Predicting Failure Using Machine Learning and Statistical-Based Method: A Production Machine Case Study

Aug 5, 2026

Upload by Rowan

This research investigates failure detection models that combine machine learning and statistical methods to reduce unplanned downtime in a food production company. Sensor data is used to identify early failure symptoms, capturing temporal and sequential dependencies through an LSTM autoencoder (LSTM-AE). Performance is compared with a classical multivariate EWMA statistical approach. Both models detect failures, while LSTM-AE lowers false alarms and issues true alarms with a longer time-to-failure. Results support the feasibility of failure prediction using limited data and improved operational productivity.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages13

Reading Time33 min

AccessFree to download

Introduction

Maintenance and downtime challenges

Data-driven predictive failure detection

Case study objectives and structure

The Data

DOI: 10.12962/j20882033.v36i1.22501 | Submitted 01-30-2025; | Reviewed 03-11-2025; | | --- | --- | Accepted 03-18-2025 ORIGINAL RESEARCH Predicting Failure Using Machine Learning and Statistical-Based Method: A Production Machine Case Study Effi Latiffianti* | Stefanus Eko Wiratno | Samuel Aditya Christianta 1Dept. of Industrial and Systems Engineering, Institut Teknologi Sepuluh Nopember, Surabaya, Indonesia Correspondence *Effi Latiffianti, Dept of Industrial and Systems Engineering, Institut Teknologi Sepuluh Nopember, Surabaya, Indonesia. [Email: effi_l@ie.its.ac.id](Email: effi_l@ie.its.ac.id) Present Address Gedung Teknik Industri, Kampus ITS Keputih Sukolilo, Surabaya 60111, Indonesia Abstract This research investigates the applicability of failure detection models based on machine learning and statistical approaches to reduce unplanned downtime in a food production company. Sensor data is utilized to for identifying early failure symptoms. To capture temporal and sequential dependencies in time-series data, we employ oneof potential network based method so called the Long Short Term Memory (LSTM) Autoencoder. Furthermore, we contrast the performance of the result with the traditional statistical method, the multivariate Exponentially Weighted Moving Average (EWMA) . While both models successfully detected all failures, LSTM-AE demonstrated superior performance by reducing false alarms and providing true alarms with a longer time-to-failure. The findings highlight the potential of leveraging limited data for failure prediction, demonstrating the effectiveness of both models in detecting anomalies while emphasizing their role in enhancing productivity through early failure detection. KEYWORDS: Anomaly detection, failure, fault, Long Short Term Memory Autoencoder, Multivariate Exponentially Weighted Moving Average. 1 INTRODUCTION Maintenance is crucial for optimal performance and lifetime of industrial infrastructure and equipments. In general, it includes regular inspections, repairs, strategic asset management, and life cycle planning to reduce risks, prevent failures, and improve performance. As industries adopt new technologies, effective maintenance strategies are vital for maintaining operational efficiency and competitiveness. Unplanned downtime due to sudden failures and emergency maintenance affect around 90% of organizations in European companies [1], with traditional maintenance methods often failing to prevent these disruptions [2] . Even with planned maintenance, unexpected breakdowns can still occur, leading to inefficiencies. Detecting a failure before its occurrence can prevent sudden machine breakdowns, which contribute to significant production losses. Early detection allows for proactive maintenance, reducing downtimes and managing costs more effectively. Data-driven failure detection models have been extensively applied to wind turbines, leveraging sensor data collected through Supervisory Control and Data Acquisition (SCADA) systems to enhance operational reliability [3–6] . Similarly, recent studies by [7] and [8] demonstrate the potential of predictive models in forecasting failures within manufacturing machinery. Building on these advancements, this study aims to implement and evaluate failure detection models within a selected case study to further explore their applicability and effectiveness. Our case study focuses on a production machine of a leading poultry-based food industry in Indonesia. The maintenance team has observed several instances of abnormal machine behavior, including increased noise levels, reduced performance, and elevated temperatures, which have resulted in breakdowns. Some of these breakdowns cause the production line to come to a complete halt, leading to significant financial losses. In some cases, the unavailability of replacement parts further prolongs the downtime, exacerbating the impact on production. This highlights the critical

FAQs

What goal does the study pursue for the production company?

The study aims to apply and evaluate failure detection models to reduce unplanned downtime by detecting early failure symptoms from sensor data.

Which machine learning method is used to model time-series dependencies?

The study uses an LSTM autoencoder (LSTM-AE) to capture temporal and sequential dependencies in the sensor time-series data.

How does LSTM-AE performance compare with the statistical EWMA method?

Both models detect all failures, but LSTM-AE achieves superior results by reducing false alarms and providing true alarms with a longer time-to-failure.

Aug 5, 2026

Upload by Rowan