System and method of controlling a temperature control system based on both system reliability and battery state
A method and system for controlling a temperature control system based on both system reliability and battery state is disclosed. Batteries may supply power to places where the power grid is unavailable. The batteries may be housed in containers, with the temperature and humidity of the containers being controlled using a temperature control system. The system and method disclosed may select the temperature setpoint and humidity setpoint in order to operate the batteries within prescribed temperature and humidity values, while considering both the state of health of the batteries (e.g., the battery state of charge) and the state of health of the temperature control system (e.g., the failure of the compressors). In this way, the batteries may be operated within prescribed limits while accounting for both the health of the batteries and the health of the temperature control system.
1 . A control system configured to control a temperature control system, the temperature control system configured to control one or both of temperature or humidity for batteries used in a container, wherein the batteries have at least one associated temperature range for operation or at least one associated humidity range for the operation, the control system comprising:
a communication interface; and
at least one controller in communication with the communication interface and configured to:
access a battery state indicator indicative of a state of at least one aspect of the batteries;
access a temperature control system reliability indicator indicative of reliability of at least one aspect of the temperature control system;
determine one or both of setpoint temperature that is selected to be within the at least one associated temperature range or setpoint humidity that is selected to be within the at least one associated humidity range based on a combined analysis of a predicted effect on the battery state indicator of controlling the container at the one or both of the setpoint temperature or the setpoint humidity and a predicted effect on the temperature control system reliability indicator in operating the temperature control system at the one or both of the setpoint temperature or the setpoint humidity; and
send one or more controls, indicative of the one or both the setpoint temperature or the setpoint humidity, via the communication interface to the temperature control system so that the temperature control system to control the container based on the one or both of the setpoint temperature or the setpoint humidity for the operation of the batteries in the container.
2 . The control system of claim 1 , wherein the at least one controller is configured to determine the one or both of the setpoint temperature or the setpoint humidity by:
inputting one or more inputs to one or more neural networks, the one or more neural networks configured to generate the predicted effect on the battery state indicator and the predicted effect on the temperature control system reliability indicator; and
selecting the one or both of the setpoint temperature or setpoint humidity using a container temperature or humidity optimizer.
3 . The control system of claim 2 , wherein the one or more neural networks comprise:
a first neural network configured to output the predicted effect on the temperature control system reliability indicator; and
a second neural network configured to output the predicted effect on the battery state indicator.
4 . The control system of claim 3 , wherein the temperature control system reliability indicator comprises a predicted indicator regarding one or more compressors in the temperature control system.
5 . The control system of claim 4 , wherein the predicted indicator regarding one or more compressors in the temperature control system comprises a predicted mean time between failures (MTBF) for the one or more compressors.
6 . The control system of claim 3 , wherein the predicted effect on the battery state indicator comprises an estimated effect on a battery state of charge (SoC).
7 . The control system of claim 3 , wherein the at least one controller is configured to select the one or both of the setpoint temperature or setpoint humidity using the container temperature or humidity optimizer by:
inputting the predicted effect on the temperature control system reliability indicator and the predicted effect on the battery state indicator to the container temperature or humidity optimizer in order for the container temperature or humidity optimizer to generate a plurality of solutions indicative of different setpoint temperatures or setpoint humidities for controlling the temperature control system; and
evaluating the plurality of solutions based on both the temperature control system reliability indicator and the predicted effect on the battery state indicator.
8 . The control system of claim 7 , wherein the at least one controller is configured to evaluate the plurality of solutions based on both the temperature control system reliability indicator and the predicted effect on the battery state indicator by evaluating whether a respective solution improves one or both of the temperature control system reliability indicator and the predicted effect on the battery state indicator.
9 . The control system of claim 8 , wherein the at least one controller is configured to evaluate whether the respective solution improves the temperature control system reliability indicator and whether the respective solution at least maintains the predicted effect on the battery state indicator.
