Get in Toucharrow

CONTACT US

Predictive Maintenance

Case Study-1

Client: One of the Leading Oil and Gas Producer Company

Overview:

The client is one of the leading oil and gas producers in the Asia-Pacific region, serving the energy needs of homes, businesses, and major industries across Asia. The Company reports annual revenues of USD 2 billion.

Business need:

When a critical asset fails in one of the Company’s operations, the result can be lost revenues and more time on the road for engineers. Could they find a way to identify faults before failures occur?

Solution:

The client developed predictive models that ingest data from our IoT-connected assets and many other sources, giving early warning of potential failures and opportunities to improve uptime and production.

Benefits:

By alerting engineers to asset downtime, the analytics solution helps cut travel time, resolve problems faster and boost production output – potentially saving in excess of USD 4 million each year.

 Case Study-2

Client: Industrial Equipment and Software Company:

Challenges:

Our client designs, manufactures, and leases industrial equipment and provides software to remotely monitor equipment operations. They looked to Faststream Technologies to accelerate the prototyping of intelligent features in the software platform that could reduce downtime of machinery by predicting and preventing faults.

Working closely with our client’s software engineering team, Faststream built a predictive analytics prototype that consists of 2 wire current sensors, an Anemometer, equipment sensor time-series data to predict and prevent machine downtime. The system alerts field teams about units at risk of faulting, so they can proactively take action before any failure. The client’s team is able to continue the work on their own, maintain the code, and conduct further experiments using the data processing pipeline and machine learning framework we created.

One sensor that we used was an inertial sensor that includes a Machine Learning Core (MLC) and a Finite State Machine (FSM). A revolutionary aspect of this sensor is that it has an embedded Machine Learning Core. Our team could configure specific parameters of the decision tree with  Weka, an open-source collection of machine learning algorithms.

 

Solutions:

In the ramp-up phase, Faststream Technologies met with the project’s business sponsor and the software development organization to more explicitly define the project’s business and technology objectives. We helped our client translate business objectives into a product specification.

(1)DATA ENGINEERING AT SCALE

We audited our client’s data to better understand their data sources, quality, and resolution. The bulk of the ETL effort involved merging multiple data sources in varying formats from the client’s data lake. We devised a data engineering strategy that sourced terabytes of data—including real-time streaming sensor data, hardware-specific demographic information, human-generated maintenance reports, and external weather data.

(2)DATA PREPARATION

After assessing the data and defining predictable targets, we performed feature engineering on the data stream to create appropriate inputs for the time-series forecast problem. We varied the look-back and prediction horizon windows, and carefully created training and validation data sets so as to avoid data leakage.

(3)MODEL ENGINEERING

We followed our standard process to evaluate multiple model architectures, from logistic regression to tree-based ensemble techniques to neural networks. We used convolution neural networks without feature engineering as a baseline for accuracy but settled on two classes of tree-based models (random forests and gradient boosted trees) because they demonstrated better performance, were easier to tune and were also more interpretable. We worked with the client’s engineers to define expert features (e.g., pressure and temperature ranges), to optimize the model  accuracy, and to interpret the output of the model (e.g., feature importances)

(4)BUSINESS INTEGRATION

After optimizing the hyperparameters for a family of models, we created a prediction job that

updated a database with daily predictions.

RESULTS:

Our client now has a predictive maintenance prototype. The software engineering team has the

tools and skills to develop the product further and deploy it to the field operations organization.

Case Study-3

Client: Petrochemical Company

Faststream Technologies IoT-powered predictive maintenance solutions have also made an indispensable impact on oil refining and petrochemical companies.

Problem:

The major challenge faced by the oil refinery Company is that the physical inspection of the equipment located on deep-ocean floors is a very dangerous and inefficient process. Therefore the oil refining industry has always been in need of a better method not only for predictive maintenance, to identify potential failure, but also for better asset tracking. Oil fields generally have assets fitted with sensors, to assimilate the vast amount of data. But most of this data is never utilized.

Solutions:

Faststream Technologies advanced predictive maintenance solutions helped mitigate the challenges associated with the huge volume of data generated by Oil refining and petrochemical companies.  Our big data analytics ensures a huge volume of data is managed in a scalable and cost-effective way thus shooting down the maintenance cost. These advanced solutions also incorporate methods like data storage in the central repository and efficient remote monitoring. The solution was designed to compare the real-time data with historical failure rate models and identify potential equipment failures. This helped inefficient resource maintenance, without the need for equipment replacement due to permanent damage.

