A laser system

Abstract

A Laser System In certain applications it is desirable to monitor the output laser power of a laser system to assess system performance but also to provide an early indication of system degradation or failure. This is typically achieved using an internal detector configured to pick off a portion of the laser's output beam to measure laser power accurately. This naturally reduces the useable output power of the laser system. There is described a laser system that includes sensors which monitor the temperature and electrical usage characteristics, e.g. voltage across and / or current through, of components of the laser system, and a machine learning model trained to predict the laser's optical output power using only the temperature and electrical usage readings as inputs. The model can predict laser output power from thermal and electrical measurements alone thus obviating the need of a pick-off system.

Description

[0001] A Laser System

[0002] For certain applications it is desirable to monitor the output laser power of a laser system to assess system performance but also to provide an early indication of system degradation or failure.

[0003] The use of a machine learning (ML) algorithm to aid laser design is described in S. Gocheva-Ilieva, et al, "Stacking Machine Learning Models using Factor Analysis to Predict the Output Laser Power," 2022 International Conference on Electrical, Computer, Communications and Mechatronics Engineering (ICECCMEf Maldives, 2022, pp. 1-6. The ML model is designed to predict a laser output power based on design parameters of the laser, such as the inside diameter of the laser cavity tube, the gas pressure and pulse repetition frequency and supplied electric power with an assumed 50% efficiency drop.

[0004] Many known systems that monitor output laser power achieve this through inclusion of an internal detector configured to pick off a portion of the laser’s output beam to measure laser power accurately. There is a consequent reduction in the useable output power of the laser system, often by around 1%. In applications where the beam is required to travel long distances, such as for the purposes of optical communications, range finding, or for a Directed Infrared Counter Measure (DIRCM) system, this has particular disadvantage as any loss in power significantly reduces the maximum operating range. Examples of systems that use optical sensors to monitor the output laser power directly are described in W092 / 05608, US20050226288A1, US20040258108A1 and JP2001300750A

[0005] A method of determining the laser output power without the corresponding loss in utilisable laser power is therefore desirable.

[0006] According to a first aspect of the invention there is provided a laser system comprising: a laser; a set of sensors comprising: one or more current sensors and / or one or more voltage sensors to monitor voltage across and / or current draw of the laser and / or one or more electrical power drawing components of the laser, and / or one or more temperature sensors to sense the temperature of one or more components of the laser; and a processing system adapted to receive an output signal from each sensor of the set of sensors and, using the output signals, execute a predictive model to estimate an output laser power value of the laser.

[0007] In another aspect there is provided a laser system comprising: a laser; a set of sensors comprising: one or more sensors to monitor voltage across and / or current draw of the laser system or part thereof, and one or more temperature sensors to sense the internal temperature of one or more heat generating components of the laser system; and a processing system adapted to receive output signals from the set of sensors and configured to execute a predictive model trained to calculate and predict the output laser power of the laser system from inputs comprising the outputs of the set of sensors.

[0008] The following applies to either aspect of the invention.

[0009] The invention allows for the prediction of the laser output power without the need to sense light emitted from the laser. Thus avoiding the need for a pick off system, for accurately monitoring laser power and the corresponding reduction of usable optical power.

[0010] The processing system may be configured to output a predictive output laser power value ahead of a firing event of the laser, using the output signals received prior to firing event of the laser. Additionally or alternatively, the processing system may be configured to estimate an output laser power value of the output during firing, and / or provide the output laser power value of a previous firing event using the output signals received during and / or after firing of the laser.

[0011] In a system configured to operate at maximum power, this allows monitoring of the system performance over time.

[0012] In a system having a maximum laser output power that is greater than a required laser output power for an intended application, the system may comprise a feedback mechanism configured to alter parameters of system components of the laser system to increase the output laser power of the laser in response to a decrease in the estimated output laser power value output from the predictive model.

[0013] Where so, the feedback mechanism of the laser system may be configured to limit the laser output power when the laser system initially enters service to value below its maximum, and to adjust system components from time-to-time to compensate for natural degradation of laser system performance to maintain operation at the application requirement power level. As this arrangement allows the laser system to operate, at least initially, at below its maximum power, it may increase the usable service life of the laser system.

