Water Quality Wire

Automation in water and wastewater now produces a second product alongside control. Each pump start, tank level change, and alarm acknowledgment can leave a timestamped entry in a historian, an event list, or a maintenance platform. Those entries are often treated as a byproduct of operations. For a technology group, they are closer to evidence, and the first step in using them well is deciding which question each record can answer.

Three Records That Get Blended

A useful separation has three parts. The first is the command record, which shows what the system was told to do: a setpoint changed, a pump enabled, a valve sent to a position. The second is the state record, which shows what sensors and equipment reported back, including levels, pressures, run status, and alarm conditions. The third is the human record, which shows who acknowledged an alarm, who placed a device in manual, and whether anyone made a local change at a panel. A command log can show that a valve was ordered to close while saying nothing about whether the valve actually moved.

The distinction matters because vendors and staff use the word visibility loosely. Nan Johnson's column in Water Finance and Management's tech perspectives column frames the shift with the phrase "Operational Visibility Has Become Just As Important As Equipment Control." The summary, which is the only part of the column this article relies on, describes an older view in which automation mainly improved how equipment ran, with pumps starting and stopping on their own and tank levels held steady. The newer demand is that staff can see the system's state clearly enough to judge it. That is a records question as much as a display question, because what a team can later reconstruct depends on what the system chose to retain.

Where Blending Causes Trouble

Confusion between records usually shows up after an event, not during one. A review might find a high-level alarm acknowledged at the console while the state record shows no matching change in inflow. Staff may read the acknowledgment as action when it was only attention. A setpoint change in the command record may be credited to an operator when it came from a scheduled sequence, a remote integration, or a technician working at a local panel. None of these readings is automatically wrong, but each assumes a relationship between records that the system may not actually establish.

Clock discrepancies make the problem worse. If the historian, the controller, and the alarm server take time from different sources or drift apart, the sequence of events can invert. A pump that appears to start before an alarm clears may simply reflect a clock offset. The fixes carry tradeoffs. Synchronizing clocks, retaining high-resolution data, and keeping full alarm event lists all cost configuration effort and storage, and that cost is justified only if someone will review the result.

What to Write Down Before the Next Change

Before an automation upgrade or a new remote-access arrangement, a team can write a short map. For each control point, name the record that captures commands, the record that captures state, and the record that captures human acknowledgment or override. Note where each record is stored, how long it is kept, and which clock stamps it. Note also which actions bypass the record entirely, such as local manual operation or a maintenance mode that suppresses alarms. This is less a documentation exercise than a test of whether the system can answer the questions it will eventually be asked.

The map also exposes tradeoffs that should be decided on purpose. Dense logging improves reconstruction but can bury operators in alarm noise and increase the review burden. Suppressed alarms reduce nuisance but leave gaps that look like silence. Central collection simplifies review but concentrates access and needs its own controls. A utility that chooses one design should be able to say which questions it has chosen not to answer.

Limits of This Analysis

This article does not assume how any particular platform handles acknowledgments, overrides, or time sources, since those depend on vendor configuration and local practice. Staff reviewing their own systems should confirm each record against the actual configuration rather than a diagram, and should treat a log as a claim about the system until its timing and sources have been checked.