# Keep the process in control and the machines serviced

> Statistical process control and preventive maintenance use the data PLCGateway already collects: control charts from the history of a measurement, maintenance counters from a running bit or a cycle counter. Both raise alarms, both reach the notifications, and both can be set up by an AI assistant. SPC and maintenance form the Quality & Maintenance module of PLCGateway SCADA and Embedded; the trial includes it.

## Control charts from live history

Pick a measurement and PLCGateway computes the chart on the server: I-MR for individual values, X̄-R for subgroups, p for the proportion defective per hour, shift or day, c for defects per period. Control limits are calculated from the data or fixed, specification limits give Cp, Cpk, Pp, Ppk and the out-of-spec ratio. The Western Electric and Nelson rules 1 to 8 mark the point that completes a pattern.

- Two-pane charts with the range or moving range below
- Rule violations marked with the rule number
- CSV, PNG and PDF export from the dashboard

![SPC report dashboard with I-MR and X̄-R control charts, limits, Cpk and rule violations](https://plcgateway.com/assets/img/shots/spc-1500.webp)

## SPC monitors that watch continuously

A monitor evaluates a chart over a sliding window at a fixed interval and writes the result to virtual tags: a violation flag, the rule that fired, the mean and the Cpk. An alarm with acknowledgement follows the flag, so an out-of-control process shows up in the alarm list, in the notifications and on the operator screen, not only in the quality office.

- Windows from minutes to days, intervals from 10 seconds
- Alarm priority and group per monitor
- Run now with the full chart and the capability figures

![Quality page with continuously evaluated SPC monitors and their last result](https://plcgateway.com/assets/img/shots/quality-1500.webp)

## Maintenance plans from machine data

A plan belongs to a machine and counts since the last service: run hours from a running bit, cycles from a counter, calendar days, or the time a condition such as a vibration level is true. Progress, remaining time and a due flag are virtual tags for screens, an alarm fires when the service is due, and the operator records the service with a note; the interval restarts.

- Warning threshold before the interval is reached
- Instructions and priority per plan
- Work orders kept in a data source that screens can list

![Maintenance page: plans with run hours, cycles and calendar days, progress bars and work orders](https://plcgateway.com/assets/img/shots/maintenance-1500.webp)

## Work orders to the CMMS

- **Webhook to the maintenance system:** Fiix, MaintainX, UpKeep, Limble, Odoo or a ticket system receive a JSON work order through a notification channel with the API key in its headers.
- **Generic or custom body:** Send the standard fields or a template with placeholders shaped for the receiving API; the closing event follows when the service is recorded.
- **E-mail as a fallback:** Systems that create work orders from incoming mail get the same order on the maintenance mailbox.
- **Set up by AI:** An assistant defines plans and monitors from a conversation: 'lubrication every 500 run hours of the extruder', 'watch the melt temperature with an I-MR chart'.

## At a glance

| Function | Input | Output |
|---|---|---|
| Control chart | Logged measurement, optional counters for p and c charts | Points, limits, rule violations, Cp / Cpk / Pp / Ppk |
| SPC monitor | Chart definition, window, interval | Virtual tags violation / rule / mean / Cpk, alarm |
| Maintenance plan | Running bit, counter, calendar or condition expression | Virtual tags due / percent / remaining, alarm, work order |
| CMMS connection | Notification channel, generic or custom JSON | Work order created, pushed and closed in the CMMS |

## Quality and maintenance: common questions

### Can AI define SPC monitors and maintenance plans?

Yes. 'Watch the melt temperature with an I-MR chart and alarm me on rule 1 and 4', 'lubrication of the extruder every 500 run hours, work order to the CMMS', 'which machines are due for service this week' are all MCP tool calls the assistant makes for you. It reads the tag descriptions and units to pick the right signals, explains the chart result in plain language and records a service or pushes a work order when an operator asks.

### Which control charts does PLCGateway offer?

I-MR (individuals and moving range), X̄-R (subgroup means and ranges with subgroups of 2 to 10), p (proportion defective per period), c (defects per period) and means of equal time slots. Limits can be calculated or fixed; specification limits give Cp, Cpk, Pp and Ppk; the Western Electric / Nelson rules 1 to 8 are checked.

### Can PLCGateway replace a CMMS?

No, and it does not try to. It keeps the plans next to the machine data, knows when a service is due from real run hours and cycles, and hands the work order to the CMMS you already use through a webhook or e-mail. The CMMS keeps the spare parts, the technicians and the history.

### Which CMMS can be connected?

Any system with a REST endpoint for creating work orders or a mailbox that turns mails into orders: Fiix, MaintainX, UpKeep, Limble, Odoo maintenance, Jira or ServiceNow style ticket systems and in-house APIs. The JSON body can be reshaped with a template for the receiving API.

### Do operators need the quality office to see a violation?

No. The SPC monitor's violation flag is a virtual tag with an alarm, so the operator screen can show a lamp, the alarm list shows the rule, and the notifications send it to the people responsible.

## See your process in control

Send us a measurement history or a running bit of a machine; we show the control chart and the maintenance plan on the demo plant. → https://plcgateway.com/contact

---
Source: https://plcgateway.com/quality-maintenance