U.S. patent application number 15/747115 was filed with the patent office on 2018-09-20 for isolation monitoring device and method.
The applicant listed for this patent is Sendyne Corporation. Invention is credited to Nicolas CLAUVELIN, Victor MARTEN, Ioannis MILIOS.
Application Number | 20180267089 15/747115 |
Document ID | / |
Family ID | 62624735 |
Filed Date | 2018-09-20 |
United States Patent
Application |
20180267089 |
Kind Code |
A1 |
CLAUVELIN; Nicolas ; et
al. |
September 20, 2018 |
Isolation monitoring device and method
Abstract
Wide deployment of high voltage battery systems in traction,
industrial and renewable energy installations is raising the
concerns for human safety. Exposure to hazardous high voltages may
occur due to deterioration of insulation materials or by accidental
events. It is thus important to monitor for such faults and being
able to provide timely warnings to affected persons. For this
purpose it has become mandatory for electrified passenger vehicles
(CFR 571.305) to maintain high isolation values which can be
continuously monitored by electrical isolation monitoring devices.
The task of monitoring isolation resistance within the electrically
noisy car environment is not a trivial task and the solution to
this problem has become quickly a field of research and innovation
for all affected industries.
Inventors: |
CLAUVELIN; Nicolas; (New
York, NY) ; MARTEN; Victor; (Flushing, NY) ;
MILIOS; Ioannis; (New York, NY) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Sendyne Corporation |
New York |
NY |
US |
|
|
Family ID: |
62624735 |
Appl. No.: |
15/747115 |
Filed: |
October 31, 2017 |
PCT Filed: |
October 31, 2017 |
PCT NO: |
PCT/IB2017/056746 |
371 Date: |
January 24, 2018 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
62436350 |
Dec 19, 2016 |
|
|
|
62436358 |
Dec 19, 2016 |
|
|
|
62452966 |
Jan 31, 2017 |
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G01R 27/025 20130101;
G01R 31/50 20200101; B60L 58/10 20190201; B60L 3/0069 20130101;
G01R 27/18 20130101; B60L 3/0046 20130101; Y02T 10/70 20130101;
G01R 31/52 20200101; G01R 31/007 20130101 |
International
Class: |
G01R 31/00 20060101
G01R031/00; B60L 3/00 20060101 B60L003/00 |
Claims
1. A method to estimate a change in values of isolation impedance
in an isolated ground (IT) electrical system comprising a power
source, the method comprising: modeling a first isolation path
between a first reference point and a second reference point and
modeling a second isolation path between a third reference point
and a fourth reference point, thereby creating a theoretical model
of the isolated ground electrical system; providing an initial
value of a first isolation resistance for the first isolation path
and an initial value of a second isolation resistance for the
second isolation path; measuring an initial value of a voltage
between the first reference point and the second reference point
and storing the measured initial value in a storage medium;
measuring an initial value of a voltage between the third reference
point and the fourth reference point and storing the measured
initial value in the storage medium; measuring a subsequent
different value of the voltage between the first reference point
and the second reference point and storing the measured subsequent
value in the storage medium; measuring a subsequent value of the
voltage between the third reference point and the fourth reference
point and storing the measured subsequent value in the storage
medium; entering the measured initial values of the voltages, the
measured subsequent values of the voltages, the provided values of
the isolation impedances and an elapsed amount of time between the
initial measurements and the subsequent measurements into a
mathematical function stored in the storage medium; wherein the
mathematical function minimizes the discrepancy between the
measured change in values of the voltages and the modeled
theoretical values by adjusting values of modeled isolation
impedances associated with the isolation paths in the electrical
system; extracting estimated values of isolation impedances
associated with the isolation paths in the electrical system by
application of the mathematical function; and storing the estimated
values in the storage medium.
2. The method of claim 1, wherein the second reference point is a
chassis ground of the electrical system.
3. The method of claim 2, wherein the fourth reference point is the
chassis ground of the electrical system.
4. The method of claim 1, further comprising: comparing the
estimated values of isolation resistance with a range of acceptable
values and communicating that the estimated value of resistance for
an isolation path is outside the range of acceptable values.
5. The method of claim 1, further comprising: communicating an
amount of estimated energy stored in the isolation impedances.
6. The method of claim 1, wherein the power source is a battery and
wherein the first reference point and the third reference point are
positive and negative terminals of the battery.
7. The method of claim 1, wherein the power source is a
supercapacitor.
8. The method of claim 1, wherein the power source is a DC
charger.
9. The method of claim 1, further comprising: identifying a minimum
resistance path from the estimated values of isolation
resistance.
10. The method of claim 9, further comprising: communicating a
value of resistance for the minimum resistance path in the
electrical system.
11. The method of claim 9, further comprising: associating the
minimum resistance path with one of the power source terminals.
12. The method of claim 1, wherein the theoretical model of the
electrical system is an equivalent circuit model.
