Load identification AC power supply with control and method
阅读说明:本技术 具有控制的负载识别ac电源及方法 (Load identification AC power supply with control and method ) 是由 马克·特勒富斯 哈里·罗德里格斯 塔尔·凯西 克里斯·凯西 于 2017-08-11 设计创作,主要内容包括:描述了一种改进的AC电源。所述电源通过监控与AC干线相关的电流和电压的波形和相位来识别负载。在以下条件下进行比较:通过使用位于在AC干线和负载之间的电路和中性线中的控制开关来可编程地改变提供给负载的功率。控制开关的程序是变化的,以优化区分相似负载类型的能力。该开关可以进一步用于控制提供给负载的功率,其根据基于负载特性的一组规则而变化。在优选的实施例中,该设计使得以可以完全地集成到硅上的最少的部件得到高的效率。(An improved AC power supply is described. The power supply identifies the load by monitoring the waveform and phase of the current and voltage associated with the AC mains. The comparison was carried out under the following conditions: the power supplied to the load is programmably varied by using control switches located in the circuit and neutral between the AC mains and the load. The program controlling the switches is varied to optimize the ability to distinguish between similar load types. The switch may further be used to control the power supplied to the load, which varies according to a set of rules based on the characteristics of the load. In a preferred embodiment, this design results in high efficiency with a minimum of components that can be fully integrated onto silicon.)
1. A power supply for connecting an AC power source to an electronic load and identifying the load, the power supply comprising:
a.an AC to DC converter, and a.C.,
b. an electronic switch, wherein the switch comprises a switch controller and the switch controller provides phase angle modulation of a voltage from the AC power source to the load, and,
c. a first voltage sensor for monitoring a voltage of the AC power source, and,
d. a second voltage sensor for monitoring a voltage applied to the load, and,
e. a current sensor for monitoring the current consumed by the load, an
f. A microprocessor powered by an AC-to-DC converter and programmed to accept inputs from the first voltage sensor, the second voltage sensor, and the current sensor and control the switch controller such that a first set of waveforms for the first voltage sensor, the second voltage sensor, and the current sensor are acquired during a first time period after the load is connected to the power source and a second set of waveforms for the first voltage sensor, the second voltage sensor, and the current sensor are acquired during a second time period after the load is connected to the power source, each of the first set of waveforms and the second set of waveforms having an amplitude and a phase offset relative to each other, and
g. reducing, by the switch, the voltage to the load during a second time period using phase angle modulation of the AC voltage to the load, an
h. The microprocessor is further programmed to identify the load by comparing the first set of waveforms to the second set of waveforms.
2. The AC power source of claim 1, wherein identifying the load comprises identifying the load as a selected one of: pure resistive loads, constant power resistive loads, pure reactive loads, and constant power reactive loads.
3. The AC power source of claim 1, wherein power to the load is controlled by the switch based on a preselected set of rules identifying the load and associated with the identification of the load.
4. The AC power supply of claim 1, wherein all components of the AC power supply are integrated on silicon.
5. The AC power supply of claim 1, wherein said AC to DC converter is made according to fig. 7.
6. The AC power supply of claim 1, wherein said electronic switch is made according to fig. 8.
7. The AC power source of claim 1, wherein the load is galvanically isolated from the AC power source.
8. The AC power source of claim 1, wherein the second voltage sensor and the current sensor are galvanically isolated from the load and the AC power source.
9. The AC power source of claim 1, wherein the AC power source is located in a power panel.
10. The AC power source of claim 1, wherein the AC power source is located in a socket box.
11. The AC power source of claim 1, wherein said AC power source is located in an extension line.
12. The AC power source of claim 1, wherein said electronic load is a plurality of electronic load devices.
13. The AC power source of claim 1, wherein comparing the first set of waveforms to the second set of waveforms comprises:
a. comparing a phase of the waveform of the second voltage sensor and a phase of the waveform of the current sensor with a phase of the waveform of the first voltage sensor, and
b. comparing the product of the amplitude of each of the second voltage sensors and the amplitude of the current sensor during the first time period with the product of the amplitude of each of the second voltage sensors and the amplitude of the current sensor during the second time period.
14. The AC power source of claim 1, wherein comparing the first set of waveforms to the second set of waveforms comprises:
a. during the first time period, comparing a high frequency component of the waveform of the first voltage sensor to a high frequency component of the waveform of the second voltage sensor, and
b. during the second time period, comparing a high frequency component of the waveform of the first voltage sensor with a high frequency component of the waveform of the second voltage sensor, and
c. comparing the high frequency components during the first and second time periods generates a pattern characteristic of a characteristic of the load.