10 . The control system of claim 9 , wherein the temperature control system reliability indicator comprises a predicted mean time between failures (MTBF) for one or more compressors in the temperature control system;
wherein the predicted effect on the battery state indicator comprises a battery state of charge (SoC); and
wherein the at least one controller is configured to evaluate whether the respective solution improves the MTBF and whether the respective solution at least maintains the battery SoC.
11 . The control system of claim 2 , wherein the one or more neural networks are configured to input an anomaly indicator indicative of one or more failures in the temperature control system; and
wherein the one or more neural networks are configured to use the anomaly indicator in order to generate the predicted effect on the battery state indicator and the predicted effect on the temperature control system reliability indicator.
12 . The control system of claim 11 , wherein the anomaly indicator is indicative of a number of anomalies in the temperature control system.
13 . The control system of claim 12 , further comprising an anomaly detection model configured to determine the number of anomalies in the temperature control system.
14 . The control system of claim 13 , wherein the anomaly detection model comprises Principal Component Analysis (PCA) algorithm configured to indicate faults in the temperature control system, a signal-to-noise ratio (SNR) process configured to determine a probability of a fault occurrence for the indicated faults, and a spectral residuals (SR) analysis configured to identify one or more most likely faulty components in the temperature control system based on the probability of the fault occurrence.
15 . The control system of claim 13 , wherein the anomaly detection model is configured to determine a number of a first type of anomalies in the temperature control system and a number of a second type of anomalies in the temperature control system, the first type of anomalies being indicative of a first type of fault, the second type of anomalies being indicative of a second type of fault, the first type of fault being different from the second type of fault; and
wherein the one or more neural networks are configured to input the number of the first type of anomalies in the temperature control system and the number of the second type of anomalies in the temperature control system; and
wherein the one or more neural networks are configured to use the number of the first type of anomalies and the number of the second type of anomalies in order to predict generate the predicted effect on the battery state indicator and the predicted effect on the temperature control system reliability indicator.
16 . The control system of claim 15 , wherein the one or more neural networks are further configured to:
determine a first predetermined number of anomalies;
determine a second predetermined number of anomalies;
compare the number of the first type of anomalies with the first predetermined number in order to determine whether the first type of fault has occurred or will occur;
compare the number of the second type of anomalies with the second predetermined number in order to determine whether the second type of fault has occurred or will occur; and
generate the predicted effect on the battery state indicator and the predicted effect on the temperature control system reliability indicator based on determining whether the first type of fault has occurred or will occur and based on determining whether the second type of fault has occurred or will occur.
17 . The control system of claim 11 , wherein the one or more neural networks are further configured to input temperature inside the container housing the batteries, temperature outside of the container, humidity inside the container housing the batteries, humidity outside of the container, battery state of charge (SoC) and compressor cycling.
18 . The control system of claim 17 , wherein the one or more neural networks comprise:
a first neural network configured to output the predicted effect on the temperature control system reliability indicator;
a second neural network configured to output the predicted effect on the battery state indicator;
wherein the first neural network is configured to input the battery SoC; and
wherein the second neural network is configured to input the compressor cycling.
19 . The control system of claim 1 , wherein the batteries are housed within the container in power banks;
wherein the container comprises a plurality of temperature sensors configured to generate sensor data;
wherein the at least one controller is configured to select a respective one of the plurality of temperature sensors based on analysis of reliability of the respective one of the plurality of temperature sensors and based on the respective one of the plurality of temperature sensors sensing a most problematic power bank of the batteries within the container.
20 . The control system of claim 1 , wherein the temperature control system comprises a plurality of compressors; and
wherein the at least one controller is further configured to determine differential loading of the plurality of compressors.
21 . The control system of claim 20 , wherein the at least one controller is configured to determine differential loading of the plurality of compressors responsive to determining sufficiency to perform the determination of the differential loading.
22 . The control system of claim 20 , wherein the at least one controller configured to determine differential loading of the plurality of compressors based on: a health indicator of one or both of the plurality of compressors; and at least one aspect of the one or both of the plurality of compressors.
23 . The control system of claim 22 , wherein the at least one aspect of the one or both of the plurality of compressors comprises one or more of compressor cycling, compressor operation history or maintenance, or previous compressor lead/lag operation.
