Pulse — an intelligent production optimization system for the field

01Data acquisition

Continuous pressure telemetry: wellhead, flowline, annulus

Every well carries three self-powered LoRa pressure sensors: wellhead, flowline and annulus. Data flows into the database around the clock with no operator involvement.

• Import from field formats. SQLite and CSV exports are accepted (including local encodings and decimal commas) — the system digests the data as it comes.
• Sensor registry. Sensors are mapped to wells and roles (wellhead / flowline / annulus), with a history of reassignments and data-freshness monitoring that raises an alert when a channel goes silent.

02Signal conditioning

The signal is cleaned layer by layer, with the raw data preserved

Raw readings are never overwritten. Every calculation is built on four layers: raw signal → automatic cleanup → manual operator corrections → smoothing.

• False-zero filter. Sensors occasionally send a «0» instead of a real value — the system detects and excludes such points at every level (zero pressure on a gas well is physically impossible).
• Smart gap filling. Short gaps (threshold configurable, 10 minutes by default) are interpolated; long ones honestly remain «holes» — the system does not fabricate data.
• Operator masks. The engineer manually flags intervals of unreliable data (a closed line, hydrate, a gauge failure, spikes during blowdowns) — the field operator's knowledge outranks the sensor. The result — false data is excluded from the rate calculation. All reports and rate calculations are built strictly on the corrected layer.
• Correct ΔP aggregation. The differential pressure is filtered row by row before averaging — this rules out mixing producing hours with downtime hours, which would inflate the calculated rate.

03Visualization

The whole well on one screen

The well page is the engineer's workspace: synchronized charts of pressure, differential and gas rate, events on a timeline, statuses and equipment.

• Synchronized charts. Wellhead and flowline pressure, ΔP and calculated rate move together: zooming and panning are linked, and the cursor shows a slice across all axes.
• Events right on the chart. Treatments, blowdowns, status changes — markers on the time axis; clicking a table row highlights the corresponding range on the chart.
• Global timeline. A summary screen of all wells: who is in observation, who is in optimization, where the events are, where a sensor has gone silent.
• Field map. Wells with coordinates and statuses on a map — the basis for planning runs and crew routes.

04Events · Sync · Telegram bot

Events are logged where they happen — in the field

The field operator sends an event to the Telegram bot — a treatment, a blowdown, a measurement — and it immediately becomes part of the whole picture: it lands on the charts, in the analytics and in the reports.

• Event log. A single feed: treatments (type and quantity), blowdowns (the full three-phase cycle: venting → pressure build-up → restart), measurements, well work.
• Life-cycle stages. Observation → Adaptation → Optimization: statuses are kept in the log, shown on the timeline and drive the logic of reports and financial documents.
• Automatic well sync. A new well appears in the system automatically on its first event — with coordinates and equipment mapping.
• Group broadcasts. Messages from the dashboard to private chats and Telegram groups; configurable recipient subgroups (supervisors, engineering staff, management) and a delivery log.

05Alerts · Trends · Routes

The system spots the problem before it turns into downtime

The well reports on itself. Acceptable ranges and notification recipients are set for every parameter — when a limit is crossed, the right people get an alert in Telegram within minutes, not at the next route visit.

• Configurable ranges. Each well has its own thresholds for pressure, differential and rate. Recipients are chosen by group: field supervisor, engineering staff, management. Different events go to different people.
• Sensor silent. Telemetry freshness is monitored: if a channel stops sending data, an alert goes out with anti-spam protection (cooldown) — the person on duty learns about the blind spot at once.
• Well loading up on itself. A falling differential and a stalled inflow are recognized from the pressure signature — the «well has loaded up» alert arrives without a trip to the pad.
• Liquid-loading forecast. From the degradation trend of rate and ΔP the system calculates when the well will approach a shut-in — the problem is visible days ahead, not after the fact.
• Treatment and blowdown recommendation. The forecast plus the accumulated treatment-interval statistics indicate when to act: when it is time to treat, when to blow down. Chemical is consumed on demand, not on a schedule.
• Route by priority. Field map + event feed = the run route: critical wells first, the rest picked up along the way. The crew drives where it is needed most.

The route can be built two ways — the system does both. «By event priority»: a failure (P1) and a shut-in (P2) always jump the queue, then blowdowns (P3) and treatments (P4). «By shortest path»: pure mileage optimization — the crew picks up the nearest points; the cost of that choice is immediately visible on the map — a distant shut-in well waits longer and has time to escalate into a failure. Switch the mode while the run is under way — the route is rebuilt instantly from the crew's current position.

A failure here is not a separate event but the state that follows a shut-in the crew did not reach in time: in the simulation well 130 shuts in during the run, and if the crew does not make it, the shut-in escalates into a failure before your eyes. Any significant event during the run — a new shut-in, a blowdown request — recalculates the live route; serviced points turn green.

06Mobile app

Mobile app: the well opens on your phone

The mobile app syncs with the shared database and the event log: the field operator on location and the manager in the office see the same numbers — minutes apart, not a day apart.

