Property equipment whole life cycle management system based on multi-source data fusion

Abstract

This invention relates to the fields of industrial internet and intelligent sensing technology, and discloses a property equipment full lifecycle management system based on multi-source data fusion. The system synchronously collects mechanical operating condition and power distribution network electrical quality characteristic data, and utilizes an edge computing hub to construct a cascaded judgment logic for a first feature value and a frequency domain interference scalar. When interference exceeds a threshold, the system generates a first control command and an associated first timing marker, forcing the target electrical load to maintain a preset output within a defined time interval. This decouples spectral aliasing in the physical time domain, achieving high-fidelity extraction of the mechanical stator current feature vector and outputting a control signal accordingly. This invention overcomes the problem of signal extraction distortion caused by cross-domain interference in shared power grids, significantly improving the diagnostic accuracy and system resilience of equipment management in complex environments.

Description

Technical Field

[0001] This invention relates to the fields of industrial internet and intelligent sensing technology, and in particular to a property equipment full lifecycle management system based on multi-source data fusion. Background Technology

[0002] In the micro-failure diagnosis of large-scale property equipment clusters, current technologies for monitoring the degradation of mechanical equipment (such as sewage pumps) typically rely on the extraction of electrical domain characteristics of the stator current of drive motors. In modern building microgrids, the grid-connected operation of numerous nonlinear electrical loads (such as charging and discharging equipment) deeply injects high-energy switching harmonics into the power distribution trunk lines. The weak fluid anomalies in the early stages of electromechanical equipment modulate the frequency bands of these high-frequency harmonics, resulting in severe physical coherence and spectral aliasing.

[0003] To address the aforementioned spectral aliasing phenomenon, existing technologies generally employ high-order digital filtering algorithms to attempt to remove noise. However, constrained by the underlying physical limitations of the Porter's integral theorem, under extremely low signal-to-noise ratio conditions, digital filters with strong frequency selectivity inevitably introduce group delay and phase distortion into the signal processing chain. This distortion causes microsecond-level mechanical transient characteristics to diffuse and collapse in the time domain, resulting in missed detections at the feature extraction layer.

[0004] The aforementioned signal extraction failures triggered a series of destructive consequences along the physical shared network: on the mechanical side, weak fluid anomalies (such as false dry pumping caused by high-viscosity mud) continued to develop due to missed detection, and the passive liquid level sensor feedback was lagging, causing the motor to suffer physical thermal damage; on the electrical side, uncontrolled continuous high-frequency harmonics caused skin heating effect at the busbar nodes of the distribution cabinet, and the continuous accumulation of thermal stress formed a "thermal ratchet cycle", which accelerated the fatigue loosening of fasteners.

[0005] In summary, existing technologies face a core underlying bottleneck under multi-physics interference conditions: simply relying on "algorithm optimization in the software data dimension" has reached the physical boundary and cannot fundamentally decouple cross-domain environmental interference in the shared power grid, so that microscopic signal distortion will inevitably evolve into macroscopic cross-domain cascade physical failure. Summary of the Invention

[0006] This invention provides a property equipment lifecycle management system based on multi-source data fusion to solve the technical problem that existing management systems, under the condition of multiple devices sharing a physical network, are limited by the physical boundary limits of pure data processing algorithms, and cannot effectively decouple cross-domain environmental interference, causing the failure of micro signal feature extraction to easily evolve into macro-level cross-domain cascading physical failure.

[0007] In view of the above problems, the present invention provides a property equipment full life cycle management system based on multi-source data fusion, comprising: a data acquisition layer, a control center and an execution layer, wherein the execution layer includes mechanical equipment and target electrical loads; The data acquisition layer is configured to: acquire environmental operating condition data of the mechanical equipment and stator current data of the mechanical equipment; and simultaneously acquire electrical quality characteristic data of the common connection point of the power distribution network shared by the mechanical equipment and the target electrical load; The control center is configured to execute the following control logic: Based on the environmental condition data, calculate the first characteristic value of the mechanical equipment; When the first feature value is greater than the first preset threshold, the frequency domain interference scalar of the distribution network common connection point to the stator current data is calculated based on the electrical quality feature data. Calculate the ratio of the frequency domain interferometric scalar to the first eigenvalue; When the ratio is greater than the second preset threshold, a first control command for the target electrical load is generated, and a first timing mark associated with the first control command on the time axis is generated. Within a first time interval defined by the first timing marker, feature extraction is performed on the stator current data to generate a second feature vector; The target electrical load is configured to maintain a preset electrical parameter output state within the first time interval in response to the first control command. The control center is configured to output operation control signals for the mechanical equipment based on the second feature vector.