15. A method of identifying an electronic load connected to an AC power source, the method comprising:
a. acquiring a waveform of the voltage of the AC power source during a first period of time,
b. acquiring a waveform of the voltage across the load during a first time period,
c. obtaining a waveform of the current through the load during a first time period, and,
d. reducing a voltage of the AC power to the load during a second time period, and,
e. acquiring a waveform of the voltage of the AC power source during a second period of time,
f. acquiring a waveform of the voltage across the load during a second time period,
g. acquiring a waveform of the current through the load during a second time period, an
h. Each waveform has an amplitude and a phase relative to each other, an
i. The load is identified by comparing a waveform acquired during a first time period to a waveform acquired during a second time period.
16. The method of claim 15, wherein identifying the load comprises identifying the load as a selected one of: pure resistive loads, constant power resistive loads, pure reactive loads, and constant power reactive loads.
17. The method of claim 15, further comprising: controlling power to the load based on a preselected set of rules that identify the load and are associated with the identification of the load.
18. The method of claim 15, wherein comparing the first set of waveforms to the second set of waveforms comprises:
a. comparing a phase of the waveform of the second voltage sensor and a phase of the waveform of the current sensor with a phase of the waveform of the first voltage sensor, and
b. comparing the product of the amplitude of each of the second voltage sensors and the amplitude of the current sensor during the first time period with the product of the amplitude of each of the second voltage sensors and the amplitude of the current sensor during the second time period.
19. The method of claim 15, wherein comparing the first set of waveforms to the second set of waveforms comprises:
a. during the first time period, comparing a high frequency component of the waveform of the first voltage sensor to a high frequency component of the waveform of the second voltage sensor, and
b. during the second time period, comparing a high frequency component of the waveform of the first voltage sensor with a high frequency component of the waveform of the second voltage sensor, and
c. comparing the high frequency components during the first and second time periods generates a pattern characteristic of a characteristic of the load.
Technical Field
The present invention relates to an AC power source and method for identifying a connected electronic load and controlling AC power to the load based on the identification of the load.
Background
A conventional method of supplying AC power from the AC mains to the devices within the home is through a plug-in outlet. Typically the socket does not include active electronics and is simply a connector. Newer outlets include fault detection circuitry, but rarely provide any means to measure or control the AC power delivered to the connected device. Recent approaches for improving power distribution within a home include home area networks that interact with communication means including, for example, wired and wireless local area networks. Typically controlled by an application programmed on a personal computer and personal device, such as a smartphone or tablet. Another approach is by including additional electronics in the wall switch. Newer programmable thermostats are used to control central heating and air conditioning. While these devices provide improved control and feedback of energy usage, they rely on improved electronics within the device itself, without improving the distribution of AC power to older, legacy devices plugged into a wall outlet.
Government agencies such as the U.S. department of Energy through government agencies such as Energy
Program (Energy Star is a trademark registered by the U.S. department of Energy) establishes standards for new devices and electrical appliances certified for low Energy consumption. In many cases, the AC power provided to a device is intelligently managed through reduced energy usage when the device is in an idle or sleep mode, thereby reducing energy consumption. As such, the focus is on new equipment and appliances, without taking steps for a large number of installed equipment. Intelligent control of new devices typically requires knowledge of the nature of the device. The power consumed by the device is managed by a built-in set of rules programmed on a microprocessor located within the device that controls the AC power. For example, during idle times, the washing machine may be completely disconnected from power and wait for the next load to be manually turned on. However, the refrigerator cannot be turned off as such because the power must be maintained to monitor the temperature and the compressor activated to maintain the set point. Other devices and appliances, such as televisions, computers, displays, and printers, may have a set of rules that are enacted by monitoring usage and time of day. When it is known from past usage history that a device is typically not in use, power to the device may be significantly reduced. In some cases, there are some user settings that allow the user to select the speed at which the device enters a low power sleep mode. Also, all of these energy usage improvements are typically incorporated into the device itself. Improving power consumption by external control of the power supply, while possible, requires knowledge of the nature of the device. In some cases, the general variety of types of devices, such as lighting, refrigerators, etc., is sufficient to provide work that will be reduced byA set of rules that improve performance is consumed. The first step is to be able to identify the load device. There is a need for a power supply in the form of a smart socket or connector between the AC mains and the electrical load device and which includes a means of identifying the devices connected to each other so that the identification can be used to control the power supplied to the devices.Conventional means of identifying the load are insufficient. Waveform analysis to look for phase shifts caused by loading is well known. More exotic systems use waveform analysis, including pattern matching of high frequency patterns present on the current and voltage waveforms. Despite these improvements, difficulties still exist in distinguishing between similar loads, such as two loads that are primarily resistive or two loads that each include an electric motor. The deep learning method applied to the high frequency components of the waveform is still insufficient to fully identify the connected electronic load. Improved waveform analysis is needed to identify loads connected to an AC mains power supply. There is a need for an AC power supply that can be fully integrated into any of the following: a power panel, a socket box connected to the power panel, a terminal block, or an extension cord attached to the socket box. There is a need for a load identification and control system that can be fully integrated onto silicon.