• Events from the job site. A treatment, a blowdown, a measurement — created right from the phone at the well and immediately entering the shared database, the charts and the reports.
• Well regime chart. Wellhead and flowline pressure, ΔP, rate and life-cycle stages — the same charts as in the dashboard, adapted to a phone screen.
• Current values and trend. The latest readings for every well and the change over the period: growth, degradation, stability — visible at a glance, without a laptop.
• Alerts on the phone. Notification ranges are configured right in the app — threshold breaches, sensor silence and signs of liquid loading arrive as a push notification to whoever is responsible for the well.
• One account. The same roles and permissions as in the web version: everyone sees their own area and their own wells.

07Rate calculation

Gas rate from physics, not from a lookup table

Gas rate is calculated from the differential pressure across the choke for each well, with what-if scenario analysis and a full log of engineering corrections.

• Calculation scenarios. Several parameter sets for one well, results compared side by side, daily aggregates and cumulative volume.
• Corrections without spoiling the data. Excluding intervals, interpolation, a manual value, range limiting — all applied on the fly, while the raw data stays untouched.
• Blowdown by phase. Every cycle is identified phase by phase: venting → pressure build-up → restart. Duration and pressures are recorded for each phase (before / bottomhole / after) — the blowdown stops being a black box.
• Blowdown quality over time. Cycles are compared with each other and month over month: how the total duration changes, whether the pressure build-up time is growing — an early sign that the well condition is deteriorating and a reason to revisit the operating regime.
• Gas lost to blowdown. Gas vented to atmosphere is calculated for every cycle and totalled for the period — the real cost of each blowdown is visible in the report and is taken into account when choosing between a blowdown and a treatment.
• Auto-detection and downtime. Cycles are recognized both from field events and from the pressure signature; the entire blowdown cycle is correctly counted as downtime (pressure build-up is not production). Downtime is distinguished from missing data.
• Regression control. Reference calculations on benchmark wells are re-run on every change to the algorithms — the numbers cannot drift unnoticed.

08Chemical performance analytics

Every treatment gets a score: is the chemistry working or not

The system splits the well history into treatment intervals — from one treatment to the next treatment or blowdown — and scores each interval against seven metrics, combining them into a single score from 0 to 100.

• Response segmentation. On the smoothed differential the phases rise → plateau → decline are detected automatically: you can see how long the chemical effect holds.
• Metrics + Score. Rate gain and stability, duration of the effect, rate of degradation and other indicators — with configurable weights and thresholds calibrated against real data.
• Chemical comparison. A ranking of chemicals by well and across the field: the top formulation for each well is justified by numbers, not by impressions.
• Transparency for the operator. Every treatment-interval card shows how the Score was calculated: the method is open, so the decisions can be trusted.

09Reporting

Reports are assembled from blocks — and always match the screen

A step-by-step wizard walks the engineer through the chapters: customer data, observation stage, adaptation, period reports. Any analytical block — chart, table, conclusion — is attached to a chapter in one click.

• Period selection. Reporting periods are set flexibly: a calendar month auto-trimmed to stages, arbitrary dates, comparison with any previous periods of the well.
• HTML = PDF. A hard invariant of the system: the printed version is identical in content to the on-screen one — the same paragraphs, tables and charts. The customer sees exactly what the engineer saw.
• Print-ready quality. The PDF is built through LaTeX: a title page with signature and seal, pagination, appendices — the document is ready to hand over with no manual touch-up.
• Data snapshots. A saved block freezes the figures as of the moment of analysis — the report is reproducible months later, regardless of new data.
• Fast rebuild. A cache of completed chapters: a full report is reassembled in seconds instead of minutes.
• Auto-conclusions. Narrative conclusions (trend, performance, optimal operating window) are generated from the data with cautious wording and the option to edit by hand.

10Economics

Project economics

Production economics in real time. Pulse brings production and financial indicators together and shows the real performance of the project — from overall payback down to the result of each individual well.

• Project economics. Incremental production, revenue, costs, profit and payback for the project as a whole.
• Well performance. The contribution of each well to total production and to the financial result.
• Weak spots. Wells with low performance or negative economics come first on the list, together with the interventions applied.
• Field leaders. The best-performing wells and interventions come first on the list.
• Intervention assessment. The result before and after a workover, a regime change, adaptation or optimization.
• Trend over time. Changes in production, costs and profit across periods, with comparison against any earlier period.

1Project economicsIncremental production, revenue, costs, profit and payback

4Detailed well economicsFull breakdown of indicators, costs and result for the selected well

11Chemical inventory

Chemical stock tells you when to reorder and how much

Receipts, consumption by treatment events and remaining stock are reconciled automatically. The system forecasts when each chemical runs out and builds a purchasing plan.

• Real-time balance. Remaining stock for every chemical and every warehouse is updated from field events — with no manual reconciliation.
• Run-out forecast. From the average daily consumption the system calculates «days to zero» and the date by which the order must be placed (allowing for delivery time).
• Purchasing plan. A supply horizon is set (60 days, for example) — the system produces a «what and how much to order» list with priorities: critical / watch / normal.
• Top consumers. For every chemical you can see which wells consume the most — and whether that consumption is justified by the result.

12Access and reliability

Role defines access; the audit log records every action

• Roles and permissions. Administrator, engineer, observer: approving baselines and reports and editing reference data are limited to authorized users.
• Login log. Who worked with the system, when and from where — with visual statistics.
• Protected customer pages. Stable screens for external access are separated from the internal tools.
• Backup. Backup and restore-point procedures; raw data is immutable — any calculation can be repeated.

◆Summary

What you get