{"id":6738,"date":"2023-11-06T12:22:21","date_gmt":"2023-11-06T10:22:21","guid":{"rendered":"https:\/\/supperundsupper.com\/?post_type=avada_portfolio&p=6738"},"modified":"2023-11-06T12:22:51","modified_gmt":"2023-11-06T10:22:51","slug":"industrial-anomaly-detection-in-manufacturing","status":"publish","type":"avada_portfolio","link":"https:\/\/supperundsupper.com\/en\/usecases\/industrial-anomaly-detection-in-manufacturing","title":{"rendered":"Industrial Anomaly Detection in manufacturing"},"content":{"rendered":"

<\/div><\/div><\/span>

Industrial Anomaly Detection in manufacturing<\/h1><\/span>
<\/div><\/div><\/div>

a Supper & Supper Use Case<\/p>\n<\/div><\/div><\/div><\/div><\/div><\/div><\/div><\/div>

<\/span><\/div>
<\/div><\/div><\/span>

Project Goal<\/h2><\/span>
<\/div><\/div><\/div>

The goal of this project was to detect anomalous behavior in machinery as well as mechanical and industrial equipment without prior information<\/strong> about what anomalous behavior consists of.<\/strong> Preemptively detecting anomalies in manufacturing and production processes paves the path towards a more efficient future for industries and factories. In this project, we used state-of-the-art machine learning tools to ensure precise anomaly detection in industrial processes, allowing for early identification of such anomalous behavior.<\/strong><\/p>\n<\/div>

<\/div><\/div><\/span>

What are anomalies in manufacturing?<\/strong><\/h3><\/span>
<\/div><\/div><\/div>

Anomalies in manufacturing refer to deviations in the operation of a (manufacturing\/industrial engineering\/technical) system\u2019s operation from its intended or normal behavior. Such deviations can decrease performance<\/strong>, leading to instabilities<\/strong>, security issues, defects,<\/strong> and even system failure.<\/strong> Given the intricate dynamics of these systems, pinpointing the causes of these anomalies can be challenging.<\/p>\n<\/div>

<\/i><\/i><\/span>Dataset used<\/span><\/a><\/h2><\/div>
\n
\n

The data consisted of multivariate<\/strong> time<\/strong> series<\/strong> collected from sensors installed on a machine test bed. Anomalous data was obtained from experiments in which the machine\/system setup was deliberately manipulated<\/strong> over certain periods of time. The algorithms used work equally well when data is available only from the normal behavior of a control system – in which case anomalies in the datasets can be synthetically created to test the anomaly detection approach. The training set consists of 8,125 data points, of which 337 are anomalous.<\/strong><\/p>\n

If we know which data points are anomalous, we can take advantage of this information using (semi-)supervised learning methods. In most applications, however, this information is unavailable, and we need to apply unsupervised learning. For this project, we used unsupervised learning methods,<\/strong> which means that the labels of the anomalous data points in our dataset are only used to test the performance of our methods, not to train our models.<\/p>\n<\/div>\n<\/div><\/div><\/div>

<\/i><\/i><\/span>Challenges<\/span><\/a><\/h2><\/div>
\n
    \n
  1. One major challenge is that there are only few <<5%)<\/strong> anomalous data points in the training set, , presenting a scarcity of information sources to learn from.<\/li>\n
  2. Furthermore, there is a tradeoff between the number of false positives (falsely detected anomalies) and false negatives (missed anomalies)<\/strong> when choosing the best algorithm. In this project, as in many industrial applications, the cost of a missed anomaly is higher than that of a false alarm. Therefore, the optimal method needs to minimize false negatives while maintaining a good overall performance.<\/li>\n
  3. Lastly, many established machine learning methods are computationally expensive and time-consuming. The optimal method should be efficient and lightweight<\/strong> enough to run on embedded\/edge devices.<\/li>\n<\/ol>\n<\/div><\/div><\/div>

    <\/i><\/i><\/span>Applied Methods (Implementation)<\/span><\/a><\/h2><\/div>
    \n

    We approached the problem by testing established machine learning methods for anomaly detection (OneClassSVM, iForest) against state-of-the-art models that address the above challenges (ECOD, COPOD).<\/p>\n