[0008] Preferably, the mechanical equipment is a sewage pump, and the target electrical load includes a charging device with an inverter or rectifier topology.

[0009] Preferably, the environmental condition data includes precipitation probability parameters and liquid level time integral values; the control center is configured to use a preset first weighting factor and a second weighting factor to perform weighted summation calculations on the precipitation probability parameters and the liquid level time integral values ​​respectively, and output the weighted summation result as the first feature value.

[0010] Preferably, the electrical quality characteristic data includes the measured total harmonic distortion rate; when the control center calculates the frequency domain interference scalar, the input parameters extracted include the measured total harmonic distortion rate, a first preset constant, a second preset constant, and the coherence coefficient between the first harmonic frequency value of the distribution network common connection point and the preset characteristic frequency value of the mechanical equipment.

[0011] Preferably, the control center is configured to calculate the frequency domain interference scalar through the following logic: The result of the power operation of the ratio of the measured total harmonic distortion rate to the first preset constant is used as the reference input, wherein the exponent of the power operation is the second preset constant; The coherence coefficient is used as the gain coefficient, which is then multiplied by the reference input and the result is calculated using a negative exponent. The frequency domain interferometric scalar is generated by subtracting the negative exponent from the constant and multiplying it by the first eigenvalue.

[0012] Preferably, the control center calculates the frequency domain interference scalar using the following formula: in, For the frequency domain interference scalar, The first feature value, It is an exponential function with the natural constant as its base. The coherence coefficient is... The measured total harmonic distortion (THD) is... It is the first preset constant. It is the second preset constant and .

[0013] Preferably, the control center is equipped with a read-only memory containing a one-dimensional static data table; when performing the negative exponentiation operation, the control center extracts the integer part of the product of the coherence coefficient and the reference input quantity and converts it into an address pointer, reads the reference convergence value from the one-dimensional static data table, extracts the fractional part of the product and performs linear interpolation operation on adjacent reference convergence values, and outputs an approximate exponentiation result to participate in the calculation of the frequency domain interferometric scalar.

[0014] Preferably, the first control instruction includes a device communication identifier, a target power limit value, and an execution time period; the start timestamp of the first timing marker is the same as the start timestamp of the execution time period, and the length of the first time interval is equal to the length of the execution time period.

[0015] Preferably, when the second feature vector matches a preset abnormal feature pattern, the control center sends a second control command to the mechanical equipment; the second control command includes alternating forward and reverse motor drive electrical signals.

[0016] Preferably, when the time interval since the last generation of the first control command is greater than a preset time period and the electrical quality characteristic data is greater than a third preset threshold, the control center restarts and executes the calculation and judgment control logic of the frequency domain interference scalar.

[0017] The technical solution provided in this application has at least the following technical effects: Addressing the underlying technical problem of cross-domain environmental interference, this invention abandons the traditional digital filtering link. Instead, it calculates the frequency domain interference scalar of the measured total harmonic distortion rate and a preset characteristic frequency, and uses this to trigger the issuance of a first control command to the nonlinear electrical load, thus opening a physical-level transient silent observation window on the time axis. This mechanism utilizes the coordinated action of space equipment to forcibly isolate anomaly feature extraction from high-frequency noise in the physical time domain, bypassing the phase distortion limit of software algorithms and achieving high-fidelity extraction of transient features under extremely low signal-to-noise ratio conditions.

