Reliability Research of Concrete Pump Main Oil Hydraulic System Based on T-S Fuzzy Fault Tree

Qing Zhao, Lijun Li, Hongbin Tang2017 · DOI

Abstract

For improving the reliability of concrete pumps, fault diagnosis is proposed based on the main oil hydraulic pumping system T-S fuzzy fault tree method. The proposed method uses fuzzy numbers to describe the probability of failure of hydraulic components and systems to address failure probability uncertainty. T-S fuzzy gates describe the contact between events to solve ambiguity and failure mechanisms associated with events. Furthermore, fuzzy numbers describe problem severity, considering the effect of the degree of fault on the system. The method is applied to the fault diagnosis of a concrete pump's main oil hydraulic system, achieving significant results.

Introduction

Predicting failures in complex industrial machinery, such as the main hydraulic systems of concrete pumps, is a critical challenge for maintenance and safety. Traditional reliability models often rely on rigid binary logic—assuming a component is either perfectly functional or completely failed. However, real-world mechanics are rarely so clear-cut. This paper introduces a more nuanced approach by leveraging T-S (Takagi-Sugeno) Fuzzy Fault Trees to model the inherent uncertainty and varying degrees of fault severity in hydraulic systems.

The Limitations of Traditional Fault Tree Analysis

Fault Tree Analysis (FTA) is a standard tool in industrial engineering for tracing system-level failures ("top events") to root causes ("bottom events"). Despite its utility, traditional FTA suffers from two primary weaknesses:

  1. Data Uncertainty: It requires precise failure probabilities for components, which are often unavailable or fluctuate in field conditions.
  2. Binary Logic: It uses simple AND/OR gates that cannot account for "semi-fault" states, such as minor leaks or pressure drops that degrade performance without causing total shutdown.

Methodology: T-S Fuzzy Logic Integration

To address these gaps, the researchers integrated T-S Fuzzy Logic into the fault tree structure. This approach replaces binary states with a spectrum of failure possibilities.

Trapezoidal Fuzzy Numbers

Instead of a single probability value, the model employs trapezoidal membership functions. These functions describe a range of probabilities and the "blurry" boundaries of a fault's severity, allowing for a more realistic representation of component health.

Trapezoidal fuzzy membership function and its mathematical definition

T-S Fuzzy Gates

Traditional logic gates are replaced by T-S fuzzy gates, which utilize IF-THEN rules. These rules allow the model to capture complex interactions, such as: "If Component A has a mild fault (0.5) AND Component B has a severe fault (1.0), the subsystem has a 60% chance of a severe fault." This expert-knowledge-driven approach captures mechanical nuances that purely mathematical models might overlook.

Case Study: HBT60 Concrete Pump

The methodology was applied to the main oil hydraulic system of an HBT60 concrete pump across a three-level hierarchy:

  • Level 1: Solenoid and hydraulic valves.
  • Level 2: Relief valves and control circuits.
  • Level 3: The main pump, cylinders, and overall system pressure.
Schematic of the concrete pump main oil hydraulic system

The study utilized failure data from a concrete pump manufacturer, identifying mean failure probabilities for six key components:

  • Main cylinder (x1): 0.057
  • Main pump (x2): 0.0067
  • Relief valve (x3): 0.0142
  • Control circuit (x4): 0.0192
  • Solenoid valve (x5): 0.0017
  • Hydraulic operated valve (x6): 0.0017

These mean values were converted into trapezoidal fuzzy numbers to account for uncertainty during the T-S algorithm processing.

Results and Insights

The model analyzed the top event: "Main oil system without pressure." It defined three states of severity: 0 (no fault), 0.5 (mild fault/instability), and 1 (complete failure).

  • Mild Fault (0.5): The probability of the system entering a semi-fault state was calculated as a fuzzy range between 0.0023 and 0.035.
  • Complete Failure (1.0): The probability of total system failure was calculated as a fuzzy range between 0.139 and 0.229.

These results demonstrated that the probability of a total system failure is significantly higher than the failure of any single minor component, reflecting the cumulative nature of mechanical wear. The findings were in "good agreement" with actual field data, validating the T-S fuzzy approach.

Conclusion and Future Work

The T-S fuzzy fault tree provides a robust framework for diagnosing complex hydraulic systems by embracing uncertainty and severity levels. While effective, the model faces challenges such as "rule explosion" in larger systems and a reliance on subjective expert experience. Future research could integrate real-time sensor data to update fuzzy probabilities dynamically, further enhancing the model's predictive accuracy over the machine's lifecycle.

Summary updated

Abstract

For improving the reliability of Concrete pump, fault diagnosis is proposed based on the main oil hydraulic pumping system T-S fuzzy fault tree method. The proposed method used fuzzy numbers to describe the probability of failure of hydraulic components and systems to address the failure probability of uncertainty; described by T-S fuzzy door contact between events, to solve the ambiguity and failure mechanisms associated events; fuzzy numbers description of the problem severity, considering the effect of the degree of fault on the system. The method is applied to the fault diagnosis of pumping concrete main oil pump hydraulic system, and achieved good results.
BibTeX
@misc{zhao2017reliability,
  author = {Qing Zhao and Lijun Li and Hongbin Tang},
  title = {{Reliability Research of Concrete Pump Main Oil Hydraulic System Based on T-S Fuzzy Fault Tree}},
  year = {2017},
  doi = {10.2991/iccia-17.2017.167},
}
Figure