- Actual
An actual value of production or actual loss represents a quantity determined from data generated by sensors in the plant. An actual value indicates real conditions in the plant.
Compare actual and expected values to help identify potential problems.
- Aggregated
A value representing a condition of the entire plant that was determined by grouping together values measured for individual components.
- Array
The photovoltaic modules in the plant. Modules in the array are connected electrically in series to form a string. Strings are connected electrically in parallel to form the array.
Modules in the array are arranged physically into rows.
For some arrays, Solar AI Analytics groups rows of modules into sections of rows. A section is a group of modules that share characteristics that affect soiling such as tilt and azimuth angle, proximity to a source of soiling such as an agricultural field, or proximity to a source of shading such as a hill.
- Available Data Record Length
The number of days covered by the available data is shown on the Data and Sensors page.
- Capacity
The size of the plant, array, single inverter, or single module under nominal operating conditions. Capacity is a measurement of electrical power that is often expressed in Watts (W), kilowatts (kW), or megawatts (MW), where 1 MW = 1,000 kW = 1,000,000 W.
Plant and inverter capacity is typically an AC power rating. Module and array capacity is typically a DC power rating.
- Clipping
See Inverter Clipping.
- Combiner
A junction box that connects DC conductors in the photovoltaic input circuit. The photovoltaic input circuit includes all conductors between photovoltaic modules, and between the modules and the connection point of the DC system.
- Commissioning
See Software Commissioning.
- Corrective Action
A maintenance, repair, or management task that can be performed to recover lost production. Examples of corrective actions include:
- Washing modules to reduce soiling loss.
- Placing a claim with the energy off-taker to stop or reduce grid curtailment loss.
- Inspecting and repairing combiner connections to reduce underperforming strings loss.
For each corrective action, Solar AI Analytics calculates the weekly impact on revenue, the cost of taking the corrective action, and the payback period in weeks required to recover the cost of the corrective action. Solar AI Analytics considers a corrective action to be cost-effective if the payback period is less than the appropriate threshold on the Settings page.
Each corrective action is associated with a loss in production. That production is recoverable production if the corrective action that eliminates the cause of the loss is cost-effective.
Each corrective action category is associated with a loss category. For example, the “backtracking schedule” corrective action category describes actions that can be taken to reduce losses in the “shade” category.
The weekly impact of a corrective action is the weekly reduction in revenue caused by the loss in production associated with the corrective action. It is calculated from the lost production in kWh and the power price, accounting for any time-of-delivery pricing adjustments.
The cost of a corrective action is the total cost of taking the action.
The payback period is the number of weeks it would take for the corrective action to pay for itself in revenue. It is the action cost divided by the weekly impact.
- Data Record Length
See Available Data Record Length.
- DC Health KPI
An indicator of the performance of DC strings, but at the inverter level. DC health is calculated in the same way as underperforming strings, but determined for each inverter rather than for each group of connected strings.
The DC health of the best-performing inverter is defined as 100%, so the DC health of the remaining inverters is in relation to the best-performing inverter.
The best-performing inverter is defined as the inverter whose production is in the 90th percentile of all inverter’s production during a period when there are no shadows on the array, clipping, curtailment, or other factors negatively affecting inverter performance.
- DC/AC Ratio
The ratio of nominal photovoltaic array DC capacity to nominal plant AC capacity.
- Design Max
See Production
- Down String
A down string is a string of modules that is inoperative. Down strings can be indicated by a number of down strings, or by a loss in kilowatt-hours.
- Downtime
Downtime is a period of time when the plant or one or more inverters are not production power.
- Expected (Losses)
Values specified by the user on the Settings page as reference values to indicate the expected losses according to a preconstruction energy model (e.g. PVsyst). These are purely referencial and not used in any internal calculations, and should be adjusted to best suit the expectation of the asset manager. Compare expected and actual values to help identify potential problems.
- Expected (Power)
Baseline Expected This model is a simple regression model where irradiance is temperature adjusted according to the PV module technology and then scaled to each inverter’s expected power during normal unconstrained operating periods during clear sky intervals using data from the most recent days. The majority of Solar AI Analytics losses are calculated from this reference.
Machine Learning Expected This model is based on a Random Forest machine learning model that is continuously trained on historical AC power readings per inverter during unconstrained periods of operation.
Physical Expected This model is based on a simple physical model similar to NREL’s PVWatts.
- Feeder
The length of electrical conductor that carries power from the combiner or recombiner to the inverter DC input.
- Flag
A number used to identify potential problems in a data set. A flag is usually a maximum or minimum acceptable value of a number.