[0018] To prevent cascading failures on the mechanical side by cutting off signal interference, this invention logically concatenates the dimensionality-reduced environmental characteristic values ​​with a frequency-domain interference scalar containing a negative exponential convergence function, establishing a safety control boundary against floating-point overflow. Upon diagnosing an abnormal state, the system bypasses the delayed liquid level feedback and directly sends alternating forward and reverse motor drive electrical signals, utilizing pulsed mechanical shear force to physically break up the fluid medium, thus constructing a physical self-healing closed loop from confidence-based early warning to physical intervention.

[0019] While eliminating the aforementioned signal and mechanical failures, this invention introduces a time-period-based active recalculation mechanism to prevent cascading failures on the electrical side. The system utilizes the extremely high temporal tolerance of the electrical load to compensate for the extremely low physical tolerance of the mechanical equipment. This periodic cross-domain physical scheduling action, while purifying the extraction environment, objectively and periodically interrupts the continuous temperature rise process of the distribution busbar caused by the skin effect of high-frequency harmonics. This releases the thermal stress of physical nodes at the control level, generating a physical engineering benefit that synergistically extends the remaining lifespan of the distribution network's physical nodes. Attached Figure Description

[0020] Figure 1 This is an architecture diagram of a property equipment lifecycle management system based on multi-source data fusion, as described in this embodiment of the invention. Figure 2 This is a flowchart of the core control logic of the property equipment lifecycle management system based on multi-source data fusion in an embodiment of the present invention. Detailed Implementation

[0021] The above technical solutions will now be described in detail with reference to the accompanying drawings and specific embodiments to provide a better understanding of them. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments used only to explain the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. Furthermore, it should be noted that, for ease of description, only the parts related to the present invention are shown in the drawings, not all of them.

[0022] Example: Please refer to Figure 1 and Figure 2 The property equipment lifecycle management system, based on multi-source data fusion, is constructed using a heterogeneous physical network on the building's foundation. The data acquisition layer includes high-frequency current transformers installed inside the motor protectors of mechanical equipment, and a power quality analyzer connected in series at the incoming end of the distribution network's common connection point. The control center is deployed within an edge computing gateway. The execution layer includes target electrical loads supporting open charging protocols and mechanical equipment controllers. A physical connection is established between the data acquisition layer and the control center via an RS485 serial bus. The control center and the execution layer rely on an Ethernet link, encapsulating the Modbus-TCP protocol for the transmission and response of industrial-grade control data packets.

[0023] With the parallel input of multi-source physical signals, the control center executes data flow actions based on causal timing. The control center reads environmental operating condition data and calculates and outputs a first characteristic value of the mechanical equipment. When the first characteristic value is greater than a first preset threshold, the control center extracts electrical quality characteristic data and calculates the frequency domain interference scalar of the distribution network's point of common coupling to the stator current data. The control center divides the frequency domain interference scalar with the first characteristic value to generate a dimensionless ratio. Controlled by Boolean logic that the ratio is greater than a second preset threshold, the control center generates a first control command to the target electrical load and synchronously generates a first timing marker on the time axis. Within a first time interval defined by the first timing marker, the control center performs feature extraction on the stator current data to generate a second feature vector, and outputs an operating control signal to the mechanical equipment based on the second feature vector.

[0024] For the generation of the first feature value, the control center performs feature dimensionality reduction calculations on the environmental condition data. The control center obtains precipitation probability parameters by polling at fixed time steps through a pre-defined application programming interface. Simultaneously, the data acquisition layer acquires the absolute displacement values ​​of the float sensors within the sump. The control center performs mathematical integration on the absolute displacement values ​​along the time axis to generate the liquid level time integral value. The precipitation probability parameters and the liquid level time integral value together constitute the discrete input data column of the environmental condition data.