13. The method of claim 1, further comprising: extracting an
estimated value of at least a first capacitance associated with an
isolation path by application of the mathematical function and
storing the estimated value in the storage medium.
14. The method of claim 13, further comprising: comparing the
estimated values of capacitance with a range of acceptable values
and communicating that the estimated value of the at least first
capacitance is outside a range of acceptable values.
15. The method of claim 13, wherein the power source is a battery
and the at least first reference point in the electrical system is
a terminal of the battery.
16. The method of claim 15, wherein the measured voltage values are
measurements of a varying voltage within the electrical system
while the electrical system is operating and measurements of a
voltage signal source while the electrical system is idle.
17. The method of claim 16, wherein the mathematical function
stored in the storage medium is a least square estimator which
produces a least squared error estimate.
18. The method of claim 17, wherein the least squared error
estimate is performed over a predetermined number of voltage
measurements and corresponding voltage predictions, thereby
minimizing a deviation between the measured voltage values and the
estimated voltage values, and thereby producing a corresponding
number of present value estimates and associated uncertainties for
the present value estimates.
19. The method of claim 18, wherein the method steps are performed
iteratively.
20. The method of claim 19, wherein the present value estimates are
expressed as a vector and the associated uncertainties are
expressed as a covariance matrix for the vector.
21. The method of claim 20, further comprising a stochastic filter,
wherein the extracted estimated values are fed to the filter and
the filter maintains the most likely present value estimates and
associated uncertainties for the present value estimates.
22. The method of claim 21, wherein the stochastic filter is a
Kalman filter.
23. The method of claim 22, further comprising: receiving as inputs
to the Kalman filter a set of previous present value estimates and
associated uncertainties; receiving as inputs to the Kalman filter
a set of estimated values, including the estimated value of the
resistance change and the estimated value of the capacitance
change; outputting a new set of values for the most likely present
value estimates and associated uncertainties by application of the
filter; and updating the present value estimates and associated
uncertainties stored in the storage medium.
24. The method of claim 23, wherein the method steps are performed
iteratively.
25. An apparatus for estimating unknown values of isolation
impedance in an isolated ground (IT) electrical power system,
comprising: a power source having a positive terminal and a
negative terminal, said terminals connected in circuit to at least
one additional electrical component and isolated from a chassis
ground within the electrical system; wherein the electrical system
contains an isolation impedance between each of the terminals and
the chassis ground; a storage medium; means measuring an initial
value and a subsequent different value of a voltage between the
chassis ground and a first reference point and between a second
reference point and a third reference point in the electrical
system; means storing the measured initial values and the
subsequent different values in the storage medium; a mathematical
function stored in the storage medium, whereby application of the
mathematical function extracts estimated values of isolation
impedances associated with the voltage measurements by using a
model of the electrical system and minimizing an error
function.
26. The apparatus of claim 25, further comprising: wherein the
electrical system contains at least one capacitance between each of
the terminals and the chassis ground, and wherein the mathematical
function extracts an estimated value of the capacitance associated
with the voltage measurement.
27. The apparatus of claim 25, wherein the power source is a
battery.
28. The apparatus of claim 25, wherein the power source is a power
conversion system.
29. An apparatus for estimating a change in values of isolation
impedance in an isolated ground (IT) electrical power system,
comprising: a power source having a positive terminal and a
negative terminal, said terminals connected in circuit to at least
one additional electrical component and isolated from a chassis
ground within the electrical system; wherein the electrical system
contains an isolation impedance between each of the terminals and
the chassis ground; a storage medium; means measuring an initial
value and a subsequent different value of a voltage between the
chassis ground and a first reference point and between a second
reference point and a third reference point in the electrical
system; means storing the measured initial values and the
subsequent different values in the storage medium; a mathematical
function stored in the storage medium, whereby application of the
mathematical function extracts an estimated change in values of
isolation impedances associated with the voltage measurements by
using a model of the electrical system and minimizing an error
function.
30. A method to estimate a change in values of isolation impedance
in an isolated ground (IT) electrical system comprising a power
source and a load, the method comprising: modeling a first
isolation path between a first reference point and a second
reference point and modeling a second isolation path between a
third reference point and a fourth reference point, thereby
creating a theoretical model of the isolated ground electrical
system; at a time when power from the power source is being
dissipated in the load, measuring an initial value of a voltage
between the first reference point and the second reference point
and storing the measured initial value in a storage medium; at a
time when power from the power source is being dissipated in the
load, measuring an initial value of a voltage between the third
reference point and the fourth reference point and storing the
measured initial value in the storage medium; at a time when power
from the power source is being dissipated in the load, measuring a
subsequent different value of the voltage between the first
reference point and the second reference point and storing the
measured subsequent value in the storage medium; at a time when
power from the power source is being dissipated in the load,
measuring a subsequent value of the voltage between the third
reference point and the fourth reference point and storing the
measured subsequent value in the storage medium; entering the
measured initial values of the voltages, the measured subsequent
values of the voltages and an elapsed amount of time between the
initial measurements and the subsequent measurements into a
mathematical function stored in the storage medium; wherein the
mathematical function minimizes the discrepancy between the
measured initial values of the voltages, the measured subsequent
values of the voltages and the modeled theoretical values by
adjusting values of modeled isolation impedances associated with
the isolation paths in the electrical system; extracting estimated
values of isolation impedances associated with the isolation paths
in the electrical system by application of the mathematical
function; and storing the estimated values in the storage
medium.