- Generation
See Lifetime Generation.
- Intra-Array Shading
Shadows on modules in the array caused by other modules.
- Inverter
A component of the plant that converts DC electricity from the array to AC electricity. An Inverter’s control system allows it to adjust its output power, restart, and shut down in response to signals from sensors in the plant or from an operator.
- Inverter Clipping
Reduction in an inverter’s production caused by the input voltage exceeding the rated range, or by the input power exceeding the rated maximum. Inverters can reduce input power by increasing the DC input voltage, which causes the modules connected to the inverter to operate at less than their maximum power point.
- Inverter Efficiency
Inverter DC to AC power conversion efficiency. Inverter efficiency significantly lower than the plant average inverter efficiency may indicate that an inverter requires maintenance or repair.
- Irradiance
The solar energy incident on the surface of a module in watts per square meter (W/^2^).
- Lifetime Generation
The total generation of the entire fleet in gigawatt-hours over the entire period covered by the data. This value is not affected by the time range you select.
- Loss
The difference between the expected and actual power of the plant or a component of the plant. Solar AI Analytics reports different loss categories to help identify potential problems in the plant.
All of the loss categories are shown on the Loss Breakdown page. More detailed analysis of individual loss categories, such as soiling losses, shading losses, downtime, etc. is shown on the other input pages.
See the table in the Loss Breakdown chapter for a definition of each loss category.
- Loss Category
A loss from a single cause. For example, soiling loss and inverter clipping loss are two different loss categories.
- Module
An electronic device that converts sunlight into DC electricity. A typical module is large, flat, and rectangular and consists of photovoltaic cells, front and back covers, and electrical conductors and diodes that connect the cells. Modules are connected together to form the array.
- Obstruction
An object such as a tree or building that casts a shadow on modules in the array.
- Outage
An event that causes a string, inverter, or the entire plant to stop generating power.
- Partial capacity loss
Energy loss due to an inverter not operating at its rated nameplate capacity. Some inverters are designed with multiple power conversion units that may fail individually. As an example, a 2 MW inverter with 4 internal units, at the failure of one of them, will exhibit clipping behavior at 1.5 MW, which is different from regular clipping because it is due to the equipment operational state that may be improved upon, rather than due to the design of the system DC/AC ratio.
- Performance Ratio
The average weather-corrected performance ratio of all plants in the fleet over the selected time range.
The performance ratio is the ratio of the energy generated by the plant to the energy that would have been generated if the plant consistently converted sunlight to electric energy at the level that would be expected from the DC nominal rating (NREL, April 2013). The weather-corrected performance ratio, indicated by “wc” accounts for seasonal differences in irradiance and temperature, so that performance ratio values can be compared across different seasons. The graph shows the expected and actual performance ratio over the available data record length.
The IEC standard for PV Plant Performance (61724-1, 2017) offers additional guidelines on data selection and operating conditions.
- Plant
The solar power plant is all of the equipment in a power generating asset and includes solar photovoltaic modules, trackers, inverters, combiners, wiring and other hardware.
- Portfolio
A portfolio is a collection of power generating assets.
- Portfolio Size
The sum of the capacities of all plants in the portfolio.
- Power
For the purposes of this documentation, power is electricity generated by the plant.
- PR
Abbreviation of Performance Ratio.
- Production
The electrical energy output of the plant measured at the inverter output.
Metered Production is the energy counter of the power plant, that continues to log accumulated energy delivered even during periods of data communication outages. The metered value is the one typically used for financial asset management and energy valuation, but is not always onboarded into Solar AI Analytics depending on the deployment scenario.
Analyzed production is the sum of the production at the inverter output for all inverters in the plant over a given time range. Actual production excludes intervals with bad data, and therefore may differ from data measured at the grid meter or reported by real-time monitoring and control applications that includes data from all periods when inverter power readings are available. The analyzed production may be greater than or less than the metered production, due to the degree to which time series data quality and sensor accuracy impact the available interval data. In theory, it may be relevant to note that in theory, the sum of inverter power and delivered energy shall always be greater than metered production, due to downstream AC wiring losses between the inverters and metered point of interconnection.
Achievable production is the production as calculated by a model of the plant. The model includes the effects for weather conditions, soiling, shading, inverter clipping. It does not include curtailment, downtime, or other losses that can be reduced by taking a corrective action.
