Fault detection and diagnostics software we provide

How many faults are running in your buildings right now?

Stuck dampers, leaking valves, and overridden schedules waste energy for months before anyone notices. We build fault detection and diagnostics software that uses AI to find the root cause, quantify the waste, and tell your team what to fix first.

Fault Detection and Diagnostics Software_2880x1760

Why alarms don’t tell you what’s actually broken

A building automation system raises alarms when a value crosses a limit. It doesn’t explain why. A warm zone could mean a failed actuator, a bad sensor, or a chiller sequence fighting itself. Operators chase symptoms, and many faults never trigger an alarm.

Svitla builds fault detection and diagnostics (FDD) software that continuously reads your trend data. It tests equipment against engineering rules and learned baselines, isolates the likely cause, and estimates each fault’s cost. Your team gets a short, ranked list instead of an alarm flood.

Who we serve

Energy and sustainability managers

You need to know what to fix today and where to start. FDD points to the equipment and the probable cause, not just the symptom.

Facility engineers and operators

You run ongoing commissioning for many clients. FDD gives you one place to monitor every site and prove your service value.

Central plant and campus utility teams

You run chillers, boilers, and thermal storage for many buildings. FDD watches plant efficiency and flags drift before it shows up in kW per ton.

Challenges we address

FDD has a strong track record. Berkeley Lab found that organizations using FDD tools reached 9% median annual energy savings after two years of use. The gap is between finding faults and fixing them.

Hundreds of alarms a day, most of them noise

What’s really going on: Alarm limits were set at commissioning and never tuned. Real faults hide in the flood.

How Svitla solves it: The system groups related symptoms into one fault, suppresses duplicates, and ranks what’s left by cost and comfort impact.

A zone that’s always too warm

What’s really going on: The cause could sit two levels up, in the air handler or the plant. Nobody has time to trace it.

How Svitla solves it: Diagnostics follow the system hierarchy from zone to air handler to plant and name the most likely root cause.

Energy use that creeps up every year

What’s really going on: Overrides, leaking valves, and simultaneous heating and cooling build up quietly. None of them trip an alarm.

How Svitla solves it: Rules and learned baselines catch slow drift, and each fault shows its estimated waste in kWh and dollars.

Faults found but never fixed

What’s really going on: Findings live in a separate dashboard. They never become work orders, so they stay open.

How Svitla solves it: Each confirmed fault opens a work order in your CMMS. The system checks the data after repair and closes the loop.

Our process at Svitla

STEP 1

Connect and map

We start with the decisions your team makes every week and the questions behind them. Each one becomes a requirement.

STEP 2

Tune the rules to your buildings

We map source systems, data owners, and quality gaps. Where we can, we fix problems at the source.

STEP 3

Act and verify

We agree on definitions with operations and finance, then build each one once in the shared data model.

Schedule a call

Semantic point mapping

Maps vendor point names to Brick or Project Haystack tags.

Suggests tags automatically with AI and lets engineers confirm them.

Builds the equipment hierarchy from zone to plant.

Flags missing, frozen, or out-of-range sensors.

Fault rule library

Covers air handlers, rooftop units, VAV boxes, chillers, and boilers.

Includes economizer, valve leakage, and simultaneous heating and cooling rules.

Lets engineers edit thresholds and add custom rules.

Versions every rule change for audit.

AI anomaly detection

Learns normal behavior for each piece of equipment.

Adjusts for weather, occupancy, and schedule.

Flags slow drift and unusual patterns.

Explains which signals drove each anomaly.

Root cause diagnostics

Links symptoms across zone, air handler, and plant levels.

Ranks possible causes by likelihood.

Suppresses downstream faults caused by one upstream problem.

Shows the trend evidence behind each diagnosis.

Fault cost and priority scoring

Estimates energy waste per fault in kWh and dollars.

Weighs comfort, equipment wear, and occupant complaints.

Sorts the fix list by impact.

Tracks how long each fault has been open.

Work order loop and savings verification

Creates work orders in your CMMS with the diagnosis attached.

Confirms the fix from post-repair data.

Reopens faults that come back.

Reports verified savings by site and fault type.

Is custom fault detection and diagnostics software the best option for your needs?

Answer a few simple questions to find out whether you should choose a custom solution or a pre-built fault detection and diagnostics platform.

1 / 9

Do your operators ignore most BAS alarms because there are too many?

Do you run buildings with BAS systems from more than one vendor?

Are your point names inconsistent across buildings?

Have you tried an FDD tool that flagged faults nobody acted on?

Do you need faults to create work orders in your CMMS automatically?

Do you have custom equipment or sequences that standard rule libraries miss?

Would AI-based anomaly detection help you catch faults without rules?

Do you need to prove verified energy savings to leadership or clients?

Important integrations for fault detection and diagnostics software

BAS and controls networks

Trend and live data over BACnet, Modbus, or vendor APIs.