[0025] To map the discrete input data series into a dimensionless control scale, the control center invokes a weighted summation operator. The control center's storage unit contains a first weighting factor and a second weighting factor. The first weighting factor is statically calibrated based on the slope of the linear fit between historical precipitation and basement water accumulation rate; the second weighting factor is engineering-calibrated based on the volumetric critical value corresponding to the rated head of the sewage pump. The control center multiplies the precipitation probability parameter by the first weighting factor and simultaneously multiplies the liquid level time integral value by the second weighting factor. With the generation of these two independent product terms, the control center performs an algebraic addition operation on the two product terms. The weighted summation result is then assigned the first eigenvalue, which constitutes the fundamental value for the control center to calculate the frequency domain interferometric scalar.

[0026] Along with the generation of the first feature value, the control center performs a Boolean logic comparison based on a first preset threshold. When the first feature value is greater than the first preset threshold, the control center triggers a cross-domain solution process for the frequency domain interference scalar. The control center reads the preset feature frequency value associated with the mechanical equipment from its internal read-only memory node. Taking the physical scenario of fluid cavitation anomaly in a sewage pump as an example, the preset feature frequency value is calibrated as the high-frequency continuous band of 15kHz to 20kHz corresponding to the fluid cavitation anomaly. The extraction of the preset feature frequency value establishes a digital alignment target for the subsequent power grid frequency domain analysis actions of the control center.

[0027] Using a preset characteristic frequency as a comparison benchmark, the control center synchronously receives electrical quality characteristic data uploaded by the power quality analyzer via a serial bus. This electrical quality characteristic data includes a measured high-frequency harmonic array injected by the target electrical load into the distribution network's common connection point. The control center invokes the floating-point unit within its microprocessor to perform a Fast Fourier Transform algorithm on the measured high-frequency harmonic array, separating the fundamental frequency and extracting the first harmonic frequency value. Along with the output of the first harmonic frequency value, the control center maps the first harmonic frequency value and the preset characteristic frequency value to the same frequency domain coordinate dimension, driving the numerical integrator to perform frequency band overlap integral calculation. The area value calculated from the overlap integral is converted into a coherence coefficient, which, at the physical dimension level, characterizes the interference coverage strength of high-frequency noise in the power grid on abnormal current signals.

[0028] The control center uses the power operation result of the ratio of the measured total harmonic distortion rate to a first preset constant as the reference input, where the exponent of the power operation is a second preset constant. The control center multiplies the reference input by the coherence coefficient as the gain coefficient, takes the negative exponent, subtracts the negative exponent result from the constant 1, and multiplies it by the first eigenvalue to generate a frequency domain interferometric scalar. The specific calculation logic formula is as follows: In the formula, For frequency domain interference scalar, The first eigenvalue, It is an exponential function with the natural constant as its base. The coherence coefficient, To measure the total harmonic distortion rate, This is the first preset constant. It is the second preset constant and .

[0029] The control center compares the frequency domain interference scalar with a second preset threshold. If the frequency domain interference scalar exceeds the second preset threshold, the control center generates a first control command for the target electrical load in the memory buffer. This first control command is reassembled and encapsulated in the format of an industrial Ethernet data packet. The control center writes the device communication identifier (MTI) and media access control address of the target electrical load into the data link layer protocol header. In the application layer control load area, the control center writes the register instruction corresponding to the target power limit value, for example, issuing an instruction to clamp the upper limit of the target electrical load's output power to 3500W. Simultaneously, the control center writes an execution time period parameter to the end of the data packet. The device communication identifier, the target power limit value, and the execution time period are collectively fixed into the complete data frame structure of the first control command.

[0030] Simultaneously with issuing the first control command, the control center generates a first timing marker associated with the first control command on the timeline. The control center invokes a precise time protocol mechanism to detect the network transmission delay of the Ethernet link and retrieves the physical switching response time parameters of the power electronic insulated gate bipolar transistors (IGBTs) within the target electrical load from the device capability database. The control center performs a mathematical summation operation on the network transmission delay and the physical switching response time parameters, injecting the summation result as a time offset compensation value into the starting timestamp of the first timing marker. The time offset compensation mechanism aligns the starting timestamp of the first timing marker with the physical moment when the target electrical load's power drops. Within the first time interval defined by the first timing marker, the target electrical load maintains its electrical parameter output state after power reduction. The control center triggers a high-frequency analog-to-digital converter to extract features from the stator current data, outputting a second feature vector characterizing the operating state of the mechanical equipment.