 

 Case Study-4

Client: Smart Coolers/Chillers with Predictive Maintenance

Problem: 

The client company deals with commercial-grade Coolers that are considered high-value assets. An equipment failure not only causes inconvenience for their client but fetches penalties when Cooler equipment breaks down. The goal of this manufacturer was to improve profitability by improving the up-time of this equipment and reduce the cost of maintenance/downtime penalty.

Solution:

The created solution to monitor the temperature and humidity in the chillers, used for storing perishable. Alarms were generated upon fluctuations in the temperate and humidity and sent to the cloud via 4G/2G/NB-IoT with OpenCPU running on Cortex-M0 for seamless data transaction.

Also, with the help of Faststream Technologies, the manufacturer company maintained an elaborate database of the installed base as well as the history for each installed equipment. The history contains records of all the maintenance work including parts that have been replaced for each piece of equipment. In addition, all the support tickets relating to when equipment failed and where the observations are also captured in the system. The data required to develop the model was already in place.

 

Model:

Our team conducted a careful analysis of each piece of equipment, its history of failure, intervention by the organization, and outcomes. Using machine learning the team was able to develop and train a model that can predict which equipment has the highest likelihood of failure, these equipment get queued for proactive maintenance thus avoiding equipment failure, increasing up-time, and reducing cost.

Decision:

Faststream’s predictive model has enabled the manufacturer company to differentiate itself with the ‘predictive maintenance model’ and it has improved up-time and profitability for the Company.

 

Case Study-5

Client: International Operating Mechanical Engineering Company

Problem:

The client Company is pushing ahead with the digitalization of its production. One starting point:  The data-supported optimization of maintenance. The available database for this purpose contains, among other things, sensor readings that contain error and malfunction messages at different time intervals. This provides the machine manufacturer with comprehensive information, since more than one hundred machines, each with more than 200 sensors, continuously supply data.

Aim:

The primary aims of Faststream’s predictive maintenance for the laser machines were as follows:

  • Introduction of data science to utilize the data
  • Analysis of lasers whose sensors produce countless machine data on a daily basis
  • Transparent and descriptive visualization of machine data
  • Cross-role workflows to support business processes between development, service, after-sales, and external as well as internal data scientists
  • Introduction of algorithms for pattern recognition of defect images and prediction of future failures

 

Solution:

In the primary stage, Faststream Technologies helped to train its own data science team. The trained team thus combined domain and statistical knowledge and was able to successfully identify and implement the first use cases. In order to perform the complex analyses of the machine data, the open-source scripting language R was used, which has a unique range of functions for analysis,  forecasting, and visualization and was used by the engineers at the company after a short time.

The use cases consisted of evaluating the existing machine data sets, examining them for anomalies and failures, mapping the results, and predicting future problems.

The use cases consisted of evaluating the existing machine data sets, examining them for anomalies and failures, mapping the results, and predicting future problems.

Result:

As a result, a maintenance strategy is implemented to the client company that detects possible errors in advance and thus prevents unforeseen machine failures. This enables the planning of optimal maintenance and creates new value-added services based on data and algorithms.

 

Case Study-6

       Industrial Asset Tracking and Predictive Maintenance

Challenge:

Create a solution to Predict and locate the usage pattern of the express painting tools like an automatic sander, roller, washer, mixer which are used at the client-side of the responsible company.

Solution:

We used a combination of BLE and Cellular Communication to communicate with the sensors attached with the above tools and pushed the data to the cloud via MQTT broker, storing the data in  MongoDB considering the heterogeneous nature of data, resulting in a comprehensive package for the business problem. We had to manage a plethora of devices where AWS IoT and AWS Device  Management came into action and we implemented a seamless way of Device Firmware Updation and Device Management.

Result:

As a result, a maintenance strategy is implemented to the client company that detects possible errors in advance and thus prevents unforeseen machine failures. This enables the planning of optimal maintenance and creates new value-added services based on data and algorithms.

March 13, 2020

Digital Twin


Digital Twin
Comments 0
March 13, 2020

Integrated Home Automation Solutions


Integrated Home Automation Solutions
Comments 0