On-premises platform for closed-loop AI process control in mineral processing plants

Conundrum Industrial Ltd

Cambridge (UK)New on Amplifa

2018

10Closed-loop deployments at five sites

10–30 minRetraining interval of the process models

About Conundrum

Conundrum Industrial Ltd is a software provider based in Cambridge (UK) that develops an on-premises platform for closed-loop advanced process control (APC/MPC) in mineral processing plants. The platform runs inside the customers' plant network and builds on the existing control system (DCS, PLC, SCADA) via OPC UA, MQTT or SQL, or replaces it. The process models are retrained every 10–30 minutes, so no periodic re-identification or step tests are required. Connected process stages such as comminution, grinding, flotation and thickening are controlled as one system rather than individually. The focus is on iron ore beneficiation as well as base and precious metals such as copper, nickel, platinum group metals and gold, including sulfide concentrators and smelting processes. Customers include a large Australian iron ore processing company and three major sulfide concentrators (copper-gold and nickel-copper).

Site: Cambridge, Cambridgeshire CB1 2GE, United Kingdom
Processes: Closed-loop AI process control (APC/MPC) for processing plants, Value Estimate (free of charge), Data Assessment (Historian data review under NDA), Closed-Loop Pilot, Platform with Python SDK and graphical interface, Soft Sensors, Pipeline framework (YAML-DAGs), Training/enablement of the customers' own engineers
Materials: Model Predictive Control, Soft Sensors, Python SDK & graphical interface, Pipeline Framework
Industries: Mineral Processing, Iron Ore Beneficiation, Base Metals (Copper, Nickel), Precious Metals (Platinum Group Metals, Gold), Sulfide Concentrators, Smelting Processes (Flash Smelting)

01
Value Estimate
02
Data Assessment (Historian data review under NDA)
03
Closed-Loop Pilot
04
License & gradual expansion
05
Training/enablement of the customers' engineers

Platform & Methods

MPC

Model Predictive Control

Core control algorithms of the platform for closed-loop control of connected process stages.

Soft Sensors

Model-based estimation of measured variables that are difficult to capture online.

SDK

Python SDK & graphical interface

Tools that enable the customers' engineers to develop new controllers, soft sensors and pipelines themselves.

YAML-DAGs

Pipeline Framework

YAML-defined data pipelines in near real time for cleansing, estimation, identification and control.

Locations

Cambridge

Registered Office (C/O Mishcon De Reya, Four Station Square)

Services

**Conundrum Platform: Plant control with control-room dashboards**Closed-loop APC · Mineral processing

**Platform connection to control system and historian**On-premises deployment · OPC UA/MQTT/SQL

Inside the production

Conundrum – Closed-Loop AI Process Control

Industries

Mineral ProcessingIron Ore BeneficiationBase Metals (Copper, Nickel)Precious Metals (Platinum Group Metals, Gold)Sulfide ConcentratorsSmelting Processes (Flash Smelting)Mineral Processing

Additional services

Closed-loop AI process control (APC/MPC) for processing plantsValue Estimate (free of charge)Data Assessment (Historian data review under NDA)Closed-Loop PilotPlatform with Python SDK and graphical interfaceSoft SensorsPipeline framework (YAML-DAGs)Training/enablement of the customers' own engineers

Reference customers

Large Australian iron ore processing companyThree major sulfide concentrators (copper-gold and nickel-copper)

History

2018
Company registration in England and Wales under number 11540644 (derived from company number)