24 . The control system of claim 22 , wherein the at least one controller is configured to dynamically determine the differential loading of the plurality of compressors.
25 . The control system of claim 24 , wherein the at least one controller is configured to dynamically determine the differential loading of the plurality of compressors based on input indicative of one or both of maintenance scheduling or budgetary information.
26 . A method for controlling a temperature control system, the temperature control system configured to control one or both of temperature or humidity for batteries used in a container, wherein the batteries have at least one associated temperature range for operation or at least one associated humidity range for the operation, the method comprising:
accessing, using at least one controller, a battery state indicator indicative of a state of at least one aspect of the batteries;
accessing, using the at least one controller, a temperature control system reliability indicator indicative of reliability of at least one aspect of the temperature control system;
determining, using the at least one controller, one or both of setpoint temperature that is selected to be within the at least one associated temperature range or setpoint humidity that is selected to be within the at least one associated humidity range based on a combined analysis of a predicted effect on the battery state indicator of controlling the container at the one or both of the setpoint temperature or the setpoint humidity and a predicted effet on the temperature control system reliability indicator in operating the temperature control system at the one or both of the setpoint temperature or the setpoint humidity; and
sending, using the at least one controller, one or more controls, indicative of the one or both the setpoint temperature or the setpoint humidity, via a communication interface to the temperature control system so that the temperature control system to control the container based on the one or both of the setpoint temperature or the setpoint humidity for the operation of the batteries in the container.
27 . The method of claim 26 , wherein determining, using the at least one controller, the one or both of the setpoint temperature or the setpoint humidity comprises:
inputting one or more inputs to one or more neural networks, the one or more neural networks configured to generate the predicted an effect on the battery state indicator and the predicted effect on the temperature control system reliability indicator; and selecting the one or both of the setpoint temperature or setpoint humidity using a container temperature or humidity optimizer.
28 . The method of claim 27 , wherein the one or more neural networks comprise:
a first neural network configured to output the predicted effect on the temperature control system reliability indicator; and
a second neural network configured to output the predicted effect on the battery state indicator.
29 . The method of claim 28 , wherein the temperature control system reliability indicator comprises a predicted indicator regarding one or more compressors in the temperature control system.
30 . The method of claim 29 , wherein the predicted indicator regarding one or more compressors in the temperature control system comprises a predicted mean time between failures (MTBF) for the one or more compressors.
31 . The method of claim 28 , wherein the predicted effect on the battery state indicator comprises an estimated effect on battery state of charge (SoC).
32 . The method of claim 28 , wherein selecting, using the at least one controller, the one or both of the setpoint temperature or setpoint humidity using the container temperature or humidity optimizer comprises:
inputting the predicted effect on the temperature control system reliability indicator and the predicted effect on the battery state indicator to the container temperature or humidity optimizer in order for the container temperature or humidity optimizer to generate a plurality of solutions indicative of different setpoint temperatures or setpoint humidities for controlling the temperature control system; and
evaluating the plurality of solutions based on both the temperature control system reliability indicator and the predicted effect on the battery state indicator.
33 . The method of claim 32 , wherein evaluating, using the at least one controller, the plurality of solutions based on both the temperature control system reliability indicator and the predicted effect on the battery state indicator comprises evaluating whether a respective solution improves one or both of the temperature control system reliability indicator and the predicted effect on the battery state indicator.
34 . The method of claim 33 , wherein the respective solution is evaluated, using the at least one controller, to determine whether there is improvement in the temperature control system reliability indicator and to determine whether the respective solution at least maintains the predicted effect on the battery state indicator.
35 . The method of claim 34 , wherein the temperature control system reliability indicator comprises a predicted mean time between failures (MTBF) for one or more compressors in the temperature control system;
wherein the predicted effect on the battery state indicator comprises a battery state of charge (SoC); and
wherein the respective solution is evaluated, using the at least one controller, to determine whether there is improvement of the MTBF and whether the respective solution at least maintains the battery SoC.