31. The method of claim 30, wherein the second reference point is a
chassis ground of the electrical system.
32. The method of claim 31, wherein the fourth reference point is
the chassis ground of the electrical system.
33. The method of claim 30, further comprising: comparing the
estimated values of isolation resistance with a range of acceptable
values and communicating that the estimated value of resistance for
an isolation path is outside the range of acceptable values.
34. The method of claim 30, further comprising: communicating an
amount of estimated energy stored in the isolation impedances.
35. The method of claim 30, wherein the power source is a battery
and wherein the first reference point and the third reference point
are positive and negative terminals of the battery.
36. The method of claim 30, wherein the power source is a
supercapacitor.
37. The method of claim 30, wherein the power source is a DC
charger.
38. The method of claim 37, further comprising: identifying a
minimum resistance path from the estimated values of isolation
resistance.
39. The method of claim 38, further comprising: communicating a
value of resistance for the minimum resistance path in the
electrical system.
40. The method of claim 38, further comprising: associating the
minimum resistance path with one of the power source terminals.
41. The method of claim 30, wherein the theoretical model of the
electrical system is an equivalent circuit model.
42. The method of claim 30, further comprising: extracting an
estimated value of at least a first capacitance associated with an
isolation path by application of the mathematical function and
storing the estimated value in the storage medium.
43. The method of claim 42, further comprising: comparing the
estimated values of capacitance with a range of acceptable values
and communicating that the estimated value of the at least first
capacitance is outside a range of acceptable values.
44. The method of claim 42, wherein the power source is a battery
and the at least first reference point in the electrical system is
a terminal of the battery.
45. The method of claim 44, wherein the measured voltage values are
measurements of a varying voltage within the electrical system
while the electrical system is operating and measurements of a
voltage signal source while the electrical system is idle.
46. The method of claim 45, wherein the mathematical function
stored in the storage medium is a least square estimator which
produces a least squared error estimate.
47. The method of claim 46, wherein the least squared error
estimate is performed over a predetermined number of voltage
measurements and corresponding voltage predictions, thereby
minimizing a deviation between the measured voltage values and the
estimated voltage values, and thereby producing a corresponding
number of present value estimates and associated uncertainties for
the present value estimates.
48. The method of claim 47, wherein the method steps are performed
iteratively.
49. The method of claim 48, wherein the present value estimates are
expressed as a vector and the associated uncertainties are
expressed as a covariance matrix for the vector.
50. The method of claim 49, further comprising a stochastic filter,
wherein the extracted estimated values are fed to the filter and
the filter maintains the most likely present value estimates and
associated uncertainties for the present value estimates.
51. The method of claim 50, wherein the stochastic filter is a
Kalman filter.
52. The method of claim 51, further comprising: receiving as inputs
to the Kalman filter a set of previous present value estimates and
associated uncertainties; receiving as inputs to the Kalman filter
a set of estimated values, including the estimated value of the
resistance change and the estimated value of the capacitance
change; outputting a new set of values for the most likely present
value estimates and associated uncertainties by application of the
filter; and updating the present value estimates and associated
uncertainties stored in the storage medium.
53. The method of claim 52, wherein the method steps are performed
iteratively.
Description
BACKGROUND OF THE INVENTION
[0001] Wide deployment of high voltage battery systems in traction,
industrial and renewable energy installations is raising the
concerns for human safety. Exposure to hazardous high voltages may
occur due to deterioration of insulation materials or by accidental
events. It is thus important to monitor for such faults and being
able to provide timely warnings to affected persons. For this
purpose it has become mandatory for electrified passenger vehicles
(CFR 571.305) to maintain high isolation values which can be
continuously monitored by electrical isolation monitoring devices.
The task of monitoring isolation resistance within the electrically
noisy car environment is not a trivial task and the solution to
this problem has become quickly a field of research and innovation
for all affected industries.
[0002] The function of the isolation monitoring device is to
determine the value of the isolation resistance between either of
the battery terminals and the chassis. Furthermore it must issue an
alarm if the isolation resistance becomes lower than a certain
value. This value is determined by the human body tolerance to
electrical current. The table of FIG. 1 shows typical human body
reaction to current passing through the body. The resistance values
shown for the different paths in the human body are typical and can
vary widely based on the condition of the skin and other factors.