[0014] Another application of the invention is to predict laser power performance whilst the laser is in an operationally ready state without the need to fire the laser. This is beneficial where it is wanted to test the output laser power of the laser system in environments where it would not be safe or desirable to fire the laser, e.g. during initialisation of a laser, or in scenarios where premature firing could forfeit a tactical advantage.

[0015] The laser system may comprise means to output an alarm signal in response to the predicted output laser power value being below a threshold value. For example, the threshold value may be a value below an ideal output power level. The alarm signal may be output by either the controller 7 or processing system 8. The alarm signal could simply comprise a flag generated and stored in an operational log of the processing system 8 of the laser system 1. Alternatively, or additionally, the alarm signal could comprise a message or other alert that is transmitted to a user of the system, or a higher level computer control system.

[0016] The alarm may signal to the user (and / or a higher level computer control system) that the laser system requires attention, e.g. an inspection or servicing, or that its maximum performance characteristics are reduced.

[0017] Typically the predictive model comprises a machine learning model, for example implemented using an artificial neural network configured to receive the output signals as an input and output the estimate output laser power value as an output. Nevertheless, the predictive model could instead be an analytical representation of the “wall-plug efficiency” of the laser system, that describes the conversion of input electrical power into output optical power, by way of a transfer function. The advantage of using a trained ML model is the automatic adaptation of the model to any laser system, without prior knowledge of the transfer function.

[0018] The laser may be a solid state laser with a solid gain medium but it could equally be applied to other laser types.

[0019] The laser system may only one or more current sensors, or one or more voltage sensors, though typically it will comprise at least one if not more of both.

[0020] The one or more electrical power drawing components may include electrical devices and / or semiconductor components. Examples includes one or more of: Q switch voltage; pump diode voltage; thermal electrical cooler current; pump (e.g. pump diode) “fire mode ” current; pump (e.g. pump diode) “ready to fire ” current; controller and / or processing system voltage.

[0021] Preferably the system comprises multiple temperature sensors each configured to sense the temperature of a different component of the laser.

[0022] The one or more temperature sensors may be configured to sense the temperature of one or more of: optics temperature, heat exchanger temperature, pump diode temperature; resonator temperature; and optical parametric oscillator temperature.

[0023] According to another aspect of the invention there is provided a method of predicting the output power of a laser system comprising receiving output signals from each of a set of sensors that comprise: one or more current sensors and / or one or more voltage sensors to monitor voltage across and / or current draw of the laser system or one or more electrical power drawing components of the laser, and one or more temperature sensors to sense the temperature of one or more components of the laser system; and execute a predictive model configured to estimate the output laser power of the laser system from the output signals. According to a further aspect of the invention there is provided a method of training a machine learning model to predict the output power of a laser of a laser system from output signals from a set of sensors of the laser system comprising: one or more current sensors and / or one or more voltage sensors to monitor voltage across and / or current draw of the laser and / or one or more electrical power drawing components of the laser, and / or one or more temperature sensors to sense the temperature of one or more components of the laser; the method comprising training the machine learning model using output signals collected from multiple of the sets of sensors of multiple of the laser systems whilst the laser systems are operating during environmental testing during which the laser systems are subjected to temperature cycling and / or vibration stress testing.

[0024] The invention will now be described by way of example with reference to the Figure.

[0025] Figure 1 is a schematic of a laser system 1. The laser system may, for example, form part of a free space optical communication system, a range finder, or DIRCM system. In each case the optical output has a power sufficient to transmit a beam at least 1 km through the earth’s atmosphere, but typically much greater distances.

[0026] The laser system 1 comprises an optical resonator 2, a pump 3, e.g. a diode pump, to transfer energy into a gain medium (not shown) of the resonator 2, a thermal electric cooler 4 configured to regulate the temperature of the pump 3, a power supply 5, an optical parametric oscillator (OPO) 6, an optical output 15, a controller 7, and a processing system 8 comprising a processor 9 and a non-volatile computer readable store 10. The store 10 holds a neural network machine learning model (MLM) 11, and a laser output power value log 12.

[0027] The controller 7 and processing system 8 are realised by suitably programmed computer hardware.