[0031] With the generation of the second feature vector, the control center initiates a matching procedure based on the Dynamic Time Warping (DTW) algorithm. The fault feature database pre-stores reference feature sequences for various abnormal states. The control center maps the second feature vector and the reference feature sequences onto a two-dimensional distance matrix, calculates the cumulative distance between each sampling point and its corresponding point in the reference feature sequence, and searches for the minimum cost path in the distance matrix. When the calculated minimum cumulative distance falls within a preset matching threshold, the control center determines that the current mechanical equipment is abnormal.

[0032] After determining that the mechanical equipment is in an abnormal state caused by high-viscosity mud, the control center generates a second control command. This second control command contains characteristic parameters of a high-frequency alternating torque electrical signal. Upon receiving the second control command, the execution layer drives the mechanical equipment's motor to perform a forward / reverse switching action. The pulsed mechanical shear force generated by this action acts on the mud medium surrounding the impeller of the mechanical equipment, reducing the mud's viscosity through physical fragmentation and clearing the blockage. This process achieves a closed-loop control system from electrical signal feature identification to changes in the physical state of the equipment.

[0033] For steady-state conditions where mechanical diagnostic actions have not been triggered for an extended period, the control center incorporates a system clock node. This system clock records the timestamp of the last time the first control command was generated. The control center calculates the difference between the current physical time and this timestamp. When this difference exceeds a preset time period, and the electrical quality characteristic data uploaded by the data acquisition layer is greater than a third preset threshold, the control center bypasses the first characteristic value's judgment restriction and restarts the extraction and calculation process for the frequency domain interference scalar. This periodic triggering mechanism establishes the system's proactive detection capability of the electrical and physical environment during long-term stable operating conditions.

[0034] The physical benchmark for the first preset constant is established above the signal-to-noise ratio avalanche critical point of the Hall current sensor. As the total harmonic distortion (THD) rate increases in the distribution network, the noise floor of the Hall current sensor tends to expand, causing the characteristic signal to be submerged by noise. The control center selects the distortion rate value at which the sensor's THD rate drops to the failure threshold as the first preset constant. In the calculation, the measured THD rate is compared with the first preset constant, achieving normalization of the physical dimensions.

[0035] The second preset constant, the sensitivity order, is used to limit the steepness of the change in the interference scalar with environmental variations. The sensitivity order is calibrated between 1.5 and 3.0. This range is derived from the physical experimental calibration process of injecting harmonics into charging devices of different power levels. When the sensitivity order equals 1, the rate of change of the interference scalar in the critical region is gentle, causing a delay in control action; when the sensitivity order is greater than 3, the system is sensitive to harmonic fluctuations in the distribution network, leading to high-frequency command triggering. The value range of 1.5 to 3.0 achieves an engineering balance between system response sensitivity and hardware action stability.

[0036] The control center establishes the physical boundaries of hardware control actions at the arithmetic level. Executing mathematical limit derivations, when the measured total harmonic distortion (THD) approaches positive infinity, the ratio of the measured THD to the first preset constant also approaches positive infinity. Accompanying algebraic operations with a sensitivity order greater than 1, the product term within the negative exponential function approaches negative infinity. Under the mapping of the natural logarithm base, the exponential term containing the negative infinity independent variable converges to 0. Controlled by this, the algebraic subtraction result within the brackets converges to 1. The calculation result of the frequency domain interference scalar is clamped and equivalent to the first eigenvalue. This convergence mechanism, at the underlying hardware instruction level, prevents floating-point overflow anomalies in the microprocessor's arithmetic logic unit. The output numerical domain of the frequency domain interference scalar is enveloped within a closed interval from 0 to the first eigenvalue. Since the system uses the ratio of the frequency domain interference scalar to the first eigenvalue as the final trigger criterion, this closed interval property ensures that the ratio parameter involved in the judgment strictly converges to the positive real number interval from 0 to 1. This ratio determination logic completely eliminates the control dead zone loophole caused by the absolute value fluctuation of the first characteristic value, and establishes the numerical boundary for generating the trigger condition of the first control command.