Nevertheless, one may notice that the order of magnitude of
resistances is relatively small.
[0003] When an electrical system is not connected to the Earth, as
in the case of an electrical vehicle, the system is said to have a
"floating" ground. The abbreviation IT comes from the French term
"isole terre" (isolated earth) and it is used by IEC (International
Electrotechnical Commission) to describe a power system with
"floating" ground. FIG. 2 shows such a system as it is implemented
in an electrified vehicle.
[0004] In this system, the high voltage battery and all the car
systems connected to it are isolated from the Chassis ground which
consists of the metal body of the car that passengers come in
constant contact with. The battery of an electric vehicle is
connected through DC to AC converters to motors, generators, which
are typically the same motors acting as generators when the car is
decelerating or moving downhill, and the various car auxiliary
systems through DC to DC converters. The two capacitors shown on
the right represent either capacitors placed with the purpose of
reducing EMI (electromagnetic interference) noise or the small
parasitic capacitances that exist in any electrical system.
[0005] This type of grounding serves an important purpose for the
safety of a car and those in contact with it. If for example the
negative of the battery was connected to the Chassis, and an
isolation fault occurred to one of the positive cables, an
immediate short would be created at the battery terminals causing
fuses to blow. This could result in immediate loss of power in the
electric vehicle--including loss of braking power--which could
result in accidents or other problems.
[0006] In contrast, in an IT power system as shown in FIG. 3, a
single isolation fault 340 would not cause an immediate power
failure. It also would not cause any danger for the car passengers,
as long as they don't touch the unaffected terminal which is
shielded from passenger access. Instead the driver will receive a
warning and will then be able to drive the car safely and in full
capacity to a service location. In case of an accident, which may
cause itself an isolation fault, the emergency first responders
will be warned to take safety precautions and to avoid touching
affected parts of the automobile.
[0007] An isolation fault may also occur through excessive
deterioration of insulating materials resulting from extreme
hot-cold cycles, by sparks and other electrical hazards or even by
rodents.
[0008] To address these potential risks, the National Highway
Traffic Safety Administration (NHTSA) of the Department of
Transportation (DOT) issued a final rule amending the electrical
shock protection requirements of Federal Motor Vehicle Safety
Standard No. 305 (49 CFR 571.305), which mandates for a DC voltage
system a minimum number of ohms/volt of isolation of a high voltage
source. In essence it specifies the maximum current that can pass
through the isolation resistance path, which cannot be more than 2
mA without an isolation monitoring device (500 ohm/volt) or 10 mA
with one (100 ohm/volt). As an example, for a DC voltage source of
400 volts with isolation monitoring, the specification is 100
ohm/volt. This translates into a minimum isolation resistance of 40
kilohms. Without isolation monitoring the minimum value for the
same system would be 200 kilohms.
[0009] The difference between the allowed values of ohms/volt
depending on whether an isolation monitoring device is present can
be appreciated by understanding the absence of one. If there is no
isolation monitoring device, the isolation resistance can be
measured only during scheduled service of the vehicle. The only
method the vehicle manufacturer has to ensure conformance in this
case is by using highly rated materials and wiring protection and
hoping that isolation faults will only happen gradually and will be
discovered during scheduled inspections. An isolation monitoring
device, in contrast, operates all of the time making it the
preferred method for future vehicle builds.
[0010] One can see the impact of these specifications by looking
again at FIG. 3. The isolation fault 340 would be the isolation
resistance (40 kilohm or 200 kilohm). The connection 315 would be
the resistance path of a human body touching the other terminal.
There would be a closed electrical circuit through the battery and
the chassis with the isolation resistance dominating the current
value as it is much higher than the human body resistance. At the
limit of 40 kilohms, the person touching the positive terminal
would still be safe albeit lightly shocked.
[0011] Together with establishing these requirements, the CFR
571.305 safety standard specifies a method for calculating the
isolation resistance. Referred to as the "voltage" method and
prescribed for use in vehicle service stations, the method will be
described in more detail below. Some improvements to this "voltage"
method also exist, about which more will be said below.
The "Voltage" Method
[0012] In the drawings of FIG. 5 the resistances 550 and 560
represent the isolation resistances between the negative and the
positive battery terminals to the chassis. Two voltmeters measure
the corresponding voltages U1 and U2. If there is a difference on
the readings, a known resistance 570 is connected to the side with
the higher voltage reading and using Ohm's law the smaller
isolation resistance is calculated. This method is referred to as
the "voltage" method and it is prescribed for use in vehicle
service stations. Although simple and straightforward the method
has serious drawbacks that limit its efficiency when used
continuously in an operating vehicle.
[0013] First, the insertion of a resistance in the isolation path
may adversely affect the isolation of the IT system. This is
because resistor 570 cannot have an arbitrarily high value as the
method relies on its effect upon the measured voltage. The
measurement method itself may jeopardize isolation. Also, switching
loads on and off in a high voltage system requires expensive
components. The measurement assumes that the battery voltage
remains constant during successive measurements. In an operating
vehicle this condition is rarely true (less than 20% of operating
time).