[0028] The laser system may be a solid state laser system having a solid gain medium, which may, for example be in rod or fibre form. The laser system 1 will also typically also includes other optical components, e.g. one or more of a Q switch, telescope, mirrors, wavefront modifiers and lenses, to manipulate and / or direct the beam, e.g. between the resonator 2 and OPO 6, and between the OPO 6 and the optical output 15 of the laser system 1. Such components, being unconnected to the invention, are not shown in Fig 1.

[0029] Associated with each of the resonator 2, pump 3 and OPO 6 is a separate thermal sensor 13 adapted output a signal indicative of the temperature of its respective system component to the processing system 8.

[0030] The laser system 1 also comprises a plurality of sensors 14 to monitor voltage across and / or current draw of the laser system 1 and parts thereof, including: the voltage across the thermal cooler 4; current draw of thermal cooler 4; and voltage across and current draw of the pump 3. Each of the plurality of sensors 14 is configured to output a sensor signal indicative of the sensed voltage and / or current to the processing system 8.

[0031] Table 1, below, provides a non-exhaustive list of variables that may be sensed. The trained MLM 11 has a separate weight assigned to each variable indicating the variable’s influence on the predicted output optical output power at output 15.

[0032] Table 1

[0033] The controller 7 functions as a logic controller, controlling the operating function of the components and also as a power controller, regulating power to the other electrical components of the system 1.

[0034] Following initialisation of the laser system 1, the controller 7 operates the laser system 1 in a ‘standby’ mode in which the pump diodes 3 are maintained at an elevated temperature at or near a temperature required for laser firing, but with electrical current through the diode pump controlled so as to remain below that necessary to initiate lasing within the resonator 2.

[0035] In response to the controller 7 receiving an external firing signal, e.g. as a result of an input from a user or an external control system, the controller 7 switches from operation in “standby mode” to “firing mode” in which the pump diode 3 current is increased to initiate lasing within the resonator 2, and the firing of an optical signal out of optical output 15.

[0036] Operation of the laser system 1 in the “standby mode” reduces the delay between the receipt of the external firing signal by the controller 7 and firing of the laser.

[0037] Whilst in the standby and firing modes, output signals from each of the sensors 13 and 14 are transmitted to the processing system 8. The processing system 8 may carry out a filtering process on one or more of the received signals, if required, to render the output suitable for input to the MLM 11. This may be achieved in a variety of ways known to those skilled in the art of signal processing, including application of low and / or high pass filers and / or application of pattern recognition algorithms.

[0038] Additionally, the processing system 8 receives a status signal from the control system 7 indicating whether the laser system 1 is operating in standby mode or firing mode. The system status is used to label at least some of the sensor data, including the pump diode current.

[0039] The processing system 8 executes the MLM 11. The MLM 11 uses the output signals received from the sensors 13, 14, including the labelled diode current readings, as inputs, and from them calculates a predicted laser power value of a beam at the optical output 15.

[0040] Notably the inventors have determined that the MLM 11 attributes a much stronger weighting to the diode current when in “standby mode” compared with the diode current in “fire mode”. As such, it is possible to predict with reasonable accuracy, the optical output power of the laser system 1 without the need to fire the laser. Nevertheless, increased accuracy of the output may be achieved though using the historical values of diode current in fire mode from firing events of the laser system 1.

[0041] The laser power value output of the MLM 11 is recorded in log 12, which also stores the time the output was logged and, optionally, the status of the system (e.g. standby or fire) when the determination was taken.

[0042] The log may be reviewed from time-to-time, e.g. during a routine maintenance schedule, to ensure that the laser system 1 is operating within required performance parameters.

[0043] In one variant, the processing system 8 may hold an optical output power threshold value 16, and the processing system 8 is configured to compare the output value from the MLM 11 with the threshold value 16, and to generate an alert in response to the predicted laser power output value being below the threshold value 16. The alert may comprise a flag signal stored in the log 12 to indicate to a maintenance engineer that the system is operating below desired operating parameters. Additionally, or alternatively, the processing system 8 may send an alert signal to the controller 7 for implementing a feedback loop as described below.