[0037] For the control center without a hardware floating-point unit, an approximate operation mechanism combining static lookup and linear interpolation is implemented. The control center's read-only memory has a contiguous address space containing a one-dimensional static data table mapping the discrete coordinates of the negative exponential function. During the negative exponentiation operation, the control center extracts the integer part of the product of the coherence coefficient and the reference input, converts it into an address pointer, and reads the corresponding reference convergence value from the static data table. The control center extracts the fractional part of this product and inputs it to the arithmetic logic unit (ALU) to perform linear interpolation on adjacent reference convergence values. The approximate exponentiation result output by the linear interpolation is pushed into a register and participates in subsequent algebraic subtraction and multiplication of the first eigenvalue. The execution path based on the static lookup mechanism combined with linear interpolation maps the occupancy status of the CPU registers and outputs the numerical calculation result.

[0038] The control center stores a sequence of replacement algorithms containing convergence functions in its memory unit. The control center extracts rational fractional convergence functions or hyperbolic tangent functions to replace negative exponential convergence terms in the computational model. When executing the rational fractional convergence function, the control center assigns the ratio of the measured total harmonic distortion rate to a first preset constant as the reference input, and performs a power operation based on the sensitivity order on the reference input to generate a power exponent term. The control center inputs a calculation instruction to the arithmetic logic unit to calculate the ratio parameter of the power exponent term and the power exponent term after adding 1. As the reference input value shifts upward, the ratio parameter outputs a monotonically increasing property. When the reference input value approaches positive infinity, the ratio parameter converges to a constant of 1. The control center extracts coherence coefficients to adjust the gain of the reference input and performs a multiplication operation between the ratio parameter and the first eigenvalue. The frequency domain interference scalar output by the multiplication operation exhibits a numerical evolution trajectory that monotonically increases with the increase of the reference input value and tends towards the first eigenvalue. When performing a task using the hyperbolic tangent function, the control center uses the asymptote property of the hyperbolic tangent function to map the adjusted reference input quantity, which in turn drives the frequency domain interference scalar to tend towards the first eigenvalue.

[0039] The control center's physical substrate integrates a reduced instruction set microprocessor (RISC) and a field-programmable gate array (FPGA). The RISC's internal universal asynchronous transceiver (API) receives environmental condition data and electrical quality characteristic data uploaded from the data acquisition layer. The FPGA contains hardwired digital signal processing circuitry that handles the sampling of stator current data and independently performs Fast Fourier Transform (FFT) and coherence coefficient decimation operations based on the hardwired circuitry. The coherence coefficients are transmitted to the RISC microprocessor via an internal parallel bus. The RISC microprocessor uses internal registers to perform algebraic calculations and compare binary values ​​for thresholds.

[0040] The physical architecture of the control center includes a non-volatile memory chip. The memory chip has a contiguous address space, within which an instruction set containing opcodes and operands is programmed. When the program counter of the reduced instruction set microprocessor (RISC) points to the contiguous address space and fetches the instruction set, the microprocessor drives the peripheral hardware communication ports to perform physical actions such as acquiring feature data, executing logical decisions, generating the first timing marker, and sending electrical pulse signals to the target electrical load. The level switching process triggered by the instruction set is transformed into an industrial control process that changes the output state of the target electrical load and the physical operating parameters of the mechanical equipment.