[0014] U.S. Pat. No. 9,322,867 presents a variation of the method
which overcomes the issue of the negative impact of resistor 570 by
using instead different types of current limiting devices. The
issues of high cost of high voltage switching and the battery noise
impact still remain unaddressed.
The "Pulse" Method
[0015] The pulse method is overcoming some of the problems
associated with the "voltage" method by injecting a pulse into the
DC network as shown in FIG. 6.
[0016] Variations of this method are well known and referenced at
EP 0 654 673 B1, EP 1 586 910 B1 and DE 101 06 200 C1. The main
shortcomings of this method are: [0017] The accuracy of the method
is susceptible to disturbances on the DC lines which typically
occur when the vehicle or system is operating. [0018] As a result a
large number of measurements have to be taken until reliable values
are obtained.
The "Frequency" Method
[0019] A variation of the "pulse" method is the "frequency"
injection method (U.S. Pat. No. 5,450,328, U.S. Pat. No. 9,069,025
B2). In this method an AC signal of known frequency is injected or
superimposed on the DC pulse. Through band-pass and low-pass
filtering of the resulting signal, the values of impedance and
resistance are estimated using digital signal processing
techniques.
[0020] The method requires digital signal processing capabilities
for digital filter implementation as well as the DFT/FFT processing
of the monitored responses. The accuracy also is affected by
dynamic changes in the load and achieves an acceptable level of
accuracy when load changes are small.
Characteristics of a "Good" Isolation Monitoring Device
[0021] Based on the analysis of the issues with prior art, it is
desirable for an isolation monitoring method to possess the
following characteristics: [0022] The method should not introduce
isolation hazards during measurement; neither should it drain
significantly the battery during continuous operation. [0023] It
should provide information not only for the isolation resistance
but also for the isolation capacitance which can also become
hazardous under certain conditions. [0024] The method should be
accurate not only in periods of system inactivity but under all
operating conditions [0025] The method should not produce "false
alarms". [0026] It should be fast and continuously updated. Ideally
it should be able to detect intermittent isolation faults.
The Method According to the Invention
[0027] The methods described so far are deterministic relying on a
unique known input (pulse, frequency, etc.) to produce an output
that can uniquely identify the unknown parameters of resistance and
capacitance. These methods are simple but fail in most instances
when the varying power load signal interferes.
Using the Power Load Signal as Excitation Source
[0028] It is beneficial for safety to be able to accurately
determine the isolation resistance and capacitance when the power
system is active. In the case of an electrified vehicle this would
correspond to 80% of the time the vehicle is in use.
[0029] The disclosed method uses these widely varying load signals
that naturally occur in an IT power system to identify the
isolation parameters. As a result, accurate information on the
isolation condition can be derived most of the time the system is
operational. An auxiliary excitation signal is used in periods of
system inactivity in order to ensure 100% monitoring
availability.
Method to Estimate the Element Values of the Isolation Path
[0030] According to this method, a model is used to represent the
IT power system along with the isolation resistances and
capacitances between the IT power system and the chassis ground.
The objective is to determine the values of the unknown resistive
and capacitive isolation paths between the IT power system and the
chassis ground. As shown in FIG. 6A and FIG. 6B those isolation
paths can be modeled as parallel RC circuits. Below, we will denote
R.sub.1 and C.sub.Y1 the resistor and capacitors in the RC circuit
for the isolation path on the negative side of the battery, and
R.sub.2 and C.sub.Y2 those for the isolation path on the positive
side of the battery.
The method of determining the values of R.sub.1, R.sub.2, C.sub.Y1
and C.sub.Y2 comprises: [0031] Measuring a first value of voltages
U.sub.1 and U.sub.2 associated with the isolation path between the
IT system and the ground chassis [0032] Measuring a second value of
voltages U.sub.1' and U.sub.2' associated with the isolation paths
between the IT power system and the ground chassis [0033]
Estimating the values for the model resistors R.sub.1, R.sub.2 and
capacitors C.sub.Y1 and C.sub.Y2 associated with the measurements
U.sub.1, U.sub.2, U.sub.1' and U.sub.2' through a function that
minimizes the discrepancy between the measurements and a
theoretical model for the RC circuits describing the isolation
paths [0034] Accepting the estimated values of each of R.sub.1,
R.sub.2, C.sub.Y1 and C.sub.Y2 of the isolation circuit model as
the present value estimate and storing it in a storage medium.
[0035] It should be appreciated that this method can utilize the
varying voltage of the IT power system as the measurement signal
for performing the calculations. If the voltage of the IT power
system is idle a voltage signal source can be used instead.
[0036] The function of minimizing the deviation between the voltage
measurements and the theoretical model for the RC circuits
describing the isolation paths can be a least-square error estimate
performed over a predetermined number of voltage measurements.
