Automatic threshold selection of machine learning/deep learning model for anomaly detection of connected chillers
A chiller threshold management system for a building, including one or more memory devices and one or more processors. The one or more memory devices are configured to store instructions to be executed on the one or more processors. The one or more processors are configured to determine whether chiller fault data exists in chiller data used to generate a plurality of chiller prediction models. The one or more processors are further configured to generate a first threshold evaluation value for each of the plurality of chiller prediction models using a first evaluation technique in response to a determination that chiller fault data exists in the chiller data, and generate a second threshold evaluation value for each of the chiller prediction models using a second evaluation technique in response to a determination that chiller fault data does not exist in the chiller data. The one or more processors are configured to select a first threshold for each of the plurality of chiller prediction models based on the first threshold evaluation values in response to the determination that chiller fault data exists in the chiller data, and select a second threshold for each of the plurality of chiller prediction models based on the second threshold evaluation values in response to the determination that chiller fault data does not exist in the chiller data.
1. A chiller threshold management system for a building, the system comprising:
one or more memory devices configured to store instructions that, when executed on one or more processors, cause the one or more processors to:
determine whether chiller fault data exists in chiller data used to generate a plurality of chiller prediction models;
generate a first threshold evaluation value for each of the plurality of chiller prediction models using a first evaluation technique in response to a determination that the chiller fault data exists in the chiller data;
generate a second threshold evaluation value for each of the plurality of chiller prediction models using a second evaluation technique in response to a determination that the chiller fault data does not exist in the chiller data;
select a first threshold for each of the plurality of chiller prediction models based on the first threshold evaluation values in response to the determination that the chiller fault data exists in the chiller data;
select a second threshold for each of the plurality of chiller prediction models based on the second threshold evaluation values in response to the determination that the chiller fault data does not exist in the chiller data; and
perform at least one of operating one or more chillers based on at least one of the selected first or second thresholds or causing a service action to be performed for the one or more chillers based on at least one of the selected first or second thresholds.
2. The system of claim 1 , wherein the instructions cause the one or more processors to:
receive the chiller data for the one or more chillers, the chiller data indicating a performance of the one or more chillers;
generate the plurality of chiller prediction models based on the chiller data; and
calculate and select a threshold for each of the plurality of chiller prediction models.
3. The system of claim 1 , wherein the instructions cause the one or more processors to:
in response to the chiller data comprising the chiller fault data, implement the first evaluation technique to calculate receiver operating characteristics for the chiller data; and
select a first operating point on the receiver operating characteristics to determine the first threshold.
4. The system of claim 1 , wherein the instructions cause the one or more processors to:
in response to the chiller data being deficient of the chiller fault data, implement the second evaluation technique to estimate a distribution of a measure of a probability density function; and
select a second operating point on the distribution of the measure of the probability density function to determine the second threshold.
5. The system of claim 1 , wherein the instructions cause the one or more processors to receive the chiller data for a plurality of epochs, wherein the plurality of chiller prediction models are generated for each of the plurality of epochs.
6. The system of claim 1 , wherein the instructions cause the one or more processors to generate a fault prediction by applying at least one of the selected first or second thresholds to a chiller prediction model.
7. The system of claim 1 , wherein the instructions cause the one or more processors to update a fault prediction based on a selected chiller prediction model for each epoch of a plurality of epochs.
8. The system of claim 1 , wherein the instructions cause the one or more processors to detect a shutdown if a measure of a probability density function is less than a determined threshold.
9. A method for managing chiller thresholds for a building, the method comprising:
determining whether chiller fault data exists in chiller data used to generate a plurality of chiller prediction models;
generating a first threshold evaluation value for each of the plurality of chiller prediction models using a first evaluation technique in response to a determination that the chiller fault data exists in the chiller data;
generating a second threshold evaluation value for each of the plurality of chiller prediction models using a second evaluation technique in response to a determination that the chiller fault data does not exist in the chiller data;
selecting a first threshold for each of the plurality of chiller prediction models based on the first threshold evaluation values in response to the determination that the chiller fault data exists in the chiller data;
selecting a second threshold for each of the plurality of chiller prediction models based on the second threshold evaluation values in response to the determination that the chiller fault data does not exist in the chiller data; and
perform at least one of operating one or more chillers based on at least one of the selected first or second thresholds or causing a service action to be performed for the one or more chillers based on at least one of the selected first or second thresholds.
10. The method of claim 9 , further comprising:
receiving the chiller data for the one or more chillers, the chiller data indicating a performance of the one or more chillers;
generating the plurality of chiller prediction models based on the chiller data; and
calculating and selecting a threshold for each of the plurality of chiller prediction models.
11. The method of claim 9 , further comprising:
in response to the chiller data including the chiller fault data, implementing the first evaluation technique to calculate receiver operating characteristics for the chiller data; and
selecting a first operating point on the receiver operating characteristics to determine the first threshold.
12. The method of claim 9 , further comprising:
in response to the chiller data being deficient of the chiller fault data, implementing the second evaluation technique to estimate a distribution of a measure of a probability density function; and
selecting a second operating point on the distribution of the measure of the probability density function to determine the second threshold.
13. The method of claim 9 , further comprising receiving the chiller data for a plurality of epochs, wherein the plurality of chiller prediction models are generated for each of the plurality of epochs.
14. The method of claim 9 , further comprising generating a fault prediction applying at least one of the selected first or second thresholds to a chiller prediction model.
15. The method of claim 9 , further comprising updating a fault prediction based on a selected chiller prediction model for each epoch of a plurality of epochs.
16. The method of claim 9 , further comprising detecting a shutdown if a measure of a probability density function is less than a determined threshold.
17. A chiller threshold management system for a building, the system comprising:
one or more processors configured to:
determine whether chiller fault data exists in chiller data used to generate a plurality of chiller prediction models;
generate a first threshold evaluation value for each of the plurality of chiller prediction models using a first evaluation technique in response to a determination that the chiller fault data exists in the chiller data;
generate a second threshold evaluation value for each of the plurality of chiller prediction models using a second evaluation technique in response to a determination that the chiller fault data does not exist in the chiller data;
select a first threshold for each of the plurality of chiller prediction models based on the first threshold evaluation values in response to the determination that the chiller fault data exists in the chiller data;
select a second threshold for each of the plurality of chiller prediction models based on the second threshold evaluation values in response to the determination that the chiller fault data does not exist in the chiller data; and
perform at least one of operating one or more chillers based on at least one of the selected first or second thresholds or causing a service action to be performed for the one or more chillers based on at least one of the selected first or second thresholds.
18. The system of claim 17 , wherein the one or more processors is configured to:
receive the chiller data for the one or more chillers, the chiller data indicating a performance of the one or more chillers;
generate the plurality of chiller prediction models based on the chiller data; and
calculate and select a threshold for each of the plurality of chiller prediction models.
19. The system of claim 17 , wherein the one or more processors is are configured to:
in response to the chiller data including the chiller fault data, implement the first evaluation technique to calculate receiver operating characteristics for the chiller data; and
select a first operating point on the receiver operating characteristics to determine the first threshold.
20. The system of claim 17 , wherein the one or more processors are is configured to generate a fault prediction applying at least one of the selected first or second thresholds to a chiller prediction model.