Predictive analytics for manufacturing processes

Analyse time series data comming from production, combine with integrated systems (ie. ERP) and external data sources to execute predictions.
Provide users with prediction results as recommendations or provide strict user instructions.

Must have technology for

predictive quality management
predictive and preventive maintenance
predictive energy efficency (dynamic tarrifs)

Preventive and predictive maintenance

Minimize breakdowns and unnecessary maintenance costs in manufacturing planning
Collect all history data that may have direct or indirect impact on past machine breakdowns or malfunctions and predict future machine malfunctions
Provide maintenance teams real-time insights (including future preditctions with % of probabilities) comming from machines to take actions when needed
Create actionable tasks for those responsible for maintenance to dispatch tasks and track statuses (including acceptance workflows where needed)

Example:
-> based on the history of executions and maintenance steps, is is known the specific machine malfunction is primarily caused by 3 separate factors (2 direct and 1 indirect)

-> all 3 factore are defined for edge/critical values that when achieved may cause machine malfunction

-> all 3 factors are being monitored in the productions process and likehood of malfunction in % (dedicated algorithm) is being calculated and reported

-> when malfunction probability exceed a given level, maintenance team is provided with maintenance recommendation with a list of recommended mainteannce steps

-> steps execution and status is being tracked

-> execution history data base is being updated to improve future analytics

Predictive energy efficiency

tailor your production plans with dynamic tarrifs and changing media costs
predict energy costs for energy market or own renewables and provide manufacturing plan recommendations to minimize energy costs per unit produced
calculate energy consumption per unit produced and compare between production lines, shifts or factories to find out best practices

Example:

-> there are 3 types (A,B,C) of similar products manufactured in a given production site

-> product C (due to its excesive weight comparing to A and B) requires additional heating cycle (50% additional heating media costs per unit produced vs A and B)

-> based on the history of executions - unit costs of energy, per kilogram of product is being calculated

-> production is running 24/7 and it is possible to change production plans between products A,B,C within a day

-> based on weather forecasts, predicting energy costs (energy market or own renewable source) for next day production, production of Product C is setup for production to the lower predicted energy costs

Predict production quality based on history execution

Minimize waste based on the history of executions
Raw materials quality, production site temperature or humidity, best mix of ingredients and multiple other factor may impact expected waste
Create production process "digital twin" to simulate multiple scenarios and "what if" analysis

Example:
-> new manufacturing order is comming that comprise of raw material mixtures, forming, cutting and drying processes
-> raw materials needed for the order are close to expiry date and raw material silo temperature is acceptable but abnormaly high (due to external conditions)
-> due to specific conditions, based on the history of executions, all similar executions from the pasts are analysed to identify these where wast level was the lowers
-> produciton operators are provided with produciton setup recommendations to minimize waste with a current conditions
-> execution data is being collected to improve further analytics

About this Offer

This offer relates to a SmartRDM pilot implementation for a selected manufacturing process or subprocess (acquisition and management of up to 1.000 data points).

About Smart RDM product

developed by ConnectPoint to shorten real-time production monitoring implementations
deployable on cloud, private cloud or on-prem
empowered with Azure analytics and hyper scalable technologies for multiparallel calculations
trused by Danone, Twinings, Mondi, Veolia and many more

SmartRDM - Predictive Analytics for Manufacturing data

durch ConnectPoint Sp. z o.o.