Disclosure of Invention
A load identification AC power system is described that includes electronics for identifying a load connected to an AC power source and controlling power provided to the load based on the identification. The load identifying AC power may be integrated into existing power panels, receptacle boxes, or into connectors such as wires and terminal blocks. In one embodiment, each receptacle on a wire or patch panel includes electronics that identify an attached load and control the power delivered to the load. In one embodiment, the load identification AC power source of the present invention includes voltage and current sensors, as well as load demand sensors. Real-time waveform analysis of the voltage and current supplied to and demanded by the load is done by the microprocessor. The load identifying AC power supply further includes a programmable switch in series with the load that can operate at a frequency higher than the frequency of the AC mains and can both turn the power supply on and off and use pulse angle modulation to control the power supplied to the load. Load identification AC power includes functionality whereby the AC power provided to the load is regulated based on the identification of the load. In one embodiment, the load identifying AC power source includes a microprocessor programmed to control the switches and acquire the current and voltage waveforms, then identify specific patterns and relationships in the voltage and current waveforms, and associate these patterns with a particular connected load device or devices. The waveform is analyzed by a set of rules that classify the nature of the load or by a pattern matching technique. Rule-based and pattern matching techniques are enhanced by waveform analysis with or without programmable switches to improve discrimination between different load types. In one embodiment, a voltage regulator, consisting of switches in series with an AC circuit or neutral, is modulated by chopping segments of the AC source sine wave, thereby changing the effective voltage across the load by phase angle "chopping" or Phase Angle Modulation (PAM). By applying PAM to the AC source, the AC voltage across the load will reduce the effective voltage drop proportional to the angle. The current and voltage waveforms of the load are monitored before, during and after the modulation of the supply voltage. In one embodiment, loads that include power management are distinguished from loads that do not include power management by observing the power management system's reaction to a reduced power supply voltage, as reflected in the voltage, current, and power waveforms. In another embodiment, a preselected pattern of changes in power supplied to the load is applied over a limited number of duty cycles. In one embodiment, waveforms are analyzed and classified using neural network analysis and based on both undisturbed and varying power supplied to the load. In another embodiment, the programming control of the switches is optimized based on the ability to differentiate between load types. In another embodiment, where multiple load types are connected to the same circuit, the programming of the control switches is optimized to maximize the number of distinguishable connected loads. In one embodiment, the identification of the device is limited to a range of conventional categories of devices. Non-limiting examples of classes include resistive loads, capacitive loads, inductive loads, and the three types of loads that further include power factor correction devices for maintaining constant power under varying supply voltages.
The AC power supply includes a connection to the AC mains, an AC/DC converter to provide DC power to the microprocessor, current and voltage sensors, and a programmable switch. The voltage sensor utilizes a resistive voltage divider and senses current using a current sensing resistor, a current amplifier, and a hall effect sensor. The sampled results are typically processed by a comparator, an analog-to-digital converter (ADC), and stored in a data storage element. In a preferred embodiment, both the AC/DC converter and the programmable switch use a design that enables the entire AC power supply to be integrated onto silicon.
The specific examples are not intended to limit the inventive concept to example applications. Other aspects and advantages of the invention will be apparent from the accompanying drawings and from the detailed description.
Drawings
Fig. 1 is a schematic diagram depicting aspects of prior art electronic load identification.
Fig. 2 is a schematic illustration of a first embodiment of improved electronic load recognition of the version of fig. 1.
Fig. 3 is a schematic diagram of a second embodiment of improved electronic load recognition of the version of fig. 1.
Fig. 4 is a flow chart of an improved electronic load identification method.
Fig. 5 is a block diagram of the electronics of the load identification AC power supply of the present invention.
Fig. 6 is a block diagram of an AC to DC converter used in a preferred embodiment of a load identifying AC power source.
Fig. 7 is a circuit diagram of a preferred embodiment of the AC-to-DC converter of fig. 6.
Fig. 8 and 9 are circuit diagrams of aspects of a programmable switch used in a preferred embodiment of a load identifying AC power source.