[0037] An improvement on the method can be achieved by utilizing a
stochastic filter, such as a Kalman filter, to minimize the
measurement and model noise.
The combined method consists of two steps:
[0038] 1. at first, a fixed number of measurements for the voltages
at the battery terminals and the excitation voltage are collected
and used as inputs for a least-square estimator which produces
estimates of the isolation parameters (i.e., leakage resistors and
leakage capacitors) together with uncertainties for those
estimates,
[0039] 2. the next step is a filtering step implemented using a
Kalman filter and designed to maintain the most likely values for
the isolation parameters with associated uncertainties by using the
results of the least-square estimator in time.
More details on those two steps are given below.
Isolation Parameters Prediction Using a Least-Square Estimator
[0040] During operation of the monitoring system, measurements are
collected for the values of the voltages at the battery terminals
and of the excitation voltage. The purpose of the least-square
estimator is to minimize the discrepancy between the measurements
and a theoretical model for the isolation paths modeled as RC
circuits, the latter expressing the conservation of charge in the
monitoring circuit. The least-square estimator therefore receives
as inputs a buffer containing a fixed number of the aforementioned
voltage measurements and produces as outputs predictions for the
isolation parameters, together with uncertainties for those
predictions. The predictions are expressed as a vector whose
components represent the isolation parameters, and the
uncertainties are expressed as a covariance matrix for this vector.
It follows that, as the monitoring system is operated, the
least-square estimator can be used at any time to provide a
prediction of the current isolation parameters vector and the
associated uncertainties. In the proposed method, the estimator is
used to regularly produce new predictions and uncertainties which
are then passed on to the filter described below. The number of
voltage measurements as inputs for the least-square estimator can
be predetermined or adjusted dynamically depending on the
conditions of operation.
Filtering of Predictions for Isolation Parameters Using a Kalman
Filter
[0041] It follows from the previous section that the monitoring
device uses the least-square estimator to obtain predictions and
uncertainties for the isolation parameters as a function of time.
That is, as the monitoring device is operated, time series of
predictions and uncertainties are generated by the least-square
estimator. Those time series can be seen as a stochastic process in
itself since the measured data sent to the least-square estimator
are themselves originating from a stochastic process. Therefore,
the purpose of the filter is to maintain estimates for the most
likely values for the isolation parameters vector and the
associated uncertainties. This is achieved by using a Kalman filter
implementation in which the results from the least-square estimator
are assimilated to noisy measurements of the isolation parameters.
The filter receives as inputs the previous estimates for the most
likely values of the isolation parameters and the associated
uncertainties, and predictions from the least square estimator. The
outputs are new estimates for most likely values of the isolation
parameters and the associated uncertainties.
Experimental Results
[0042] Experimental results are provided in FIGS. 12-16 based on
the apparatus shown in FIG. 8B to FIG. 10 and FIG. 17. The load
profile used was based on city driving data measured on a BMW i3
model. The load profile was accelerated 4 times in order to test
the algorithms in more adverse load conditions. Measured data along
with the produced estimates are provided in FIGS. 12-16.
BRIEF DESCRIPTION OF THE DRAWINGS
[0043] FIG. 1A is a diagram of electrical current paths through a
human body and each path's approximate value of resistance.
[0044] FIG. 1B is a table of biological effects upon the human body
of electrical currents.
[0045] FIG. 2 illustrates an IT grounding system of an electric
vehicle.
[0046] FIG. 3 illustrates an IT grounding system with a single
isolation fault.
[0047] FIG. 4 illustrates an IT grounding system of an electric
vehicle utilizing an Isolation Monitoring device.
[0048] FIGS. 5A-5C show the "voltage method" for determining
isolation resistance.
[0049] FIGS. 6A-6C show the "pulse method" improvement of injecting
pulses into the DC network.
[0050] FIG. 7 shows superposition of an injected AC signal on a DC
pulse in a network according to the "frequency method."
[0051] FIG. 8A shows the variation of battery voltage possible
during operation of an electric vehicle.
[0052] FIG. 8B shows a model as used to represent the IT power
system along with the isolation resistances and capacitances
between the IT power system and the chassis ground.
[0053] FIG. 9 shows a least-squared error method used to minimize
the error between model predicted values of the isolation path and
measurements.
[0054] FIG. 10 is a schematic representation of the two-step
process to estimate the isolation parameters and associated
uncertainties.
[0055] FIG. 11 shows a typical Kalman Filter implementation.
[0056] FIG. 12 is a plot of example results for the identification
of isolation parameters according to the invention.
[0057] FIG. 13 is a plot of voltage waveforms at the two battery
terminals in a device under test.
[0058] FIG. 14 shows an estimation of the isolation resistance
between the positive terminal and chassis.
[0059] FIG. 15 shows an estimation of the isolation resistance
between the negative terminal and chassis.