Data historians and IoT platforms

Long-term time-series data from historians or cloud data lakes.

Utility meters and submeters

Interval data to quantify energy waste per fault.

CMMS and work order systems

Forecasts and forecast accuracy by item and location.

Weather and occupancy feeds

Context for baselines, from weather services, access control, and space sensors.

Energy management dashboards

Verified savings and fault metrics shared with your reporting tools.

Lower energy use. Faults that waste energy get fixed sooner

Less simultaneous heating and cooling.

Fewer schedules left in override.

Economizers that actually use free cooling.

Faster repairs. Technicians arrive knowing what’s wrong

Less time spent on troubleshooting.

Fewer repeat visits.

Parts ordered before the truck rolls.

Longer equipment life. Equipment stops working against itself

Less short cycling and hunting.

Early warning before wear turns into failure.

Better input for replacement planning.

Fewer comfort complaints. Root causes get fixed, not just symptoms

Better occupant satisfaction.

Stable conditions in critical spaces.

Fewer hot and cold calls.

Proof of results. Every fix has measured savings attached

Verified savings for leadership reports.

Evidence for utility incentive programs.

Clear value for commissioning service clients.

Data discovery and semantic modeling

We inventory your BAS networks, point lists, and trend settings. We tag points to Brick or Project Haystack and build the equipment hierarchy. We fix trend intervals where data is too sparse to diagnose.
Rule tuning and model training

We apply a fault library to your equipment and run it against historical data. Engineers review every early finding to cut false positives. Anomaly models learn each unit’s normal behavior across seasons.
Workflow and integration build

We connect findings to your CMMS, define who gets which faults, and build the operator and manager views. We calibrate cost estimates with your meter data and utility rates.
Pilot, verify, and scale

We pilot on a few buildings, measure verified savings, and tune the rules. We then roll out to the rest of the portfolio with a repeatable onboarding process.

How much does fault detection and diagnostics software cost?

A focused FDD tool for one campus or a small client base typically costs $90,000 to $180,000. It covers one or two BAS protocols, a core rule library for air-side equipment, and CMMS ticketing. A multi-site platform with plant diagnostics, AI anomaly detection, savings verification, and multi-tenant access typically costs $200,000 to $450,000. The final cost depends on the number of protocols, equipment types, and sites.

Want to understand the cost of your fault detection and diagnostics software?

Calculate the cost

Check our client success cases

Risk Management System Modernization

Project Type Cloud Solutions, DevOps

Industry Energy, Financial Services

Discover how we modernized an ESG risk management platform by optimizing deployments, cutting disaster recovery time from an hour to 10 minutes, and enhancing observability. This led to better performance, scalability, and data protection for enterprises worldwide.

Solar Energy Billing Platform

Project Type Digital Transformation

Industry Energy

Discover how we transitioned a US solar energy supplier to a new business model focused on energy management, developing a comprehensive platform for payment estimation and billing, integrating seamlessly with external systems.

Fuel Management Analytics

Project Type Digital Transformation

Industry Energy, Logistics & Transportation

Uncover how we modernized a leading fuel management platform, transitioning from a monolithic to a microservices architecture, developing cross-platform mobile applications, and enhancing security features, significantly improving system reliability and user experience.

Consulting

We review your BAS data, alarm practices, and maintenance workflow, then define an FDD scope with a clear savings case.

Engineering

We build FDD software around your equipment, protocols, and work order process.

Modernization

We audit your current alarm setup or legacy analytics tool and plan the move. The new system adds root cause ranking the old one lacked. We migrate site by site and confirm fault accuracy after cutover.

FAQ

How is FDD different from the alarms in our BAS?

BAS alarms fire when one value crosses a limit. FDD looks at many points over time and across systems. It finds faults that never trip an alarm and explains the likely cause.

How do you keep false alarms low?

We run every rule against your historical data first. Engineers review early findings, and we tune thresholds to your equipment. A fault must also persist before it reaches your team.

Our point names are a mess. Is that a problem?

It’s the normal starting point. AI suggests tags for each point, and engineers confirm them. Once mapped, the same rules work across every building and vendor.

Can the software fix faults automatically?

It can for some faults, such as schedule overrides or setpoint drift, if you allow write access. Most clients start in read-only mode. Automatic correction is added rule by rule, with clear limits.

What happens when the cause is a bad sensor?

The system checks sensors against related readings and physical limits. The system flags a suspect sensor as its own fault. Rules that depend on it are paused, so one bad sensor doesn’t create a flood of false findings.

Can we offer FDD as a service to our own clients?

Yes. We can build multi-tenant access, client-branded reports, and per-site billing data. Each client sees only their own buildings.

Stop chasing symptoms across your buildings

Tell us about your buildings and controls. We’ll show you what a custom fault detection and diagnostics solution could find first.