Design Max is the theoretical maximum production of the plant, assuming that all corrective actions are taken, regardless of whether they are cost-effective. It is the sum of the following loss categories:
- Tracker stow (wind) recoverable energy
- Tracker stow (other) actual loss
- Snow actual loss
- Soiling actual loss
- Shading recoverable energy
- Down strings actual loss
- Underperforming strings actual loss
- Inverter efficiency recoverable energy
- VAr support actual loss
- Inverter downtime (other) actual loss
- Curtailment actual loss
- Design maximum does not include the following loss categories because they are normal losses that are determined by the system’s design rather than by its operation and maintenance:
- Module thermal loss
- Inverter clipping
- Night loss
- Nominal plant capacity
Design Max represents the O&M provider practicing perfect maintenance, in the sense of all corrective action taken immediately regardless of cost effectiveness.
Recoverable production is production that could be added to the actual production by taking all cost-effective corrective actions.
Not economically recoverable is production that could be added to the actual production by taking corrective actions that are not cost-effective.
- Recombiner {#recombiner}
A junction box that connects DC wiring in the photovoltaic output circuit. The photovoltaic output circuit includes all conductors in the DC system between the combiner and the inverter.
- Record Length
See Available Data Record Length
- Recoverable Energy
The recoverable energy is an estimate of energy available at the inverter output, assuming that all corrective actions that are determined to be economically recoverable are taken.
Recoverable energy is calculated by a model of the plant that accounts for weather conditions at the plant site, and uses daily estimates of the effects of soiling, shading, string operational state, and inverter clipping. Recoverable energy does include loss categories that have no corrective actions, such as curtailment and downtime.
- SCADA
The system that interacts with the plant’s data collection and control equipment and records data from sensors.
- Section
A group of modules in the array with the same tilt angle. If all modules in the array have the same tilt angle, the array has one section.
- Delivery Engineer
The Univers employee responsible for setting up Solar AI Analytics software at the customer’s site.
- Software Commissioning
The process of setting up Solar AI Analytics software performed by an Univers delivery engineer at the customer’s site.
- String
Modules electrically connected in series. The number of modules in a string determines its nominal voltage.
- Time Range
The number of days represented by a value.
See also Selecting Time Ranges.
- Total Losses (%)
Total losses as a percentage of recoverable energy over the available data record length. It accounts for all loss categories considered by Solar AI Analytics and includes both recoverable and non-recoverable losses. The sum of total losses and analyzed production does not equal the recoverable energy because recoverable energy only accounts for recoverable losses, such as downtime and curtailment. Total losses also includes non-recoverable losses such as clipping and thermal losses.
- Total Losses ($)
The total lost revenue is the economic value of the lost energy based on the power price for the plant. This is typically higher than the recoverable amount because it includes the value of unrecoverable losses, such as shading.
- Tracker
A piece of equipment that moves modules in the array to follow the movement of the sun. Trackers move into a stow (typically horizontal) position during high wind events to reduce mechanical load on the tracker hardware and avoid damage, which causes a reduction in plant power.
- Underperforming String
A string of modules that is producing less power than its nearby neighbors in the array, based on a statistical analysis of data measured at the DC combiner, recombiner, or DC feeder.
The underperforming string indicator is similar to DC Health KPI, but determined from data measured at the string level rather than the inverter level.
- VAr Support
Power required for reactive power control to grid system operator voltage requirements.
- wc
Abbreviation of “weather-corrected.”
- Zone
Parts of the array that are likely to experience similar soiling because of their orientation, proximity to roads and agricultural fields, and other environmental factors.
Concepts that could be overlapped with other usages in the entire Bazefield platform.
- Active Alarm
See Alarm
- Alarm
A warning message indicating the possible need for a corrective action.
An active alarm is an alarm that has not been addressed by a corrective action.
- Availability
A measure of the reliability of the plant or plant components. A component with 100% availability over a given time range is 100% reliable, meaning that it never failed during that time. A component with an availability of 95% over a time range did not operate for 5% of that time.
Solar AI Analytics measures availability at the output of each inverter, and aggregates those values to calculate the plant availability. It calculates availability over a given time range by dividing downtime in hours in the time range by the total number of hours in the range when the plane-of-array solar irradiance was greater than 50 W/m^2^
Operational time-based availability (TBAop) includes all downtime events caused by equipment failures and external events such as grid outages, snow-covering modules, and planned maintenance outages.
Technical time-based availability (TBAtech) only includes downtime events caused by equipment failures. It does not include downtime events caused by grid outages, inverter failure outside of normal operation and maintenance, or other unplanned causes. Technical availability may be comparable to contractual availability, although the contract terms that define contractual availability may differ from the Solar AI Analytics definition.
- Time-Based Availability
The average time-based availability of all plants in the fleet over the selected time range.
See Availability.
- Unavailability
The percentage of time data was not available over the data record length.