[0060] FIG. 16 displays the application of a Kalman filter on
initially noisy data in providing the estimate of isolation
resistance.
[0061] FIG. 17 is a block diagram of an isolation monitoring device
as implemented in hardware.
DETAILED DESCRIPTION OF THE DRAWINGS
[0062] When a human body contacts two points of non-identical
electrical potential, an electric current may flow through the path
between the points. Approximate values for each of these paths
through the human body are shown in FIG. 1A and the table in FIG.
1B shows typical reactions in the human body for given amounts of
electric current.
[0063] A high voltage battery system 200 with a "floating ground"
is shown in FIG. 2. Connected to the battery 201 and insulated from
the chassis ground 202 are motors 205, 215 and 225. Generators 210
and auxiliary systems 220 are also connected to the battery 201 and
insulated from the chassis ground 202. Actual capacitors and
modeled capacitances 230 and 240 are either placed or may exist
between terminals of the battery 201 and the chassis ground
202.
[0064] FIG. 3 shows a similar high voltage battery system 300 in
which the positive terminal 310 of the battery 301 is connected via
path 315 to the chassis 302. If an isolation fault 340 was to occur
between the negative terminal 305 and the chassis 302, a short
would be created between the battery terminals, causing fuses to
blow and affecting vehicle safety.
[0065] An isolation monitoring system alerts the operator and
responders to hazardous conditions that develop in an electrical
system. FIG. 4 shows a high voltage battery system 400 containing
an isolation monitoring device 440 according to the invention. The
device 440 monitors the isolation resistance between the terminals
405 and 410 of the battery 401 and the chassis ground 402, and
provides warnings of dangerous faults within the system during
vehicle operation.
[0066] FIGS. 5A-5C illustrate an implementation of the "voltage"
method in a high voltage battery system. In FIG. 5A, a battery 501
is isolated from the chassis ground 502. Resistances 550 and 560
represent the isolation resistances between the negative 515 and
the positive 525 battery terminals to the chassis 502. Two
voltmeters 555 and 565 measure the corresponding voltages u.sub.1
and u.sub.2. If there is a difference on the readings, a known
resistance 570 is connected to the side with the higher voltage
reading and, using Ohm's law, the smaller isolation resistance is
calculated. The method is illustrated in FIGS. 5B and 5C for either
the case of u.sub.2>u.sub.1 or u.sub.1>u.sub.2,
respectively.
[0067] An improvement over the "voltage" method, known as the
"pulse" method, is shown in FIG. 6A-6C. FIG. 6A shows a DC network.
In FIG. 6B, a pulse is injected into the DC network of FIG. 6A. The
impedance is then determined by monitoring the response over time,
as shown in FIG. 6C.
[0068] A variation of the "pulse" method, called the "frequency"
injection method also exists. In this method, an AC signal of known
frequency is injected or superimposed on the DC pulse. FIG. 7
illustrates superposition of DC and AC pulse injection on the
network. Through band-pass and low-pass filtering of the resulting
signal, the values of impedance and resistance are estimated using
digital signal processing techniques.
[0069] FIG. 8A shows an example of widely varying battery voltage
levels during operation of a Hybrid electric vehicle. The graph of
FIG. 8A plots battery voltage over time under different load
conditions. FIG. 8B is an illustration of a model used to represent
the IT power system along with the isolation resistances and
capacitances between the IT power system and the chassis
ground.
[0070] An exemplary step of minimizing the deviation between
measured and estimated values using a least-square estimator is
shown in FIG. 9. In the figure, a predetermined number of voltage
measurements 910 and a model for the isolation paths modeled as RC
circuits 920 are entered into a least-square estimator 930. The
least-square estimator function produces output predictions 950 for
the isolation parameters values and uncertainties.
[0071] As shown in FIG. 10, output predictions 950 from the
least-square estimator are used as inputs to a stochastic filter
1071. The stochastic filter 1071 may be, for example, a Kalman
filter. The stochastic filter 1071 also receives as inputs best
estimates 1060 for isolation parameters and uncertainties. Through
its operation, the stochastic filter 1071 outputs new best
estimates 1061 for isolation parameters and uncertainties. A second
iteration of the stochastic filter 1072 receives updated output
predictions 951 from the least-square estimator together with best
estimates 1061 for isolation parameters and uncertainties. This
second operation of the stochastic filter 1072 outputs new best
estimates 1062 for isolation parameters and uncertainties. A detail
of typical Kalman filter operation is shown in FIG. 11.
[0072] A graph of example results of the best estimate for the most
likely value of capacitor C.sub.Y1 in the monitoring circuit is
shown in FIG. 12. The best estimate is plotted as curve 1210
between curves 1205 and 1215, which themselves represent the
narrowing confidence interval for the estimated value due to
filtering. The unfiltered, noisy predictions from the least-square
estimator are shown in the background as the widely varying curve
1220.
