patent drawingFigure 1
  • EP3696462B1 patent drawingFigure 2a
  • EP3696462B1 patent drawingFigure 2b

AI summary

This disclosure specifies a method for controlling a boiler of a waste incineration plant by a trained neural network, comprising: - inputting input data into the trained neural network, wherein the input data is based on parameters that include at least a boiler drum pressure, a combustion chamber temperature, a parameter of a waste feeding system (1), a parameter of a combustion air duct system (HLA), a parameter of a flue gas discharge duct system (HNA), and a parameter of a live steam duct system (LBA); and - computing and outputting output data by the trained neural network, wherein the output data includes at least control data for at least one control element that contains at least one air damper of the combustion air duct system (HLA).