[0041] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A property equipment lifecycle management system based on multi-source data fusion, characterized in that, include: The system comprises a data acquisition layer, a control center, and an execution layer, wherein the execution layer includes mechanical equipment and target electrical loads. The data acquisition layer is configured to: acquire environmental operating condition data of the mechanical equipment and stator current data of the mechanical equipment; and simultaneously acquire electrical quality characteristic data of the common connection point of the power distribution network shared by the mechanical equipment and the target electrical load; The control center is configured to execute the following control logic: Based on the environmental condition data, calculate the first characteristic value of the mechanical equipment; When the first feature value is greater than the first preset threshold, the frequency domain interference scalar of the distribution network common connection point to the stator current data is calculated based on the electrical quality feature data. Calculate the ratio of the frequency domain interferometric scalar to the first eigenvalue; When the ratio is greater than the second preset threshold, a first control command for the target electrical load is generated, and a first timing mark associated with the first control command on the time axis is generated. Within a first time interval defined by the first timing marker, feature extraction is performed on the stator current data to generate a second feature vector; The target electrical load is configured to maintain a preset electrical parameter output state within the first time interval in response to the first control command. The control center is configured to output operation control signals for the mechanical equipment based on the second feature vector.

2. The property equipment lifecycle management system based on multi-source data fusion as described in claim 1, characterized in that, The mechanical equipment is a sewage pump, and the target electrical load includes a charging device with an inverter or rectifier topology.

3. The property equipment lifecycle management system based on multi-source data fusion according to claim 1, characterized in that, The environmental condition data includes precipitation probability parameters and liquid level time integral values; the control center is configured to use a preset first weighting factor and a second weighting factor to perform weighted summation calculations on the precipitation probability parameters and the liquid level time integral values ​​respectively, and output the weighted summation result as the first feature value.

4. The property equipment lifecycle management system based on multi-source data fusion as described in claim 1, characterized in that, The electrical quality characteristic data includes the measured total harmonic distortion rate; when the control center calculates the frequency domain interference scalar, the input parameters extracted include the measured total harmonic distortion rate, the first preset constant, the second preset constant, and the coherence coefficient between the first harmonic frequency value of the distribution network common connection point and the preset characteristic frequency value of the mechanical equipment.

5. The property equipment lifecycle management system based on multi-source data fusion according to claim 4, characterized in that, The control center is configured to calculate the frequency domain interference scalar using the following logic: The result of the power operation of the ratio of the measured total harmonic distortion rate to the first preset constant is used as the reference input, wherein the exponent of the power operation is the second preset constant; The coherence coefficient is used as the gain coefficient, which is then multiplied by the reference input and the result is calculated using a negative exponent. The frequency domain interferometric scalar is generated by subtracting the negative exponent from the constant and multiplying it by the first eigenvalue.

6. The property equipment lifecycle management system based on multi-source data fusion according to claim 5, characterized in that, The control center calculates the frequency domain interference scalar using the following formula: in, For the frequency domain interference scalar, The first feature value, It is an exponential function with the natural constant as its base. The coherence coefficient is... The measured total harmonic distortion (THD) is... It is the first preset constant. It is the second preset constant and .

7. The property equipment lifecycle management system based on multi-source data fusion according to claim 5, characterized in that, The control center is equipped with a read-only memory containing a one-dimensional static data table. When performing the negative exponentiation operation, the control center extracts the integer part of the product of the coherence coefficient and the reference input quantity and converts it into an address pointer. It reads the reference convergence value from the one-dimensional static data table and extracts the fractional part of the product to perform linear interpolation operation on adjacent reference convergence values. It outputs an approximate exponentiation result to participate in the calculation of the frequency domain interferometric scalar.

8. The property equipment lifecycle management system based on multi-source data fusion according to claim 1, characterized in that, The first control instruction includes a device communication identifier, a target power limit value, and an execution time period; the start timestamp of the first timing marker is the same as the start timestamp of the execution time period, and the length of the first time interval is equal to the length of the execution time period.

9. The property equipment lifecycle management system based on multi-source data fusion according to claim 1, characterized in that, When the second feature vector matches a preset abnormal feature pattern, the control center sends a second control command to the mechanical equipment; the second control command includes alternating forward and reverse motor drive electrical signals.

10. The property equipment lifecycle management system based on multi-source data fusion according to claim 1, characterized in that, When the time interval since the last generation of the first control command is greater than a preset time period, and the electrical quality characteristic data is greater than a third preset threshold, the control center restarts and executes the calculation and judgment control logic of the frequency domain interference scalar.

Images