[0073] Voltage waveforms at the two battery terminals of the
apparatus of FIG. 8B under test are shown in FIG. 13. Estimation of
the isolation resistance between the positive terminal and the
chassis is plotted in the graph of FIG. 14. Similarly, estimation
of isolation resistance between the negative terminal and the
chassis is plotted in the graph of FIG. 15. Both FIG. 14 and FIG.
15 show the method prediction as well as the method confidence
interval under a varying load profile for a given period of time.
The estimate of isolation resistance provided by a Kalman filter is
shown together with the raw results of the least-squared algorithm
in FIG. 16.
[0074] A block diagram for the Isolation Measurement Device
implemented in hardware is illustrated in FIG. 17.
[0075] What has been described is a method to estimate a change in
values of isolation impedance in an isolated ground (IT) electrical
system comprising a power source, the method comprising: modeling a
first isolation path between a first reference point and a second
reference point and modeling a second isolation path between a
third reference point and a fourth reference point, thereby
creating a theoretical model of the isolated ground electrical
system; providing an initial value of a first isolation resistance
for the first isolation path and an initial value of a second
isolation resistance for the second isolation path; measuring an
initial value of a voltage between the first reference point and
the second reference point and storing the measured initial value
in a storage medium; measuring an initial value of a voltage
between the third reference point and the fourth reference point
and storing the measured initial value in the storage medium;
measuring a subsequent different value of the voltage between the
first reference point and the second reference point and storing
the measured subsequent value in the storage medium; measuring a
subsequent value of the voltage between the third reference point
and the fourth reference point and storing the measured subsequent
value in the storage medium; entering the measured initial values
of the voltages, the measured subsequent values of the voltages,
the provided values of the isolation impedances and an elapsed
amount of time between the initial measurements and the subsequent
measurements into a mathematical function stored in the storage
medium; wherein the mathematical function minimizes the discrepancy
between the measured change in values of the voltages and the
modeled theoretical values by adjusting values of modeled isolation
impedances associated with the isolation paths in the electrical
system; extracting estimated values of isolation impedances
associated with the isolation paths in the electrical system by
application of the mathematical function; and storing the estimated
values in the storage medium.
[0076] Also described is an apparatus for estimating a change in
values or unknown values of isolation impedance in an isolated
ground (IT) electrical power system, comprising: a power source
having a positive terminal and a negative terminal, said terminals
connected in circuit to at least one additional electrical
component and isolated from a chassis ground within the electrical
system;
[0077] wherein the electrical system contains an isolation
impedance between each of the terminals and the chassis ground; a
storage medium; means measuring an initial value and a subsequent
different value of a voltage between the chassis ground and a first
reference point and between a second reference point and a third
reference point in the electrical system; means storing the
measured initial values and the subsequent different values in the
storage medium; a mathematical function stored in the storage
medium, whereby application of the mathematical function extracts
estimated values of isolation impedances associated with the
voltage measurements by using a model of the electrical system and
minimizing an error function.
[0078] Also described is a method to estimate unknown values of
isolation impedance in an isolated ground (IT) electrical system
comprising a power source and a load, the method comprising:
modeling a first isolation path between a first reference point and
a second reference point and modeling a second isolation path
between a third reference point and a fourth reference point,
thereby creating a theoretical model of the isolated ground
electrical system; at a time when power from the power source is
being dissipated in the load, measuring an initial value of a
voltage between the first reference point and the second reference
point and storing the measured initial value in a storage medium;
at a time when power from the power source is being dissipated in
the load, measuring an initial value of a voltage between the third
reference point and the fourth reference point and storing the
measured initial value in the storage medium; at a time when power
from the power source is being dissipated in the load, measuring a
subsequent different value of the voltage between the first
reference point and the second reference point and storing the
measured subsequent value in the storage medium; at a time when
power from the power source is being dissipated in the load,
measuring a subsequent value of the voltage between the third
reference point and the fourth reference point and storing the
measured subsequent value in the storage medium; entering the
measured initial values of the voltages, the measured subsequent
values of the voltages and an elapsed amount of time between the
initial measurements and the subsequent measurements into a
mathematical function stored in the storage medium; wherein the
mathematical function minimizes the discrepancy between the
measured initial values of the voltages, the measured subsequent
values of the voltages, and the modeled theoretical values by
adjusting values of modeled isolation impedances associated with
the isolation paths in the electrical system; extracting estimated
values of isolation impedance associated with the isolation paths
in the electrical system by application of the mathematical
function; identifying a minimum resistance path from the estimated
values of isolation resistance; and storing the estimated values in
the storage medium.
[0079] It will be appreciated that one skilled in the art of
isolated ground electrical systems, varying output power sources
and electrical systems could devise additional obvious improvements
and variations upon the invention described and claimed herein. All
such obvious improvements and variants are intended to be
encompassed by the claims which follow.
* * * * *