Fig. 10 is a block diagram of a load identification AC power source further including electrical isolation of the load from the AC source.
Fig. 11 is a block diagram of a load identification AC power source further including electrical isolation of the load sensor from other components of the load identification AC power source.
Detailed Description
Referring to fig. 1, a typical prior art method for identifying a load attached to an AC source is shown. The graph depicts the general waveforms used in prior art analysis and includes the
Referring to fig. 2, a first embodiment of an analysis method included in the load recognition AC power supply of the present invention is shown. The graph is the voltage of the
Fig. 3 shows waveform analysis in another embodiment of the present invention. The data for these graphs is the same as that depicted in fig. 2. Time t0 is the start of a baseline time or data acquisition. At time t1, the load device is connected to an AC mains circuit that includes a load identification AC power source. At time t2, the load is sensed and the series connected switches are activated to provide power to the device. The load consumes power at time t 3. At a later time t4, the power supplied to the device is changed. In this case, the power supplied to the device is reduced during the two
The waveforms of the AC mains and the voltage and current across and through the load are recorded and analysed at a sampling frequency which records a cycle time which is significantly greater than a single cycle of the AC mains. The sampling frequency of the voltage and current waveforms is selected as needed to distinguish between load types. In one embodiment, the sampling frequency is in the kilohertz range. In another embodiment, the sampling frequency is in the megahertz range. In a preferred embodiment, the programmed variation in power applied to the load is selected to optimize the discrimination of the acquisition waveform between expected load types. In one embodiment, the analysis of the waveform includes matching patterns in high frequency components of the voltage and current waveforms from the load. In another embodiment, the analysis of the waveform includes determining a delay in time of the load consuming power after the power is first applied to the load. In another embodiment, analyzing means classifying the acquired waveform, including its high frequency components, into groups representing different load types. Non-limiting examples of groups include waveforms representing the following loads: waveforms for primarily resistive loads, capacitive loads, inductive loads, loads including power factor correction, and loads including power control, such loads having a delay in the power provided to the load when the power from the power source is initially applied.
Referring now to fig. 4, a method of controlling an AC source using a load is shown. A load control appliance is installed 401. In one embodiment, the installing includes electrically connecting the load control device between the AC mains power supply and the load. In one embodiment, mounting includes mounting the load control device at a junction box. In another embodiment, the installing includes installing the load control AC source in a wall socket. In another embodiment, the installing includes installing the load control device as a power supply patch panel or smart extension cord by plugging the load control device into a conventional wall outlet, and the load will be plugged into the load control device. Once the load control device is installed 401, a load is attached to the load control device 402. The load control device detects the load 403 and provides power to the load by activating a switch within the load control device. The details of the switches and load control devices are shown in subsequent figures. Once the load is detected, data acquisition 404 is initiated. Data acquisition includes recording the time when the load is connected to the power source, when power is applied to the load, and when the load consumes power. The data acquisition further includes acquiring waveform data. Once a load specific load is detected, any data acquired is referred to as "load data". The load data includes the time the load was on and waveform data. The waveform data includes values of the acquired AC mains voltage, the load current, and the power consumed by the load as a function of time. All data is obtained at a frequency optimized for detecting the type of load. In one embodiment, the data is acquired at a frequency that is several times higher than the frequency of the AC mains power supply. In one embodiment, data is acquired at a rate of kilohertz for 50 to 60 cycles of data from an AC source. In another embodiment that relies on high frequency components of the voltage and current waveforms to identify the load, data is acquired at a megahertz rate. The acquired load data is stored 409 for analysis. In one embodiment, the memory comprises memory in a short term random access memory of the microprocessor for immediate or near immediate processing. In another embodiment, the memory includes memory in long-term memory such that the stored load data is used for subsequent pattern matching to identify the same or similar loads based on the matching of waveform patterns, i.e., the waveform pattern obtained when the load 402 is first connected (e.g., when first traversing the represented flow chart) matches the waveform pattern of the same or different loads that are subsequently connected. In one embodiment, the memory 409 comprises a memory accessible by a plurality of load control devices. Such a memory may be accessed by a device that is connected to the load control AC source, either wired or wirelessly, or by transferring stored load data from a first load control AC source to a second load control AC source. Once connected 402 and detected 403, and after initial data acquisition 404, the power provided to the device is modulated 405. Modulation means the use of programmable switches to vary the power supplied to the device. During and after modulation further load data is obtained 406 and the load is then identified 407 based on the load data. In one embodiment, the identification is based on comparing the waveform of the load data with waveforms in previously acquired load data of known load devices. In another embodiment, the load is identified based on both the timing of the power around the turn-on load and the matching of the waveform data, as already discussed. In another embodiment, neural network analysis is employed to classify load data into categories of load types by comparison with a previous load database. In another embodiment, the identification of the load means classifying the load as a particular class of load based on the phase relationship between the voltage and current waveforms of the load and the AC mains voltage waveform before, during and after modulating the power of the load using the series switches. In one embodiment, the load is identified 407 as one of:
1. a purely resistive load. The voltage and current synchronously zero-cross and peak, both before, during and after modulation of the supply voltage. The power is reduced when the voltage is reduced and returns to the level before modulation when the modulation of the supply voltage is stopped and the supply power returns to full voltage.