36 . The method of claim 27 , wherein the one or more neural networks are configured to input an anomaly indicator indicative of one or more failures in the temperature control system; and
wherein the one or more neural networks are configured to use the anomaly indicator in order to generate the predicted effect on the battery state indicator and the predicted effect on the temperature control system reliability indicator.
37 . The method of claim 36 , wherein the anomaly indicator is indicative of a number of anomalies in the temperature control system.
38 . The method of claim 37 , further comprising an anomaly detection model that determines the number of anomalies in the temperature control system.
39 . The method of claim 38 , wherein the anomaly detection model comprises Principal Component Analysis (PCA) algorithm that indicates faults in the temperature control system, a signal-to-noise ratio (SNR) process configured to determine a probability of a fault occurrence for the indicated faults, and a spectral residuals (SR) analysis that identifies one or more most likely faulty components in the temperature control system based on the probability of the fault occurrence.
40 . The method of claim 38 , wherein the anomaly detection model determines a number of a first type of anomalies in the temperature control system and a number of a second type of anomalies in the temperature control system, the first type of anomalies being indicative of a first type of fault, the second type of anomalies being indicative of a second type of fault, the first type of fault being different from the second type of fault; and
wherein the one or more neural networks inputs the number of the first type of anomalies in the temperature control system and the number of the second type of anomalies in the temperature control system; and
wherein the one or more neural networks uses the number of the first type of anomalies and the number of the second type of anomalies in order to predict generate the predicted effect on the battery state indicator and the predicted effect on the temperature control system reliability indicator.
41 . The method of claim 40 , wherein the one or more neural networks further:
determine a first predetermined number of anomalies;
determine a second predetermined number of anomalies;
compare the number of the first type of anomalies with the first predetermined number in order to determine whether the first type of fault has occurred or will occur;
compare the number of the second type of anomalies with the second predetermined number in order to determine whether the second type of fault has occurred or will occur; and
generate the predicted effect on the battery state indicator and the predicted effect on the temperature control system reliability indicator based on determining whether the first type of fault has occurred or will occur and based on determining whether the second type of fault has occurred or will occur.
42 . The method of claim 36 , wherein the one or more neural networks input temperature inside the container housing the batteries, temperature outside of the container, humidity inside the container housing the batteries, humidity outside of the container, battery state of charge (SoC) and compressor cycling.
43 . The method of claim 42 , wherein the one or more neural networks comprise:
a first neural network configured to output the predicted effect on the temperature control system reliability indicator;
a second neural network configured to output the predicted effect on the battery state indicator;
wherein the first neural network inputs the battery SoC; and
wherein the second neural network inputs the compressor cycling.
44 . The method of claim 26 , wherein the batteries are housed within the container in power banks;
wherein the container comprises a plurality of temperature sensors configured to generate sensor data;
wherein a respective one of the plurality of temperature sensors is selected, using the at least one controller, based on analysis of reliability of the respective one of the plurality of temperature sensors and based on the respective one of the plurality of temperature sensors sensing a most problematic power bank of the batteries within the container.
45 . The method of claim 26 , wherein the temperature control system comprises a plurality of compressors; and
further comprising determining, using the at least one controller, differential loading of the plurality of compressors.
46 . The method of claim 45 , wherein determining, using the at least one controller, differential loading of the plurality of compressors is responsive to determining sufficiency to perform the determination of the differential loading.
47 . The method of claim 45 , wherein determining, using the at least one controller, differential loading of the plurality of compressors is based on: a health indicator of one or both of the plurality of compressors; and at least one aspect of the one or both of the plurality of compressors.
48 . The method of claim 47 , wherein the at least one aspect of the one or both of the plurality of compressors comprises one or more of compressor cycling, compressor operation history or maintenance, or previous compressor lead/lag operation.
49 . The method of claim 47 , wherein the differential loading of the plurality of compressors is determined, using the at least one controller, dynamically.
50 . The method of claim 49 , dynamically determining, using the at least one controller, the differential loading of the plurality of compressors is based on input indicative of one or both of maintenance scheduling or budgetary information.