Anaerobic Digester Fault Diagnosis Using Model-Based Deviation Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for diagnosing faults in anaerobic digesters are inefficient and slow, often taking weeks to return to stable operation after a fault is detected, necessitating advanced and timely fault detection to maintain high biogas productivity and reduce idle times.
Innovation Solution
A method using a reverse engineering approach with MM-AD simulations to identify input variations by comparing measured and expected output parameters, setting a threshold deviation to detect operational faults in real-time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional fault detection methods using general references and handbooks are used, then the system is simple to operate, but fault detection is slow and inefficient with prolonged recovery time
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing fault conditions through simulations before actual operation. The system performs virtual experiments during a preparation stage to establish fault signatures and detection thresholds, so that when real-time monitoring occurs, the fault detection can immediately compare against pre-established criteria rather than relying on slow manual analysis of general references
Solution Approach 2:
The patent uses copying by creating virtual copies of the digester through mathematical models (ADM1 simulations). These virtual replicas are used to simulate fault conditions and generate reference data without affecting the actual physical system. The virtual fault signatures are then copied and used for comparison during real-time monitoring, enabling fast fault detection without complex physical sensors or invasive measurements
2Measurement precision
If mathematical models and simulations are used to predict biogas production, then the prediction accuracy is improved, but the system complexity increases
Solution Approach 1:
The patent applies taking out by extracting only the essential fault detection functionality from the complex mathematical model. Instead of using the full ADM1 model for continuous monitoring, the system extracts specific fault signatures and detection thresholds from preliminary simulations, then uses these simplified criteria for real-time comparison. This separates the complex modeling work (done once offline) from the simple real-time detection (done continuously online)
3Reliability
If operator supervision is used to maintain stable digester conditions, then the system is easy to operate, but the response time to faults is prolonged
Solution Approach 1:
The patent implements automated feedback by continuously comparing measured digester parameters against expected values derived from the mathematical model. When deviations indicate fault conditions, the system automatically generates alerts without waiting for manual operator assessment. This closed-loop feedback mechanism maintains process stability through rapid detection and notification, reducing the time from fault occurrence to operator awareness and enabling faster recovery
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
Examples methods may comprise a preparation stage, comprising: selecting an operation fault to be diagnosed and an operational parameter associated with the fault, running multiple simulations in a mathematical model of anaerobic digestion (MM-AD) to select an output parameter that varies with a variation of the operational parameter, and determine a threshold deviation of the output parameter that indicates that the fault is occurring. Example methods may further comprise an operation stage, comprising: obtaining measured values of the output parameter, obtaining expected values of the output parameter as provided by an MM-AD running in parallel with the operation of the digester, comparing the measured values and the expected values, over time, and determining that the fault is occurring if the deviation between the measured value and the expected value is above a threshold deviation.