2. Constant power resistive load with power correction. The peaks of the voltage and current are synchronized before modulation and the power is constant before, during and after modulation.
3. Purely reactive (resistive or inductive) loads. The voltage and current are out of phase before, during and after modulation, the supply voltage is modulated and the power is reduced, and when modulation of the supply voltage is complete and the supply voltage returns to full voltage, the power returns to the level before modulation.
4. A constant power reactive load. The voltage and current are out of phase before, during and after modulation, and the power is constant before, during and after modulation of the supply voltage.
In one embodiment, the modulation of the supply voltage may result in a reduction of the effective value of the supply voltage by an amount between 1 and 20%.
In one embodiment, identifying 407 further includes determining a confidence level (confidence level) for the identification. In one embodiment, the confidence level is determined by a goodness of fit matching the load data obtained during the data acquisition steps 404, 406 to data previously obtained and stored 410 over known loads. Once the identification step 407 is completed, the system further checks 408 whether a load has been identified and whether there are control rules associated with the load identification. In one embodiment, the checking 408 of the recognition is accomplished by comparing the confidence level in the recognition to a preselected confidence level defined as a correct recognition. If the load is correctly identified and there are pre-selected control rules associated with the identified load, control of the load is performed 409. In a preferred embodiment, the power of the load is then controlled by a switch in series with the load. Non-limiting examples of pre-selected control rules include:
1. during daytime times, purely resistive loads such as light bulbs are dimmed to reduce power usage, especially during peak demand periods.
2. In a constant power load, as the load demand decreases, the input power will also decrease accordingly to minimize the power consumption of no load/minimum load demand.
3. At a remote location (no human present), the pure resistive load and the constant power resistive load will automatically disconnect and reconnect according to the load requirements.
4. Equipment that generates an arc during normal operation (e.g., having brushes connected to the rotor) is ignored by the arc fault circuit interrupter to prevent a disconnection hazard.
In another embodiment, there is a preselected set of rules based on whether the load is one selected from the group consisting of: pure resistance, constant power resistance, pure reactance, and constant power reactance. In a non-limiting example of a preselected rule, a load identified as having included power factor correction, i.e., a constant power load, is not turned off by the controller, and a purely resistive load is turned off during a preselected time period and the power to the purely reactive load is reduced during the preselected time period.
The components in various embodiments of the load identification AC power source can be seen in fig. 5-11. Referring first to fig. 5, an
In one embodiment, the AC/DC converter may be of any type known in the art that can provide voltages and power suitable for microprocessors, sensors, and switch controllers. Such AC/DC converters include rectifier and transformer components to provide the selected voltage and power required by the sensor and microprocessor circuits. Likewise, the
AC to DC power supply
Details of the AC-to-
Fig. 7 shows a preferred embodiment of the AC to DC converter.
In the preferred embodiment of fig. 6 and 7, the AC-to-DC converter is made up of
The preferred embodiment of fig. 7 generally includes a
Switch with a switch body
Switch 508 (fig. 5) is an integral part of the present invention. The power of the
The on time constant depends on the values of the current limiting
Fig. 9 shows an embodiment that uses two switching
In another embodiment as shown in fig. 10, the
In another embodiment shown in fig. 11, the sensor connected to the
Summary of the invention
An improved AC power supply is described. The power supply identifies the load by monitoring the waveform and phase of the current and voltage associated with the AC mains. The comparison is made under the condition that the power supplied to the load is programmably varied by using control switches located in the circuit and neutral between the AC mains and the load. The program controlling the switches is varied to optimize the ability to distinguish between similar load types. The switch may further be used to control the power supplied to the load, which varies according to a set of rules based on the characteristics of the load. In a preferred embodiment, this design results in high efficiency with a minimum of components that can be fully integrated onto silicon.
- 上一篇:一种医用注射器针头装配设备
- 下一篇:电力变换装置以及电力变换系统