Method for operating a fan system and fan system with a reverse curved radial fan

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing fan systems with backward-curved centrifugal fans face challenges in precisely controlling volume flow without sensors, due to unclear relationships between motor current and generated volume flow, especially with various external influencing variables.

Innovation Solution

A method utilizing an artificial neural network to regulate the fan system by analyzing motor current, voltage, and fan speed, which learns to account for changes in outlet pressure and volume flow, allowing precise control without flow sensors through recorded and calculated operating parameters.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If mathematical models or algorithms are used to control the fan based on motor current and operating parameters, then control can be achieved without flow sensors, but the volume flow determination is relatively imprecise due to unclear relationships between motor current and generated volume flow

Engineering Contradiction:
Improvesensorless controlVSAvoidvolume flow determination precision
Core Design Contradiction:
Extent of automationVSMeasurement precision

Solution Approach 1:

The patent introduces an intermediary device (the determination device with neural network or support vector machine) that mediates between the motor current/operating parameters and the volume flow determination. This intermediary learns the complex, non-linear relationships between motor parameters and actual volume flow through training data, providing more accurate predictions than direct mathematical models while still avoiding the need for flow sensors

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent changes the approach from using simple mathematical models to using machine learning models (neural networks or support vector machines) that can capture complex parameter relationships. The determination device is trained with multiple datasets containing operating parameters and corresponding volume flow measurements, allowing it to learn and adapt to non-linear relationships between motor current, speed, and actual volume flow

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If flow sensors are used to measure volume flow for control purposes, then precise volume flow measurement is achieved, but the device complexity and cost increase

Engineering Contradiction:
Improvevolume flow measurement precisionVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts the measurement function from physical flow sensors and relocates it to a determination device that uses machine learning algorithms. Instead of measuring volume flow directly with sensors, the system extracts volume flow information from existing motor operating parameters (current, speed, voltage) through intelligent analysis, thereby eliminating the need for additional flow measurement hardware

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The motor and control system serve dual purposes: they not only drive the fan but also provide the data needed for volume flow determination. The existing sensors that monitor motor operating parameters are repurposed to feed the machine learning model, allowing the system to self-determine volume flow without requiring external flow measurement devices

Inventive Principle:
Principle #25Self-service

3Ease of operation

If traditional control methods are used without considering external influencing variables, then control simplicity is maintained, but the ability to regulate volume flow under varying external conditions deteriorates

Engineering Contradiction:
Improvecontrol simplicityVSAvoidresponse to external disturbances
Core Design Contradiction:
Ease of operationVSAdaptability or versatility

Solution Approach 1:

The patent performs preliminary action by training the machine learning model with diverse training datasets that include various operating conditions and external influencing variables before actual operation. This pre-training allows the determination device to already know how to compensate for common disturbances and changes in operating conditions, enabling adaptive control without complex real-time calculations

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3916236B1Method for operating a fan system and fan system with a reverse curved radial fan
Publication Date: 2023.03.01 EBM PAPST MULFINGEN GMBH & CO KG
  • EP3916236B1 patent drawingFigure 1~2
  • EP3916236B1 patent drawingFigure 3~4
  • EP3916236B1 patent drawingFigure 5~6

AI summary

The invention relates to a method for operating a fan system (10) and to such a fan system (10). The fan system (10) has a control unit (11) with an artificial neural network (19). The control unit (11) controls a motor (13) of a backward-curved radial fan (12). The radial fan (12) generates a gas flow (G) characterized by a flow rate, in particular the volume flow rate (Q). The flow rate is not detected by a sensor, but is determined by the artificial neural network (19) depending on input variables, and the motor (13) is controlled or regulated by the control unit (11) based on this.The motor current (I) and motor voltage (U) of the motor (13), as well as their temporal behavior, which can be determined by the time derivative (e.g., first-order gradient) or by at least one previous value at an earlier time, are fed to an input layer (30) of the artificial neural network (19). It is particularly advantageous if the artificial neural network (19) determines an output pressure value and feeds this value back to the input layer (30) internally or externally as an input.