[0044] The required output optical power for certain applications of the laser system 1 may be less than the maximum potential optical output power of the laser system 1. Where so, the laser system 1 may be configured to operate at an optical power level that is below its maximum. This may increase the service life of the laser system 1. Nevertheless, in such a system it may be still be important to ensure that the optical output power is still at or over the level required for the application. As such, the controller 7 may be configured to alter parameters of system components of the laser system 1 in response to receiving an alert signal from the processing system 8, to increase the actual or expected optical output power of the laser system 1. This may be achieved, for example, by one or more of increasing the diode current during standby and / or firing mode, controlling the temperature control systems to alter the temperature of one or more of the pump diodes, resonator 2 or OPO 6. After a period of time has elapsed to allow the changes to take effect, the MLM 11 can be re-executed using updated sensor readings from the sensors 13 14 to determine if the expected output power is at or above the threshold.

[0045] The MLM 11 is trained using sensor data and laser output power levels measured using an external power meter when the laser system is fired.

[0046] Training data is collected from multiple laser systems undergoing environmental testing prior to being put in service. During environmental testing each laser system is subjected to temperature cycling and vibration stress testing to check for any manufacturing defects. Advantageously training using data from multiple laser systems during environmental testing avoids the need to collect training data by running a single laser system in normal operational service for full operating life, which for certain lasers could exceed 15,000 hours.

[0047] Although the afore described system includes an OPO 6, this is an optional optical component.

Claims

Claims1. A laser system comprising: a laser; a set of sensors comprising: one or more current sensors and / or one or more voltage sensors to monitor voltage across and / or current draw of the laser and / or one or more electrical power drawing components of the laser, and / or one or more temperature sensors to sense the temperature of one or more components of the laser; and a processing system adapted to receive an output signal from each sensor of the set of sensors and, using the output signals, execute a predictive model to estimate an output laser power value of the laser.

2. A laser system according to claim 1, wherein the processing system is configured to execute the predictive model to output an estimate output laser power value for a firing event of the laser ahead of the firing event, using the output signals received prior to firing event.

3. A laser system according to claim 1 or 2, wherein the processing system is configured to output an estimate output laser power value of a previous firing event of the laser using the output signals received during and / or after said previous firing event.

4. A laser system according to any previous claim comprising a controller configured control one or more system components of the laser, and wherein the controller is configured to receive the estimate output laser power value from the predictive model, to compare the output laser power value with a pre-determined laserpower value and use the output of the comparison to adjust and / or set operating parameters of one or more system components to alter the output laser power of the laser towards the pre-determined laser power value.

5. A laser system according to any previous claim comprising means to output an alarm signal in response to the predicted output laser power value being below a threshold output laser power value.

6. A laser system according to any previous claim wherein the predictive model comprises a machine learning model trained to predict the output laser power value of the laser using the output signals as an input.

7. A laser system according to any previous claim wherein the one or more temperatures sensors include one or more of a: resonator temperature sensor to provide a temperature indicative of a resonator of the laser system; an optical parametric oscillator sensor to provide an indication of the temperature of an optical parametric oscillator of the laser system; and a pump sensor to provide a temperature indicative of a pump of the laser system configured to pump a gain medium of the laser system.

8. A laser system according to any previous claim wherein the set of sensors includes: a current sensor and / or a voltage sensor to monitor the voltage across and / or current draw through a thermal cooler for cooling a pump of the laser; and / or a current sensor to monitor the current draw of a pump of the laser system configured to pump a gain medium of the laser system.

9. A method of predicting the output power of a laser system comprising receiving output signals from each of a set of sensors that comprise: one or more current sensors and / or one or more voltage sensors to monitor voltage across and / or current draw of the laser system or one or more electrical power drawing components of the laser, andone or more temperature sensors to sense the temperature of one or more components of the laser system; and execute a predictive model configured to estimate the output laser power of the laser system from the output signals.

10. A method according to claim 9 wherein the predictive model comprises a machine learning model trained to estimate the output laser power of the laser system from the output signals.

11. A method of training a machine learning model to predict the output power of a laser of a laser system from output signals from a set of sensors of the laser system comprising: one or more current sensors and / or one or more voltage sensors to monitor voltage across and / or current draw of the laser and / or one or more electrical power drawing components of the laser, and / or one or more temperature sensors to sense the temperature of one or more components of the laser ; the method comprising training the machine learning model using output signals collected from multiple of the sets of sensors of multiple of the laser systems whilst the laser systems are operating during environmental testing during which the laser systems are subjected to temperature cycling and / or vibration stress testing.

Images