Stream processing

The goal: Detect and respond to system inefficiencies that constrain or interrupt production.
The solution: A rules-based engine that detects process anomalies, alerts users and automates corrective action.

Capturing and combining time-series data with complex event processing, the Savigent® Platform monitors all data generated by systems, machines, tools and sensors on the shop floor — and measures, predicts, prevents, corrects and optimizes production in real time.

The Platform also captures data for engineers, analysts, data scientists and AI systems, enabling deeper analysis. Savigent spots trends and institutionalizes corrective action across shifts, plants, departments and global operations.

A different kind of stream processor

Traditional data stream processors evaluate data in batches, with time delays your business cannot afford.

Using a sliding time window, Savigent processes and evaluates data in milliseconds. The variable data generated by enterprise and control systems is historized for analysis and continuous improvement.

Savigent stream processing

  • Automates responses to faults, process drift, equipment malfunction or other anomalies that reduce efficiency
  • Sends immediate alerts and automatically responds by self-correcting or shutting-down the operation to avoid losses and protect workers and assets
  • Recognizes conditions that trigger automated out-of-control action plans (OCAP), corrective and preventive action plans (CAPA) and other process and quality models
  • Supports a wide variety of analytics, including sampling, conditioning, filtering, SPC rules, slope and oscillation algorithms and many other methods used in trend analysis or investigations
  • Provides multi-variable analysis, giving manufacturers the capability to infer data that is not directly collected
  • Empowers extended analysis and machine learning capabilities

Dynamic sampling

In tandem with work orchestration, the Platform works to ensure product conformance.

  • The Platform can control the number and parameters of samples measured based on the performance of a process
  • Run-to-run control blocks analyze data following each new measurement and determine the set-point for the next run based on simple or complex models

Your continuous improvement tool

In addition to fault detection and predictive maintenance, the Platform captures and notes positive events that form a base camp for your best practices and repeatable standards.

Learn more